# About Name: GQS Direct,LLC URL: https://gqsdirect.com/blog # Navigation Menu - Home: https://gqsdirect.superblog.click/ - Sample Page: https://gqsdirect.superblog.click/sample-page - Search: https://gqsdirect.superblog.click/search - GQS Direct Insights: https://gqsdirect.com/ # Blog Posts ## Your Attribution Report Is Taking Credit for Sales You’d Have Made Anyway Author: Gary Q. Smith Author URL: https://gqsdirect.com/blog/author/gary-q-smith Published: 2026-10-31 Category: Marketing Insights Category URL: https://gqsdirect.com/blog/category/marketing-insights Meta Title: Your Attribution Report Is Taking Credit | GQS Direct Meta Description: Attribution divides up the credit; it doesn’t prove cause. How holdout groups and honest A/B tests show which sales your marketing actually created. URL: https://gqsdirect.com/blog/attribution-vs-incrementality-holdout-testing **I’ve seen this movie before.** In the catalog business, nobody trusted a response rate on its own. Before a circulation plan went to the printer, someone asked the question that kept the business honest: _what would these customers have bought if we hadn’t mailed them?_ We answered it the hard way. Test cells. Holdout panels. Matchback reports. It was slow, unglamorous work, and it’s the reason direct marketing earned its reputation for accountability. **Then the dashboards arrived.** Today every channel ships with a report showing how much revenue it produced. Email claims the order. Paid search claims the order. Retargeting claims the order. Add up the claims and you’ve often sold more than you actually sold. Those reports aren’t lying. They’re answering a different question from the one you need answered. I’ve planned circulation for 40 million catalogs a year, moved 44% of a company’s sales to the web, built marketing databases since the 1990s, and spoken at DMA conferences on lifetime value and segmentation. So take this from someone who has actually run the business: > “A report tells you who got the credit. Only a test tells you what you caused.” Don’t measure this by next quarter’s dashboard. Ask where your marketing budget sits **three to five years from now** if every dollar keeps flowing to the channel that’s best at _claiming_ sales. Then ask where it sits if the money follows the channels that actually _create_ them. Here are the four lessons that separate the two. ## Lesson One: Start with a question, not a report The scientific method is the oldest playbook in direct marketing. Ask. Hypothesize. Test. Measure. Learn. Improve. Then do it again. ![The scientific method as a five-step cycle: form a hypothesis, design the test, measure results, analyze and learn, improve](https://prod.superblogcdn.com/site_cuid_cmvlko1f7002t01xk5ihgbmf7/images/sm-fig1-scientific-method-cycle1-1793455383364-compressed.png) _Figure 1. The scientific method is a repeatable learning cycle._ Most teams skip the first step. They start with a report and work backward to a conclusion. Consider two questions that sound almost identical: - _Which email generated the most revenue?_ - _Does sending this email increase total purchases compared with not sending it?_ Your email platform answers the first one every morning. It can’t answer the second one at all, because that requires knowing what happened to the people who _didn’t_ get the email. That’s the whole game. Testing two subject lines is useful work, but it’s testing _routinely_, not testing _rigorously_. Rigor starts with a hypothesis written down before anything goes out: - A catalog plus an email will produce more incremental profit than the catalog alone. - Free shipping will produce more contribution margin than 15% off. - Cutting promotional frequency will lower costs without materially reducing purchases. A hypothesis forces you to say what you expect to change and how you’ll know whether it mattered. The goal isn’t to prove your latest idea right. It’s to find out what actually works. ## Lesson Two: Credit is not cause Here’s a customer journey you’ll recognize. A catalog lands. A few days later she searches your company name, clicks the paid ad, browses, gets an email, and orders. ![A multi-channel customer journey from catalog to paid search, email, social, website and purchase, with attribution and incrementality contrasted side by side](https://prod.superblogcdn.com/site_cuid_cmvlko1f7002t01xk5ihgbmf7/images/sm-fig2-attribution-vs-incrementality1-1793457173492-compressed.png) _Figure 2. Attribution allocates credit across touchpoints; incrementality asks which sales the marketing actually created._ **Who gets the credit?** That depends on the attribution model: last click, first touch, a split, or an algorithm’s best guess. Each tells you something. **None of them proves the marketing caused the sale.** The catalog may have prompted it. The ad may have helped her find you. Or she’s bought from you for ten years and was going to reorder no matter what. That’s the difference between two words that get used interchangeably and shouldn’t be: - **Attribution** divides up the credit for sales that happened. - **Incrementality** counts the sales that happened _because of_ the marketing. The gap between them is widest exactly where money is easiest to spend: loyal customers, branded search, retargeting, promotional email, and mail to people who already know you. The most famous proof came from eBay. Its economists switched off paid ads on eBay’s own brand name and watched what happened. **99.5% of the lost click traffic** came straight back through the free search listings. The ads had been claiming credit for customers who were arriving anyway. (Blake, Nosko and Tadelis, _Econometrica_, 2015.) That’s not villainy. It’s an unforced error. It also works in reverse: a channel that rarely gets the last click may be doing heavy lifting that no report will ever show you. The answer isn’t to throw out attribution. It’s to stop treating it as proof. ## Lesson Three: Do the math on the holdout Here’s a simplified example. The numbers are illustrative, but the arithmetic is real. You mail a promotion to **10,000** customers and get **500** orders, a 5% response rate. The report says the campaign drove 500 orders. Now suppose you held back a random group of comparable customers who got nothing, and they ordered at **3%** during the same window. **What the promotion actually produced** Measurement Result Customers mailed 10,000 Orders from the mailed group 500 Orders expected anyway (3% holdout rate) 300 **Incremental orders** **200** Three hundred of those orders were coming in the door regardless. The promotion earned two points of response, not five. Now put a P&L on it. Assume the piece costs **$0.80** in the mail and each order contributes **$40** after cost of goods and fulfillment. These are assumptions, so plug in your own. **Same campaign, two scorecards** Attributed Incremental Orders counted 500 200 Contribution at $40 each $20,000 $8,000 Campaign cost $8,000 $8,000 **Net contribution** **+$12,000** **$0, break-even** Same campaign, same mail drop. One scorecard says repeat it and add volume. The other says it barely paid its postage. Multiply that gap across every program you run in a year and you see why this matters. Budget, contact frequency, channel mix and the offer itself all shift once you know which sales you caused. > **A rule I live by** > > Never judge a campaign by the orders it _received_. Judge it by the orders it _created_. > > If you don’t have a control group, **admit you don’t know which is which.** ## Lesson Four: Test the details with A/B. Test the program with a holdout. These are two tools for two different jobs, and confusing them is one of the most common mistakes I see. ![Two complementary testing tools: an A/B test splits an audience between two versions; a holdout test splits it between a campaign group and a control group](https://prod.superblogcdn.com/site_cuid_cmvlko1f7002t01xk5ihgbmf7/images/sm-fig3-ab-vs-holdout-1793457331128-compressed.png) _Figure 3. A/B tests compare campaign alternatives; holdout tests compare marketing exposure with a control._ An **A/B test** answers _which version works better?_ Two headlines, two offers, two formats, two landing pages: split a random audience and run them at the same time. A **holdout test** answers _does this program add anything at all?_ Randomly withhold the campaign from part of the eligible audience, say 10% of your list for one promotional email. After a full buying cycle, compare purchases, revenue per customer and contribution profit between the two groups. If they look the same, the email is mostly taking credit, whatever the platform report says. The same principle works for catalogs, paid media and contact strategy. When randomizing individuals isn’t practical, hold out whole markets: matched geographies that don’t get the campaign. The disciplines that make either test trustworthy: - **Change one thing.** Change the headline, offer, format and audience at once, and you’ll learn which version won without learning why. - **Measure what the business cares about.** Opens and clicks diagnose. Orders, contribution margin and lifetime value decide. - **Set sample size and duration before you start.** Small samples and short windows produce confident nonsense. - **Don’t stop when you’re ahead.** Peeking daily and calling the test the moment your favorite leads is how false signals get funded. - **Write it down.** Record the hypothesis, audience, design, result, limitations and decision. Over time that log becomes institutional memory instead of a pile of campaign reports. Yes, a holdout costs you some sales during the test. Weigh that against knowing whether a program you’ll run fifty more times is actually profitable. You don’t need to test everything. Test the decisions where better evidence changes the money. > **A word to the next generation** > > The tools change every few years. The ability to design a fair test never goes out of style. > > Learn to write a hypothesis, pick a control group, and read a result honestly, including the results you didn’t want. **The people who can do that end up running the budget.** > **Your next step** > > Pick one program whose real contribution you’re unsure of: branded search, a promotional email series, a retargeting campaign. > > Write the hypothesis. Hold out 10% at random. Agree on the success metric and the measurement window _before_ launch. **Then act on what you find.** ## If I Sat on Your Board: The question I’d ask The biggest obstacle isn’t technology. It’s habit. Reporting systems are built to show what happened, not what caused it, and nobody enjoys learning that a long-standing program isn’t pulling its weight. ![Four compounding returns from better testing — reduced waste, improved profitability, challenged assumptions and new opportunities — alongside a holdout example in which a promotion lifted orders from 3% to 5%, making 200 of 500 orders incremental Delete this line once the image is in](https://prod.superblogcdn.com/site_cuid_cmvlko1f7002t01xk5ihgbmf7/images/sm-fig4-the-payoff-is-real-1793457573502-compressed.png) _Figure 4. The return compounds: less waste, better profitability, assumptions challenged, opportunities found._ But every honest test pays off in one of three ways. It confirms that something works, so you can fund it with confidence. It exposes something that doesn’t, so you can stop. Or it uncovers an opportunity your reports were hiding. All three are worth the price of the test. So look at the marketing budget and ask your team: > Of the sales our marketing gets credit for, how many would we have made anyway? If nobody can answer with a control group behind it, you’re steering by the dashboard instead of the road. Direct marketing has always had the edge of connecting what we do to what customers buy. The scientific method is how we keep it. You’re not behind. You just need to start testing again. ![ive practical steps: start with a hypothesis, use A/B testing, explore holdout groups, document and learn, make it continuous ](https://prod.superblogcdn.com/site_cuid_cmvlko1f7002t01xk5ihgbmf7/images/sm-fig5-practical-steps-1793457855842-compressed.png) _Figure 5. Five steps to put the method back to work._ **Measure what your marketing creates. I’m glad to help you get there.** > **About the author** > > **Gary Q. Smith** — President, GQS Direct, LLC > > Gary pairs four decades of P&L-accountable direct marketing leadership with practitioner-level command of today’s commerce stack. He is currently an AI evangelist and practitioner and has led four e-commerce website implementations and integrations, built marketing databases, was architect and project leader for a custom ERP system, and managed omnichannel marketing strategy and communications for several businesses. > > _Most consultants have studied these problems. GQS Direct has owned them._ > > [**Start with a complimentary consultation →**](https://gqsdirect.com/booking.html) --- This blog is powered by Superblog. Visit https://superblog.ai to know more. --- ## Your B2B Buyer Is Not a Consumer With a Bigger Cart Author: Gary Q. Smith Author URL: https://gqsdirect.com/blog/author/gary-q-smith Published: 2026-10-30 Category: E-Commerce Category URL: https://gqsdirect.com/blog/category/e-commerce Meta Title: Your B2B Buyer Is Not a Consumer | GQS Direct Meta Description: Business buyers need company accounts, negotiated pricing, approvals and POs. Shopify, Adobe Commerce and OroCommerce on capability and TCO. URL: https://gqsdirect.com/blog/b2b-ecommerce-platform-decision She needs a purchase order number, two internal approvals, and the price her company negotiated last March. She is looking at a checkout built for somebody buying a sweatshirt. So she does what every business buyer in America does when the website can’t do her job. She closes the tab, picks up the phone, and your customer service rep types the order in by hand — and now you are paying a salary to be a website. **That’s not a conversion problem. That’s an architecture problem.** Before I go further, you should know where I stand. **I’m a Shopify advocate.** I was the decision-maker and project leader on a Shopify implementation for a custom products manufacturer that required heavy customization, and it worked. That’s a bias, and you’re entitled to know about it before you read my comparison. It also means something else: when I tell you not to use Shopify, pay attention. That part’s coming. Module 04 taught you to count the three-year number. This module is about where you spend it. I’ve led four e-commerce implementations and served as architect and project leader on a custom ERP replacement, so take this from someone who has actually run the business: > “Retrofitting a consumer storefront to handle business purchasing is how you end up with a website that takes orders and a staff that fixes them.” Four lessons, two tables, and the question I’d ask if I sat on your board. ## Lesson One: Business buyers don’t shop. They procure. Consumer commerce is one person, one card, one decision. Business commerce is none of those things. Five capabilities decide whether a platform can carry your B2B business. Not fifty. Five. **Company accounts with roles.** A buyer, an approver, a controller — one login each, different permissions, under one company. And one human who buys for three subsidiaries without three logins. **Negotiated pricing.** Not a discount code. A price list that belongs to that customer, honors their contract, and never shows up in front of anyone else. **Approval workflows.** Over five thousand dollars, it routes to the manager. That rule lives in the platform or it lives in somebody’s inbox. **Quotes and RFQ.** A request comes in, a rep prices it, the buyer counters, it converts to an order. This is a negotiation, not a cart. **Terms and purchase orders.** Net 30. A PO number captured at checkout, because without it their AP department won’t pay you. Miss any one of them and you don’t have a B2B site. You have a B2C site with a manual process bolted to the back of it — and that process has a payroll cost that grows with your success. > **A rule I live by** > > Write the five down before you take a single demo. > > Every vendor can show you a beautiful storefront. Make them show you a purchase order routing for approval, with your actual dollar threshold, in your actual catalog. The demo that gets vague right there has answered you. ## Lesson Two: April 2, 2026 moved the entry price For years the answer to “I need B2B on Shopify” was “upgrade to Plus, $2,300 a month.” That stopped being true this spring. On **April 2, 2026**, Shopify moved a block of core B2B capability down to the Basic, Grow and Advanced plans at no additional cost: **company profiles, payment terms, volume pricing, ACH payments (US only), vaulted credit cards, and up to three active B2B catalogs** assigned through Markets. Read the plan prices against that. Basic is $39 a month. Grow is $105. Advanced is $399. Plus is $2,300 on a three-year term, $2,500 on one year. A wholesale operation that needs three price tiers and Net 30 can now run on a **$399-a-month** plan. Three years of that is about $14,400. Three years of Plus is nearly $83,000. That is the single most useful fact in this article, and most of the B2B comparisons you’ll read were written before it was true. Now the other half, because this is where people get hurt. These stayed behind on Plus: - **Unlimited catalogs** for customer-specific pricing — more than three price lists and you’re going up - **Direct catalog assignment** to companies and locations - **Partial payments and deposits** So the real question isn’t “do I need Plus.” It’s **“how many distinct price lists do I actually have?”** Three or fewer, and you have a very inexpensive answer. Four or more, or customer-specific pricing at account level, and you’re on Plus — and that’s a legitimate reason to be there. ## Lesson Three: The architecture is the cost driver, not the license Module 04’s argument was that the license is the smallest number. It’s just as true here, and in B2B it shows up as a question about **where your customization lives**. **Shopify** keeps the core lean and pushes specialized capability to its app ecosystem. That’s a deliberate strategy and it works — but it means evaluating Shopify honestly requires you to evaluate the apps too. Quote negotiation, sales-rep ordering, tiered approvals: those are app territory. The companion piece to this module maps three configurations. **Adobe Commerce** is now two different products wearing one name, and this is the trap. The PaaS edition is what people mean when they say Magento — full PHP control, the Marketplace extension ecosystem, deep native B2B. **Adobe Commerce as a Cloud Service**, introduced in June 2025, is the SaaS edition. Adobe runs it, patches it, upgrades it versionlessly — all the things Module 04 said you should want. And on it, **PHP modules, Luma themes, and Magento Marketplace extensions do not run.** Customization is out-of-process only: App Builder, API Mesh, webhooks. Base license limits run to 250,000 SKUs and 100 catalog variations. B2B itself is there — Adobe shipped company management, purchase orders, quote management and requisition lists as storefront drop-ins beginning January 2026, with seller-assisted buying later in the year. Coverage is real but still filling in, so verify each capability you need against the current list rather than the brochure. Here’s the honest reading: ACCS fixes the upgrade treadmill that Module 04 was about. It also removes the extension ecosystem that is most of why people chose Adobe Commerce for B2B. **You can have the maintenance relief or the extension library. Not both.** Go in knowing which one you’re buying. **OroCommerce** is the one built for B2B from the start, and its pricing model has a detail worth knowing before you call. Oro licenses by **GMV bands** — the value of goods sold through the platform — and publishes no list price. Enterprise plans **include 25 admin users**. Your buyers and customer accounts don’t count against that; your internal people do. For a distributor with a large inside sales team, that is a real number, and it is not on any comparison chart. **The capability comparison** Capability Shopify + apps Adobe Commerce OroCommerce Best fit SMB to mid-market; hybrid B2B/B2C Mid-market to enterprise Mid-market to enterprise, B2B-first Company accounts & roles Native on eligible plans Strong native controls Strong, core to the product Customer-specific pricing 3 catalogs below Plus; unlimited on Plus Shared catalogs, advanced pricing Native price lists Quotes / RFQ App territory Built-in Core workflow Sales-rep-assisted ordering Apps and integrations Native, extensible Designed for it ERP integration Broad app ecosystem; custom work likely Extensive; custom development likely Native connectors for major ERPs Where customization lives Apps + Shopify’s architecture PaaS: PHP. ACCS: out-of-process only B2B workflow config + custom code Maintenance burden Lowest Highest on PaaS; low on ACCS Depends on deployment _Strategic comparison, not a quote. Capability varies by plan, edition, deployment model and implementation partner._ ## Lesson Four: Run the RFP, not the demo The three-year numbers, as planning ranges: **Projected three-year total cost of ownership** Cost category (3 years) Shopify + apps Adobe Commerce OroCommerce Platform license $14,400 (Advanced) – $90,000 (Plus) $66,000 – $375,000+ GMV-banded; quote only Hosting & infrastructure Included $36,000 – $180,000 (PaaS); included on ACCS Included on OroCloud Initial implementation $50,000 – $200,000 (8–16 wks) $100,000 – $400,000+ (16–40 wks) $75,000 – $250,000+ B2B customization & apps $36,000 – $104,000 $26,000 – $130,000 Largely included ERP integration $20,000 – $60,000 $30,000 – $100,000 Lower; native connectors Maintenance & upgrades $36,000 – $144,000 $102,000 – $378,000 ~15–25% of license **Three-year total** **$156,000 – $598,000** **$360,000 – $1,563,000+** **~$350,000 – $900,000+** _Two honesty notes. Adding every low together and every high together produces a wider band than any real project will land in — use these to size the decision, not to budget it. And Shopify applies a variable platform fee at higher volumes that it does not publish; if you’re approaching that scale, get the threshold and the cap in writing before you sign._ So don’t shop. Procure. 1. **Document first.** Strategy, three-year objectives, critical workflows, requirements. Don’t let a demo do your thinking. 2. **Issue an RFP** and require a written three-year TCO in every response. 3. **Demand demos and references** — customers your size, in your model, who will take your call. 4. **Calculate ROI yourself.** The vendor won’t, and executive management isn’t approving a platform. They’re approving a return. For Shopify or Adobe, run the RFP through a certified solution partner rather than the vendor’s sales desk. And the order I’d evaluate in: **Shopify first.** Start here if your B2B requirements can be met natively or with modest app support — shortest timeline, lowest cost, least implementation risk. Especially if you’re hybrid B2B/B2C or D2C-first under $200 million. **Don’t choose it if a key requirement can’t be met to your satisfaction.** That’s the part people skip. **OroCommerce to the top** if you’re a B2B-first manufacturer, industrial distributor, or multi-entity wholesaler where complex corporate buying rules _are_ your revenue operation. Demo it against your actual contract pricing and approval hierarchies. **Adobe Commerce** if you need extensive enterprise customization and have the technical resources to carry it — massive catalog complexity, 100,000+ SKUs, multi-brand regional networks from one admin. And decide PaaS versus ACCS deliberately, because that choice is irreversible in practice. > **A word to the next generation** > > Learn to run a real RFP. Requirements first, scorecard second, demos last. > > Anyone can be impressed by a demo. The professional is the one who wrote down what mattered _before_ the sales engineer walked in — and that discipline will outlast every platform on this page. > **Your next step** > > Count your distinct price lists. Then count your approval rules. Then count the orders your staff types in by hand every week. > > Three numbers, one afternoon. They’ll narrow this decision faster than any comparison chart — including mine. ## If I Sat on Your Board: The question I’d ask Not which platform. > How many orders did your people key in by hand last month, and what did that cost us in salary? Three possible answers. _We don’t track that_ means the cost is real and invisible. _A few hundred_ means you have a platform problem you’ve been calling a staffing problem. _Eleven, and here’s why each one was an exception_ means your architecture fits your business. Your buyer already knows what good looks like. She uses it at home every night. **She’s not asking you to be Amazon. She’s asking you to take her purchase order.** _Next in this series —_ [**_Module 06: Your Next Customer Won’t Have Eyes_**](https://gqsdirect.com/blog/agentic-commerce-machine-readable-catalog) _Agentic commerce arrived as a discovery channel, not a checkout — what to fix in your product data before the machines decide you don’t exist._ > **About the author** > > **Gary Q. Smith** — President, GQS Direct, LLC > > Gary pairs four decades of P&L-accountable direct marketing leadership with practitioner-level command of today’s commerce stack. He is currently an AI evangelist and practitioner and has led four e-commerce website implementations and integrations, built marketing databases, was architect and project leader for a custom ERP system, and managed omnichannel marketing strategy and communications for several businesses. > > _Most consultants have studied these problems. GQS Direct has owned them._ > > [**Start with a complimentary consultation →**](https://gqsdirect.com/booking.html) **The E-Commerce Implementation Series** - [Module 01 — Re-Platforming Is a Systems Project, Not a Web Design Project](https://gqsdirect.com/blog/replatforming-is-a-systems-project) - [Module 02 — Your Redirect Map Is Worth More Than Your Theme](https://gqsdirect.com/blog/redirect-map-seo-preservation-migration) - [Module 03 — Middleware: The Key to Effective Integration and Project ROI](https://gqsdirect.com/blog/middleware-integration-project-roi) - [Module 04 — The Open-Source Paradox: Hidden Maintenance and DevOps Costs](https://gqsdirect.com/blog/open-source-paradox-three-year-tco) - Module 05 — Your B2B Buyer Is Not a Consumer With a Bigger Cart _(you are here)_ - [Module 06 — Your Next Customer Won’t Have Eyes](https://gqsdirect.com/blog/agentic-commerce-machine-readable-catalog) - [Module 07 — The Re-Platforming Payoff](https://gqsdirect.com/blog/replatforming-payoff) --- This blog is powered by Superblog. Visit https://superblog.ai to know more. --- ## The Re-Platforming Payoff Author: Gary Q. Smith Author URL: https://gqsdirect.com/blog/author/gary-q-smith Published: 2026-10-30 Category: E-Commerce Category URL: https://gqsdirect.com/blog/category/e-commerce Meta Title: The Re-Platforming Payoff | GQS Direct Meta Description: Three months after launch the board asks what you got for the money. The customer record, the capacity freed, the flows that replaced blasts. URL: https://gqsdirect.com/blog/replatforming-payoff Three months after launch, somebody on your board is going to ask the question. _So what did we get for the money?_ And if the only answer in the room is “the site is faster and it looks better,” you are about to have a very bad meeting. Because speed and design were never worth the number on that invoice. **You didn’t buy a storefront. You bought a set of things the old platform made impossible.** I’ve led four e-commerce implementations and served as architect and project leader on a custom ERP replacement. I’ve also sat on the other side of this — the Shopify implementation where I was the decision-maker and project leader, for a custom products manufacturer. So take this from someone who has actually run the business: > “The payoff was never the website. It was everything the old one wouldn’t let you do.” Four lessons on where the money actually shows up, the numbers from my own project, and the question I’d ask if I sat on your board. ## Lesson One: The data was always yours. The old platform just wouldn’t hand it over. Here is the part that stings: your legacy platform had most of this data the whole time. It had the order history. It had the browse behavior. It knew who bought twice and who bought once and vanished. It just couldn’t give any of it to you in a form you could act on while it still mattered — because it lived in tables nobody could join, behind an export that ran overnight, in a system that spoke to nothing else. Module 01 called a re-platform a systems project, not a web design project. **This is the bill coming due in your favor.** What you bought was an addressable customer record, and it has three layers: - **Identity** — one key per customer, with email, phone, location, time zone, and consent on record. Not three conflicting records across three systems. - **Transaction history** — order count, lifetime spend, average order value, time since last purchase, the SKUs and categories they actually buy. - **Behavior, in real time** — products viewed, collections browsed, internal searches run, carts opened and abandoned, checkouts started and left. That third layer is the one legacy platforms never delivered, and it’s the one that changes what marketing can do. Not what they know _about_ a customer. What they know about a customer **right now**. > **A rule I live by** > > Data you can’t reach in time is not an asset. It’s a filing cabinet. > > If acting on a customer signal requires a developer, a ticket, and a week, you do not have a data advantage. You have a report. ## Lesson Two: Your best people stopped maintaining data and started marketing This is the payoff nobody puts in the business case, and in my experience it’s the biggest one. Before the project, a meaningful share of my marketing team’s week went to moving data around. Reconciling lists. Exporting, cleaning, re-importing. Checking whether the thing that ran last night actually ran. Smart, expensive, creative people doing clerical work because the systems wouldn’t talk. Module 03 argued that middleware is where integration budgets either die or pay for themselves. Here’s what paying for themselves looks like: **the connections held, and the people who had been maintaining the data went back to marketing.** That’s not a soft benefit. Run it as arithmetic. Take the fraction of your marketing payroll currently spent on data janitorial work, multiply by the fully loaded cost of those people, and that number is a line item you recovered — every year, permanently, without hiring anybody. Then ask the second question, which is bigger: **what did those people produce once they had their time back?** In my case, over a hundred pieces of content. Which brings its own compounding return — more on that in a moment. > **A rule I live by** > > Count the capacity you free, not just the cost you cut. > > A re-platform that saves you money is a decent project. A re-platform that gives your best people their calendar back is a growth project. Most business cases only model the first one. ## Lesson Three: Relevance became a margin lever With an addressable customer record and people free to use it, the marketing changes shape. **You stop blasting.** The batch-and-blast email — same message, entire list, Tuesday morning — is replaced by flows that fire on behavior. In my implementation we built segments and a flow for each of the ones that mattered: - **New buyer** — the welcome and second-purchase sequence, where repeat-customer economics are actually won - **Abandoned cart** — the highest-intent moment in e-commerce, and the easiest money on this list - **Browse abandonment** — interest without a cart, which legacy platforms simply could not see - **Post-sale review request** — which feeds the social proof that Module 06 showed now feeds AI recommendation engines Four flows. Each one addressing a customer in a state the old platform couldn’t detect. **And the on-site experience stops being static.** We deployed personalized recommendations across the home page, collection pages, product pages and checkout — the four places a shopper actually decides. Not a hand-built “you may also like” grid that somebody updates quarterly, but placements that respond to what this visitor is doing in this session. I’ll be straight with you about the category, because this is where the hype lives. Recommendation engines are genuinely useful and they are not magic. They work by finding patterns in what people bought together and what this shopper has looked at. Several mature options exist and they are broadly comparable. **The engine is not the hard part. Having clean product data and real behavioral signal to feed it is the hard part** — which is Module 06’s argument, arriving from the other direction. What it was worth, in my project: **sales up more than 15%**, from a combination of new customer growth and higher average order value — **with significant growth in the years that followed.** For external corroboration, McKinsey’s personalization research found that doing this well typically produces a **10–15% revenue lift**, with company-specific results ranging from 5% to 25%, and that companies excelling at it draw 40% more revenue from those activities than average performers. That research also found 71% of consumers expect personalized interactions and 76% are frustrated when they don’t get them. Worth noting: **that study is from November 2021.** Those expectations have not softened in the five years since. ## Lesson Four: The money is in the second order, not the first Everything above is about the sale in front of you. The durable payoff is the next one. Once you have a clean transaction history and a live behavioral stream in one place, the system stops only reporting backward and starts forecasting. Four numbers come out of it, and they are the ones I’d put on a dashboard: 1. **Predicted customer lifetime value** — expected spend over the coming year, per customer 2. **Churn risk** — the probability this customer doesn’t come back 3. **Expected date of next order** — when they’re statistically most likely to reorder 4. **Average time between orders** — their individual purchase cadence Those four turn marketing from reactive to scheduled. You stop sending the win-back campaign after someone has already gone quiet for six months, and start sending it the week their own cadence says they’re due. You stop discounting your whole file and start spending retention budget on the customers whose churn risk says they need it. **That’s the difference between a mailing list and a customer base.** None of it was computable on the old stack. Not because the math is hard — it isn’t — but because the data to feed it lived in four systems that disagreed with each other. > **A word to the next generation** > > Learn to read a customer file the way a merchant reads a floor. > > Recency, frequency, value, cadence. Those four ideas are older than every platform you will ever work on, and every predictive model being sold today is a machine that computes them faster. Understand the ideas and the tools will never intimidate you. ## The Scorecard: What to measure, so the board meeting goes differently Before the next review, get these on one page: - **Revenue from flows versus broadcasts.** The ratio tells you how much of your marketing is behavioral and how much is still batch. - **New customers from the site**, trended against every other acquisition source you run. - **AOV, before and after**, and which placements moved it. - **Marketing hours recovered** from data maintenance, and what got produced with them. - **Organic and AI-search visibility**, trended against content published. That last one deserves a note. Those hundred-plus content pages my team produced with their recovered time did what good content always did — built site authority and organic traffic, the asset Module 02 was about protecting. But they’re now also doing something that didn’t exist when this series started: **feeding the answer engines from Module 06.** The same work, paying twice. > **Your next step** > > Write down three things your business can do this quarter that it physically could not do two years ago. > > If you can’t fill all three lines, you didn’t finish the project. You just changed platforms. ## If I Sat on Your Board: The question I’d ask Not what it cost. That’s settled, and Module 04 told you how to count it. > What can we do now that we could not do before — and what is it worth in dollars this year? Three possible answers. _The site is faster_ means you bought a redesign and called it a transformation. _We’re working on it_ means the capability is sitting there unused, which is the most expensive outcome of all. _Here are three things, here’s what each is worth, here’s the trend_ means you ran a systems project and collected on it. I’ll leave you with the one that stays with me. The business I did this for is **more than a hundred years old.** It was built on a printed catalog, a mailing list, and a telephone. Today its website is the single largest source of new customers it has. That is not a technology story. That’s a hundred-year-old company earning its next hundred years. **You didn’t suffer through the project for a prettier homepage. You suffered through it for the capability. Go use it.** _This closes the E-Commerce Implementation Series. Module 01 argued it was a systems project. Modules 02 through 06 covered the traffic, the integrations, the true cost, the platform fit, and the machine-readable catalog. If you’ve been following along and you’re planning a project of your own, start with a complimentary consultation — I’ll tell you the truth about what it will take._ > **About the author** > > **Gary Q. Smith** — President, GQS Direct, LLC > > Gary pairs four decades of P&L-accountable direct marketing leadership with practitioner-level command of today’s commerce stack. He is currently an AI evangelist and practitioner and has led four e-commerce website implementations and integrations, built marketing databases, was architect and project leader for a custom ERP system, and managed omnichannel marketing strategy and communications for several businesses. > > _Most consultants have studied these problems. GQS Direct has owned them._ > > [**Start with a complimentary consultation →**](https://gqsdirect.com/booking.html) **The E-Commerce Implementation Series** - [Module 01 — Re-Platforming Is a Systems Project, Not a Web Design Project](https://gqsdirect.com/blog/replatforming-is-a-systems-project) - [Module 02 — Your Redirect Map Is Worth More Than Your Theme](https://gqsdirect.com/blog/redirect-map-seo-preservation-migration) - [Module 03 — Middleware: The Key to Effective Integration and Project ROI](https://gqsdirect.com/blog/middleware-integration-project-roi) - [Module 04 — The Open-Source Paradox: Hidden Maintenance and DevOps Costs](https://gqsdirect.com/blog/open-source-paradox-three-year-tco) - [Module 05 — Your B2B Buyer Is Not a Consumer With a Bigger Cart](https://gqsdirect.com/blog/b2b-ecommerce-platform-decision) - [Module 06 — Your Next Customer Won’t Have Eyes](https://gqsdirect.com/blog/agentic-commerce-machine-readable-catalog) - Module 07 — The Re-Platforming Payoff _(you are here)_ --- This blog is powered by Superblog. Visit https://superblog.ai to know more. --- ## Your Next Customer Won’t Have Eyes Author: Gary Q. Smith Author URL: https://gqsdirect.com/blog/author/gary-q-smith Published: 2026-10-30 Category: E-Commerce Category URL: https://gqsdirect.com/blog/category/e-commerce Meta Title: Your Next Customer Won’t Have Eyes | GQS Direct Meta Description: Agentic commerce arrived as a discovery channel, not a checkout. Machine-readable product data is where the money is. What to fix first. URL: https://gqsdirect.com/blog/agentic-commerce-machine-readable-catalog Somebody asked an assistant for exactly what you sell this morning. Half-inch brass fittings, in stock, delivered by Friday. The assistant answered in about four seconds and named three companies. You weren’t one of them. Not because your price was wrong. Not because your site is ugly. Because when the machine went looking, your catalog didn’t carry the attribute it needed to confirm you qualified — and a machine that can’t verify doesn’t guess. **It doesn’t bounce. It just never arrives.** For thirty years we built storefronts for eyes. Banners, navigation, hero images, the fold. I’ve led four e-commerce implementations and I have argued about homepage layout in more conference rooms than I care to count. So take this from someone who has actually run the business: > “Your homepage is now a courtesy you extend to humans. Your product data is what the machines use to decide whether you exist at all.” Four lessons, one honest reckoning with the hype, and the question I’d ask if I sat on your board. ## Lesson One: The storefront didn’t disappear. The shelf did. Merchandising used to be your job. You decided what sat on the category page, what went above the fold, what got the endcap. You controlled the shelf because the shelf was on your property. When a buyer asks an assistant instead of browsing, that shelf gets assembled somewhere else, in about four seconds, out of structured data. Yours or somebody else’s. Module 02 made the case that schema stopped being a star-rating trick and became a machine-readable feed. That was true for search. It is now true for the thing that answers instead of searching — and the standard is higher, because an answer engine doesn’t rank you, it either _names_ you or it doesn’t. Forrester’s Emily Pfeiffer put the problem plainly when she said crawling and scraping is inadequate for commerce. Scraped pages carry stale stock, stale delivery promises, stale shipping terms. An agent that can’t trust the data moves to a merchant whose data it can. **That’s the whole game. Not aesthetics. Verifiability.** > **A rule I live by** > > A missing attribute is a lost sale you will never see in your analytics. > > There is no bounce, no abandoned cart, no exit page. The agent filtered you before you entered the conversation. Nothing in your reporting will ever tell you it happened. ## Lesson Two: The agents aren’t buying yet, and the numbers say so Now the part most articles on this subject won’t tell you, because it’s bad for the people selling agentic readiness. **In-chat checkout stalled.** OpenAI launched Instant Checkout on September 29, 2025, built on the Agentic Commerce Protocol with Stripe. Etsy sellers first, with over a million Shopify merchants promised to follow. By February 2026, Forrester counted roughly **thirty Shopify merchants actually live** on it. In March 2026, OpenAI pulled it back, moving checkout into Apps. Why? Walmart found that checkout completed inside ChatGPT converted at about **one-third the rate** of sending the same shopper to Walmart’s own site. Read that twice. The frictionless future converted three times worse than a link. And yet — this is the part that matters — **discovery is working.** Walmart found ChatGPT drove roughly **twice the new-customer rate** of search engines. A December 2025 Semrush survey of 1,030 US shoppers found **50% bought something after using AI to research it, while only 22% ever bought inside an AI tool.** Half your market is already being influenced by a machine. Barely a fifth is transacting with one. Meanwhile the protocol center of gravity moved. On January 11, 2026, Google and Shopify announced the **Universal Commerce Protocol** at NRF, with Etsy, Wayfair, Target and Walmart as co-developers and more than twenty endorsers including Visa, Mastercard, Stripe and American Express. By April, Amazon, Meta, Microsoft and Salesforce had joined its technical council. So the plumbing is being standardized by everyone at once, while the buying behavior hasn’t arrived yet. That is not a reason to ignore this. It is a reason to **spend on the half that’s paying today.** > **A rule I live by** > > Discover in AI, buy on your site. > > Invest in being findable and verifiable. Keep the transaction, the payment relationship and the customer record on your own property until the data tells you otherwise. Right now the data is not close. ## Lesson Three: Your platform already has the plumbing. It doesn’t have your data. Here’s the assumption I want to kill: that you need to go build an agent interface. If you’re on Shopify, **you already have one.** Shopify deploys a Storefront MCP server to every store, and agentic storefronts went live by default for eligible US merchants on March 24, 2026. Nobody opted in. It’s running. Which means the work in front of you is not protocol plumbing. It’s the same thing it has been since Module 03: **somebody has to make sure what flows through the pipe is correct.** Buy the dispatcher. Don’t hand-build the spaghetti. Where the platforms stand, as I read it: **Agentic readiness by platform** Platform My assessment Where it stands Shopify **Leader** UCP co-author; Storefront MCP on every store; agentic storefronts on by default for eligible US merchants commercetools / Saleor **Leader** API-first GraphQL; clean structured product models; custom agent interfaces are straightforward BigCommerce **Mid-tier** Strong native metadata feeds; programmatic checkout through partners Salesforce Commerce **Mid-tier** Joined the UCP technical council in April 2026; custom API work still required Adobe Commerce — ACCS **Mid-tier** API-first and out-of-process by design — the shape agents want Adobe Commerce — PaaS **Laggard** Heavy custom development for agent endpoints; manual schema refactoring WooCommerce **Laggard** No native agent endpoints; plugins and custom code _My assessment, not a published index. Verify against your vendor’s current roadmap before you decide anything._ Note the Adobe split, because it inverts Module 05. On the SaaS edition you lose PHP and the extension library — and that same out-of-process, API-first architecture is exactly what an agent wants to talk to. The thing that costs you flexibility buys you readiness. **Nothing is free. Everything trades.** ## Lesson Four: Six things to fix, in the order that pays Resequenced by what returns money this year, not by what’s most exciting. 1. **Validate your schema.** Clean JSON-LD on every product template — Product and Offer on all of them. **AggregateRating only where genuine reviews exist**, because marking up ratings you don’t have is a policy violation, not a growth tactic. 2. **Complete the catalog.** Material, size, color, dimensions, compatibility, finish. Answer engines filter on attributes. A blank field isn’t a gap in your listing — it’s a disqualification you never see. 3. **Sync price and stock in real time.** If indexed data and live checkout disagree, the agent abandons and learns not to trust you. Module 03 was about exactly this failure, and it is still the one that quietly costs the most. 4. **Build off-site trust.** Language models cross-reference third-party platforms to validate quality. Your Trustpilot profile and your review volume now feed a recommendation engine, not just a badge. 5. **Have the vendor conversation.** Ask your platform three questions: what’s on by default today, what’s on the roadmap, and what does your UCP position look like. Make them answer in writing. 6. **Agentic checkout last.** When the conversion data turns, you’ll want to be ready. Today it hasn’t. The first three are blocking and tackling. They are also where nearly all of the available money is. > **A word to the next generation** > > Learn product data. Attributes, identifiers, taxonomy, feed hygiene — the unglamorous middle of this business. > > Every generation gets a new interface. This one happens to be a machine that reads instead of looks. The people who understood the catalog underneath the interface have been winning for forty years, and that isn’t going to stop now. > **Your next step** > > Pick your ten best-selling SKUs. Open each one’s structured data and count the empty attributes. > > That number is your agentic readiness score, and it costs you nothing but an afternoon to find out. ## If I Sat on Your Board: The question I’d ask Not which protocol. Not what we’re building. > When a buyer asked an assistant about our category last month, how often were we in the answer — and who here is responsible for checking? Three possible answers. _We’ve never looked_ means you are flying blind in a channel that already influences half your market. _We looked once_ means somebody was curious but nobody owns it. _Here’s the number, here’s the trend, and here’s the SKUs we fixed_ means you have a merchandiser for the machine age. The agents aren’t coming for your storefront. **They’re coming for your data. Give them something worth finding.** _Next in this series —_ [**_Module 07: The Re-Platforming Payoff_**](https://gqsdirect.com/blog/replatforming-payoff) _The market share and the margin you suffered through the whole project to go get._ > **About the author** > > **Gary Q. Smith** — President, GQS Direct, LLC > > Gary pairs four decades of P&L-accountable direct marketing leadership with practitioner-level command of today’s commerce stack. He is currently an AI evangelist and practitioner and has led four e-commerce website implementations and integrations, built marketing databases, was architect and project leader for a custom ERP system, and managed omnichannel marketing strategy and communications for several businesses. > > _Most consultants have studied these problems. GQS Direct has owned them._ > > [**Start with a complimentary consultation →**](https://gqsdirect.com/booking.html) **The E-Commerce Implementation Series** - [Module 01 — Re-Platforming Is a Systems Project, Not a Web Design Project](https://gqsdirect.com/blog/replatforming-is-a-systems-project) - [Module 02 — Your Redirect Map Is Worth More Than Your Theme](https://gqsdirect.com/blog/redirect-map-seo-preservation-migration) - [Module 03 — Middleware: The Key to Effective Integration and Project ROI](https://gqsdirect.com/blog/middleware-integration-project-roi) - [Module 04 — The Open-Source Paradox: Hidden Maintenance and DevOps Costs](https://gqsdirect.com/blog/open-source-paradox-three-year-tco) - [Module 05 — Your B2B Buyer Is Not a Consumer With a Bigger Cart](https://gqsdirect.com/blog/b2b-ecommerce-platform-decision) - Module 06 — Your Next Customer Won’t Have Eyes _(you are here)_ - [Module 07 — The Re-Platforming Payoff](https://gqsdirect.com/blog/replatforming-payoff) --- This blog is powered by Superblog. Visit https://superblog.ai to know more. --- ## The Open-Source Paradox: Hidden Maintenance and DevOps Costs Author: Gary Q. Smith Author URL: https://gqsdirect.com/blog/author/gary-q-smith Published: 2026-10-30 Category: E-Commerce Category URL: https://gqsdirect.com/blog/category/e-commerce Meta Title: The Open-Source Paradox: Hidden Costs | GQS Direct Meta Description: The license is free; the software never is. Hosting, DevOps, security and upgrades put self-hosting at three to ten times the cost of SaaS. URL: https://gqsdirect.com/blog/open-source-paradox-three-year-tco **The license was free.** That was the whole argument, delivered across a conference table by a developer who meant every word of it. No subscription. No vendor. Download it, own it, build whatever you want. Three years later the same company was paying two engineers to keep the thing breathing, a hosting bill that had quietly tripled, and a quote for a version upgrade that cost more than the original build. Nobody lied to anybody. **The license really was free. The software never was.** In Module 01 I laid out the full migration cost stack, and line 05 was backend integration — which became Module 03. This module is the line nobody puts on the stack at all, because it doesn’t arrive as an invoice. It arrives as payroll, retainers, and nights your team spends on infrastructure instead of on your customers. I’ve led four e-commerce implementations, served as architect and project leader on a custom ERP replacement, and watched this exact arithmetic play out on both sides of it. So take this from someone who has actually run the business: > “You never buy software. You buy the obligation to keep it running. The license is just the cover charge.” Four lessons, one worksheet you can fill in tonight, and the question I’d ask if I sat on your board. ## Lesson One: The license was never the cost Here’s the arithmetic nobody puts in the deck. Shopify Plus runs **$2,300 a month on a three-year term**, $2,500 on a one-year — call it **$27,600 a year**. Hosting, CDN, patching, scaling, uptime and PCI certification of the platform are inside that number. Self-hosted, the license line goes to zero and five other lines wake up. **Three-year total cost of ownership** Cost component Managed SaaS Self-hosted (low) Self-hosted (high) Platform license $27,600 $0 $0 Hosting & CDN included $12,000 $36,000 DevOps & infrastructure included $40,000 $150,000 Security & PCI auditing included $5,000 $15,000 Version upgrades (annualized) included $25,000 $66,667 **Three-year total** **$82,800** **$246,000** **$803,000** Three to nearly ten times the cost of the thing that had a price tag on it. And notice where the money actually goes. It isn’t servers — servers are cheap now. It’s **people**. A specialized DevOps retainer, or the engineers you hire once you pass roughly $2 million and discover that “our developer handles it” is no longer a sentence anyone can say out loud. Those are my ranges, from my experience. Yours will differ. That’s the point — put your own numbers in and run it. > **A rule I live by** > > A cost that arrives as payroll is still a cost. > > If the only way to see your platform bill is to look at headcount, you don’t have a cheaper platform. You have a hidden one. ## Lesson Two: The upgrade treadmill you never budgeted for Here’s what the free license buys you that nobody mentions: a deadline you don’t control. Adobe Commerce supports each 2.4.x release for three years. Version 2.4.7 goes out of support on **May 31, 2027**. After that, security patches stop. You are not choosing whether to upgrade. You are choosing between an upgrade project and an unpatched storefront taking credit cards. **An upgrade is not a patch. It’s a project.** Every extension revalidated, every customization re-tested, every integration re-pointed. Thirty to eighty thousand dollars is a fair range for a store carrying real customization, and the three-year support window puts two or three of them in your window. POLYWOOD — the outdoor furniture manufacturer, 150,000 product variations — ran a heavily customized Adobe Commerce build and measured exactly this. Their Chief Digital Officer, Benjamin Spiegel, put the number on the record: **more than 80% of their development resources went to platform maintenance rather than creating value.** One full fifth of every year was spent on codebase version upgrades alone. Twenty percent of the year. Maintaining the thing that was supposed to be free. After moving to a managed platform, they redirected 100% of those development resources, met roughly three-quarters of their custom requirements with off-the-shelf apps, and started shipping features in hours and days instead of months. Conversion rose 22%. Average order value rose 12%. Operating cost fell by a six-figure amount. They braced for the three-to-six-month organic traffic drawdown everyone warns about — and it never came. That last part matters, and it’s the whole argument of [Module 02: the traffic loss is not a law of nature](https://gqsdirect.com/blog/redirect-map-seo-preservation-migration) It’s a consequence of a bad redirect map. ## Lesson Three: Control is real. Ask what you’re paying for it. I’m not here to tell you open source is a mistake. It isn’t. Self-hosting buys you three things that are genuinely worth money. You own the code, so no vendor can retire a feature you depend on. You can build a workflow that exists nowhere else — and if you manufacture configured product, that may not be optional. And nobody raises your rent. Those are real. Pay for them with open eyes. What I will not let you pretend is that the comparison is free money. Two honest notes, because both columns get fudged: **Card processing fees belong in neither column.** You pay those on any platform. Anyone who loads them onto one side of the comparison is selling, not counting. **PCI isn’t a problem that vanishes.** A managed platform carries the certification for the platform itself, which is most of the burden and all of the hardest part. It does not carry your apps, your checkout extensions, your call-center order entry, or your staff. Your scope shrinks. It doesn’t disappear. > **A rule I live by** > > Decide control versus cost on your product, not on your pride. > > Configured product, custom manufacturing, rules that exist nowhere else — build. A stock catalog on a standard order flow — buy. The second answer is not the lesser one. It’s the one that leaves money for marketing. ## Lesson Four: Run the three-year number before the demo Most platform decisions are made on a year-one comparison, because year one is the only year anybody models. Year one is also the only year self-hosting looks cheap, because the upgrades haven’t come due and the DevOps retainer hasn’t been signed. So model three years. Every time. It takes an afternoon. 1. **Get your real hosting bill** — not the quote, the bill, with bandwidth and scaling events in it. 2. **Price the DevOps capacity honestly.** Retainer, or salary plus benefits plus the hiring risk. 3. **Count the upgrade projects** inside the support windows of the version you’re actually running. 4. **Add security and compliance** — penetration testing, audits, remediation. 5. **Then add the line nobody adds: opportunity cost.** Every engineer-week on infrastructure is an engineer-week not spent on merchandising, site speed, or the checkout. That fifth line is the one that got POLYWOOD. Not the hosting bill. The eighty percent. > **A word to the next generation** > > Learn to read a total cost of ownership model, and learn to build one. It is the single most transferable skill in this business. > > Platforms will change three times in your career. The question “what does this actually cost us over three years” never will — and the person in the room who can answer it is the person who gets listened to. > **Your next step** > > The worksheet accompanying this module is prefilled with the ranges above and does the math itself. Overwrite the blue cells with your own quotes, retainers and hosting bills, and look at the bottom line. > > Then send it to whoever told you the license was free. Not to win the argument. To have the real one. ## If I Sat on Your Board: The question I’d ask Not which platform. Not what the license costs. > What percentage of your engineering hours last quarter went to keeping the platform alive rather than making you money — and who is tracking that number? Three possible answers, and they tell you everything. _We don’t measure that_ means you are paying the bill without reading it. _About half, we think_ means you already know and haven’t said it out loud. _Eleven percent, here’s the dashboard_ means you own your platform instead of the other way around. The license is the smallest number in this decision. **Free software. Expensive obligations. Count them both.** _Next in this series —_ [**_Module 05: Your B2B Buyer Is Not a Consumer With a Bigger Cart_**](https://gqsdirect.com/blog/b2b-ecommerce-platform-decision) _Shopify, Adobe Commerce and OroCommerce — and how to tell which one your business actually needs._ > **About the author** > > **Gary Q. Smith** — President, GQS Direct, LLC > > Gary pairs four decades of P&L-accountable direct marketing leadership with practitioner-level command of today’s commerce stack. He is currently an AI evangelist and practitioner and has led four e-commerce website implementations and integrations, built marketing databases, was architect and project leader for a custom ERP system, and managed omnichannel marketing strategy and communications for several businesses. > > _Most consultants have studied these problems. GQS Direct has owned them._ > > [**Start with a complimentary consultation →**](https://gqsdirect.com/booking.html) **The E-Commerce Implementation Series** - [Module 01 — Re-Platforming Is a Systems Project, Not a Web Design Project](https://gqsdirect.com/blog/replatforming-is-a-systems-project) - [Module 02 — Your Redirect Map Is Worth More Than Your Theme](https://gqsdirect.com/blog/redirect-map-seo-preservation-migration) - [Module 03 — Middleware: The Key to Effective Integration and Project ROI](https://gqsdirect.com/blog/middleware-integration-project-roi) - Module 04 — The Open-Source Paradox: Hidden Maintenance and DevOps Costs _(you are here)_ - [Module 05 — Your B2B Buyer Is Not a Consumer With a Bigger Cart](https://gqsdirect.com/blog/b2b-ecommerce-platform-decision) - [Module 06 — Your Next Customer Won’t Have Eyes](https://gqsdirect.com/blog/agentic-commerce-machine-readable-catalog) - [Module 07 — The Re-Platforming Payoff](https://gqsdirect.com/blog/replatforming-payoff) --- This blog is powered by Superblog. Visit https://superblog.ai to know more. --- ## Middleware — The Key to Effective Integration and Project ROI Author: Gary Q. Smith Author URL: https://gqsdirect.com/blog/author/gary-q-smith Published: 2026-10-30 Category: E-Commerce Category URL: https://gqsdirect.com/blog/category/e-commerce Meta Title: Middleware: The Key to Integration ROI | GQS Direct Meta Description: Point-to-point integrations break silently and cost you orders. How to choose middleware for your ERP and storefront, with an RFP scorecard. URL: https://gqsdirect.com/blog/middleware-integration-project-roi **Nobody calls the integration layer a project.** It’s a line item. “Connect the store to the ERP.” A developer, a couple of weeks, done. Then it’s a Monday morning and the warehouse calls to ask why there’s nothing to pick. Orders stopped posting Friday night. Nobody knew, because nothing threw an error — the script kept running, the site kept taking orders, and the ERP kept waiting for data that never arrived. The storefront was fine. The ERP was fine. **The thing between them was not.** In Module 01, line 05 of the migration cost stack was “backend systems integration — secure pipelines between checkout and your ERP, WMS, and POS.” This module is that line. I’ve served as architect and project leader on a custom ERP replacement and led four e-commerce implementations, so take this from someone who has actually run the business: > “Every system in your company can be healthy while your business is broken. That’s what the integration layer does when nobody owns it.” Four lessons, a scorecard you can send to vendors, and the question I’d ask if I sat on your board. ## Lesson One: Spaghetti is cheap to write and expensive to own Past roughly $2 million in revenue, the stack multiplies. Online store. Retail point of sale. ERP. Accounting. Warehouse management. CRM. Each one arrived to solve a real problem, and each one speaks its own dialect. The instinct is to wire them together directly — a script here, a webhook there. Point-to-point is genuinely cheap to write. That’s the trap. Six systems wired directly to each other is a bowl of spaghetti where every strand is load-bearing and none of them are documented. And here’s the part nobody warns you about: that spaghetti has an expiration date you didn’t set. Shopify releases a new API version **every three months**, named by date. Each stable version is supported for **at least twelve months**. When your script keeps asking for a version that has aged out, Shopify doesn’t slam the door — it **falls forward** and answers with the oldest version it still supports. Read that again, because it’s the whole problem. Your integration doesn’t crash. It gets a perfectly good response, in a shape it wasn’t written for, and keeps going. The failure is silent, and you find it downstream — in a warehouse that has nothing to pick, or an inventory count that’s quietly wrong. Rate limits work the same way. Shopify uses a leaky bucket: fill it and you get throttle errors until capacity drains back. Their own guidance is to back off, retry deliberately, cache what you reuse, queue and spread the work. All reasonable. All of it code somebody has to write — and point-to-point scripts almost never do, because they were a two-week line item. Middleware is what that code becomes when you build it once instead of eleven times. One dispatcher that every system talks to: it cleanses payloads, maps fields, absorbs rate limits, retries what’s retryable, holds what isn’t, and leaves an audit trail you can replay when something goes wrong. > **A rule I live by** > > Every connection needs a name on it. > > If you can’t tell me who gets the alert when an order fails to post, you don’t have an integration. You have a script with good intentions. ## Lesson Two: Six roads, one choice There’s no universally correct answer here. The right one falls out of two things: how unusual your business is, and how complex your workflows are. Run a custom product line, a heavily customized platform, and a custom-built ERP, and you will end up building. Run a stock catalog as a retailer or distributor, and one of the packaged options will serve you better and cost less. **Six middleware alternatives** Option Who it fits What it costs you **iPaaS** Mid-market brands with mostly standard processes wanting prebuilt connectors, monitoring, reusable workflows A subscription forever, and friction when an integration gets genuinely unusual **Prebuilt ERP-to-commerce connector** Common ERP plus common platform, standard catalog Flexibility — unique processes won’t fit, and you’ll wait on the vendor **Custom API integration** Custom ERP, configured products, rules that exist nowhere else Build cost up front, and you own maintenance, monitoring and version upgrades forever **Enterprise service bus** Large organizations routing many systems with complex orchestration More cost and complexity than most mid-market businesses will ever need **Direct API / webhook** A small number of flows that need near-real-time updates Tight coupling, and you hand-build every retry, error path and security control **File-based / EDI** Legacy ERP, scheduled bulk transfers, trading-partner requirements Latency, plus failed transfers, duplicates and manual reconciliation Pick from your business model and your workflows — not from a vendor’s feature grid. ## Lesson Three: Choose on five things, not fourteen I’ve watched teams build forty-row comparison matrices. They always produce the same result: a tie, and a decision made on vibes two weeks later. Five things decide this. **1\. Compatibility.** Your specific ERP, your specific version, your specific platform. Confirm the connector performs the _function_ you need. “Supports NetSuite” is marketing, not a requirement met. **2\. Business fit.** Your real order flow, not a generic one. Product configurations, custom text, artwork files, production instructions. If you’re a manufacturer, this is where standard connectors usually run out of road. **3\. Reliability.** Can it detect a failed transaction, retry what’s temporary, prevent a duplicate order, and tell a human when it needs help? Lost orders and wrong inventory are what integration failures actually cost. **4\. Total cost and ROI.** License, transaction fees, implementation, maintenance, future enhancements. Model it at two or three times your current volume, because that’s the point of the project. **5\. Future flexibility.** The next system is always coming — accounting, CRM, shipping, WMS, manufacturing. A platform that carries several applications is worth more over five years than one that connects exactly two. > **A rule I live by** > > Decide real-time versus batch per data type, not per project. > > Orders and fulfillment want real time. Catalog updates and bulk pricing rarely do. Timely batch — small, frequent, scheduled — is the most underrated option in this whole business, because it fails politely. > > Real-time is terrific. Ask what happens to your website, your middleware and your ERP when it fails. If the answer is a shrug, choose batch. ## Lesson Four: Run it like a project, not a purchase Middleware selection deserves the same rigor you gave choosing the e-commerce platform. It usually gets a fraction of it. **Assign a project leader.** If the technical expertise isn’t in the building, get it or contract for it — but the project management accountability stays in-house, with someone respected enough to make decisions and hold a timeline. **Keep the team small.** A handful of subject matter experts who know how the work actually happens. In my experience the best-performing teams on this are small and agile. **Then work the milestones:** 1. **Document today’s reality.** Every business rule and process that moves data between the site and the ERP — including the manual ones. _Especially_ the manual ones. They’re where the undocumented business logic lives. 2. **Write the requirements.** Most of them fall out of step one, along with the improvements you’ve wanted for years. 3. **Issue an RFP** to a short list. Short. Four serious vendors beat twelve polite ones. 4. **Evaluate the responses** against the scorecard, and decide which of the six roads you’re on and who your partner is. 5. **Present the decision** with cost-benefit, implementation strategy, timeline, ROI and the business risks named out loud. > **A word to the next generation** > > Learn data mapping. Product identifiers, price structures, customer records, order formats — how a thing in one system becomes the same thing in another. > > The platforms will change three times in your career. The mapping problem never does. ## If I Sat on Your Board: The question I’d ask Not which vendor. Not what it costs. > When this breaks at 2 a.m., who gets the alert — and what happens to the order that was in flight? Three possible answers, and they tell you everything. _Nobody knows_ means you’re running on hope. _An email alias_ means the alert exists and nobody reads it. _It queues, it retries, and here’s the person who owns it_ means you bought an integration instead of a script. > **Your next step** > > Turn the considerations below into a weighted scorecard, send it with your RFP, and score every vendor against your requirements rather than their demo. > > A vendor who answers all fourteen cleanly has done this before. One who dodges six of them has told you something important. **The integration RFP scorecard** Consideration What to verify **1\. ERP & platform compatibility** Your exact ERP, version and platform — by function, not by logo **2\. Business requirements** Which data flows: products, inventory, pricing, customers, orders, fulfillment **3\. Customization & manufacturing** Configurations, custom text, artwork files, production instructions **4\. Real-time vs. batch** What truly needs immediacy, and where timely batch is safer **5\. Implementation time & complexity** Time, people and expertise to get live **6\. Total cost of ownership** License, transaction fees, implementation, maintenance, enhancements — at future volume **7\. Reliability & error handling** Failure detection, retries, duplicate prevention, human alerts **8\. Monitoring & troubleshooting** Dashboards and transaction logs a non-developer can actually use **9\. Scalability & flexibility** More volume, more sites, more systems without a redesign **10\. Data mapping & transformation** Identifiers, price structures, customer records, order formats **11\. Security & data protection** Authentication, access control, encryption, compliance **12\. Vendor support & viability** Support, documentation, roadmap — and how hard it is to leave **13\. Internal technical resources** An honest match to what your team can carry **14\. Future integrations** Accounting, CRM, shipping, WMS, manufacturing Your storefront is what your customers see. Your middleware is what your business runs on. **Buy the dispatcher. Not the spaghetti.** _Next in this series —_ [**_Module 04: The Open-Source Paradox — Hidden Maintenance and DevOps Costs_**](https://gqsdirect.com/blog/open-source-paradox-three-year-tco) _Exposing the true cost of self-hosting and calculating your three-year TCO._ > **About the author** > > **Gary Q. Smith** — President, GQS Direct, LLC > > Gary pairs four decades of P&L-accountable direct marketing leadership with practitioner-level command of today’s commerce stack. He is currently an AI evangelist and practitioner and has led four e-commerce website implementations and integrations, built marketing databases, was architect and project leader for a custom ERP system, and managed omnichannel marketing strategy and communications for several businesses. > > _Most consultants have studied these problems. GQS Direct has owned them._ > > [**Start with a complimentary consultation →**](https://gqsdirect.com/booking.html) **The E-Commerce Implementation Series** - [Module 01 — Re-Platforming Is a Systems Project, Not a Web Design Project](https://gqsdirect.com/blog/replatforming-is-a-systems-project) - [Module 02 — Your Redirect Map Is Worth More Than Your Theme](https://gqsdirect.com/blog/redirect-map-seo-preservation-migration) - Module 03 — Middleware: The Key to Effective Integration and Project ROI _(you are here)_ - [Module 04 — The Open-Source Paradox: Hidden Maintenance and DevOps Costs](https://gqsdirect.com/blog/open-source-paradox-three-year-tco) - [Module 05 — Your B2B Buyer Is Not a Consumer With a Bigger Cart](https://gqsdirect.com/blog/b2b-ecommerce-platform-decision) - [Module 06 — Your Next Customer Won’t Have Eyes](https://gqsdirect.com/blog/agentic-commerce-machine-readable-catalog) - [Module 07 — The Re-Platforming Payoff](https://gqsdirect.com/blog/replatforming-payoff) --- This blog is powered by Superblog. Visit https://superblog.ai to know more. --- ## Your Redirect Map Is Worth More Than Your Theme Author: Gary Q. Smith Author URL: https://gqsdirect.com/blog/author/gary-q-smith Published: 2026-10-30 Category: E-Commerce Category URL: https://gqsdirect.com/blog/category/e-commerce Meta Title: Your Redirect Map Is Worth More Than Your Theme | GQS Direct Meta Description: A re-platform can give back the organic traffic you spent a decade earning. How to price the risk, build the redirect map, and protect schema for AI search. URL: https://gqsdirect.com/blog/redirect-map-seo-preservation-migration **The call always comes on a Tuesday.** Launch was Saturday. The site looks terrific. Everyone slept. Then somebody opens Search Console, and the organic line has gone off a table edge. Nobody in the room wants to say the word out loud. Redirects. In Module 01 I called the storefront the tip of the iceberg — 20% to 30% of a re-platform, with the rest below the waterline. This module is about the one line item down there that can cost you the most, because it’s the only one your competitors get to keep if you drop it. I’ve run Shopify migrations, and I spend my days now on search and AI-search visibility: redirect maps, schema, citations. So take this from someone who has actually run the business: > “Crawlers don’t care about your new theme. They care about whether the URL that ranked last week still leads to the same thing.” Four lessons. Then the question I’d ask if I sat on your board. ## Lesson One: Put a number on the risk before you launch The horror stories are real, but they get told badly. “Migrations tank your traffic” isn’t a fact. It’s a description of what happens when nobody owns the URLs. Price it instead. Shopify’s own risk model is honest arithmetic: > **Revenue at risk** > > **Share of revenue from organic × share of that organic that is non-branded × 15% to 30%.** > > The 15–30% is the worst case, not the forecast. Branded searches find you anyway — people typing your name land on your homepage no matter what you did to /shop/category/. It’s the non-branded traffic, the people searching for what you sell rather than who you are, that lives or dies on your redirect map. Run it on a merchant doing 20% of revenue through organic, 70% of it non-branded: **2.1% to 4.2% of revenue at risk**, recoverable in one to three months. Shopify’s own guidance is that most migrations see declines of 10% or less. That’s a number you can take into a budget meeting. “We might lose some traffic” is not. > **A rule I live by** > > Calculate the revenue at risk **before** the kickoff meeting, not after the launch. > > A board that has seen the number funds the redirect work. A board that hasn’t will cut it to pay for photography. ## Lesson Two: The redirect map is the deliverable Google’s site-move guidance says it plainly: _“Don’t redirect many old URLs to one irrelevant single URL destination, such as the home page of the new site.”_ That’s the single most common failure I see. A wildcard rule sweeps a decade of product and blog URLs onto the homepage, someone calls it “handled,” and Google reads every one of those as a soft 404. The ranking doesn’t transfer. It evaporates. The map takes five steps: 1. **Inventory every URL.** Search Console, Semrush or Ahrefs, your XML sitemap, and the legacy database. Historical URLs count — old pages still carry links. 2. **Clean it up.** Kill duplicate paths, trailing-slash mismatches, and parameterized tracking junk before you map anything. 3. **Map 1:1 with permanent redirects.** Every legacy URL to its real equivalent, 301 or 308, straight to the final destination. No chains. 4. **Check canonical parity.** Every new URL gets a self-referencing canonical that matches the redirect target. Mismatches create indexing loops. 5. **Crawl staging against the old list.** Screaming Frog against your legacy URL export, before go-live, finds the unmapped pages while fixing them is still free. On Shopify, the structure itself guarantees work. Products live under /products/, categories under /collections/, posts under /blogs/blog-name/article-title. Almost nothing on your old platform matches that, so nearly every URL needs a row in the map. Standard plans hold up to 100,000 redirects and Plus up to 20 million — if you’re over the line, prioritize by organic revenue, not by alphabetical order. Then leave them alone. Google’s advice is to keep redirects live **at least a year.** I’d say leave them forever; they cost nothing. > **A rule I live by** > > If a URL isn’t on the map, it isn’t launched. > > I’ve never regretted holding a go-live for a redirect file. I have regretted the opposite. ## Lesson Three: Structured data stopped being a rich-results trick For fifteen years, schema markup was something you did for a star rating in the search results. **That era is over, and most SEO advice hasn’t caught up.** Your structured data is now a machine-readable feed — read by Google for merchant listings, and read by ChatGPT, Gemini, Perplexity and AI Overviews when they decide which products to name in an answer. These engines pull facts from JSON-LD rather than guessing from your HTML. If your schema breaks during migration, you don’t drop a few positions. You stop being quotable. **What each markup earns you** Markup What it earns you **Product / ProductGroup** The core entity. Variants grouped so engines know which SKUs belong together. **Offer** Live price, currency, and availability. Automate it — stale data is worse than no data. **Return policy** (under Organization) Return window and fees. Engines weigh risk when recommending a purchase. **AggregateRating / Review** Proof from customers rather than from your copy. **Organization + sameAs** Your brand as an entity, linked to the profiles that confirm it. **BreadcrumbList** Your category hierarchy, so engines understand what sits where. **VideoObject** Assembly, sizing, and unboxing video surfaced as media rather than buried. **FAQPage** **No longer a Google rich result** — retired in May 2026. Keep it only if it helps AI engines read your answers. That last row is worth your attention. Plenty of migration checklists still promise FAQ rich results. Google turned them off. Budget accordingly. One honest note on timing. Google says new URLs replace old ones in its index over a few weeks or more, depending on your size. The AI engines publish no schedule at all. Anyone quoting you a precise number of days for ChatGPT to re-cite your catalog is guessing. Plan for weeks, and verify the way I do: ask the engines the questions your buyers ask, and watch for your name to come back. ## Lesson Four: The first 90 days decide the recovery Cutover isn’t the finish line. It’s the first day of the only period that determines whether this project pays for itself. **The post-launch watch** Window What you watch What you fix **Days 1–7** · War room Checkout and payment gateways, 404 logs, sitemap indexing — hourly. Every redirect miss, within 24 hours. No exceptions. **Days 8–30** · Edge cases Conversion rate, mobile cart abandonment, support tickets, session recordings and heat maps. Interface friction and page speed — the problems real traffic finds. **Days 31–90** · Recovery Rankings, impressions, AI citations, non-branded sessions. Hold the line. Measure, don’t thrash. Expect a dip in weeks two through six while engines recrawl and re-index. It isn’t a failure; it’s the cost of moving. Most disciplined migrations stabilize within one to three months, large catalogs can take up to six. And during the recovery window, resist the urge to “fix” things by rewriting content. Change five variables at once and you lose the ability to tell which one mattered. Discipline now is what buys you an honest answer later. ## If I Sat on Your Board: The question I’d ask Not “when do we launch.” Not “do we like the design.” > Show me the redirect map. Then tell me what traffic dip we’ve budgeted for, and when we expect to be whole. If the answer is a spreadsheet and a number, you have a plan. If it’s a reassurance, you have a theme and a hope. > **Your next step** > > This week, before anyone talks design: **pull the URL inventory.** Search Console, your SEO platform, the XML sitemap, the legacy database. One list, deduplicated, ranked by organic revenue. > > That list is the spine of the whole migration. Everything else in Module 01’s cost stack bolts onto it. Your theme has a shelf life — you’ll replace it in three years and barely remember this one. Your URLs don’t. **Protect the map. The rankings follow.** _Next in this series —_ [**_Module 03: Middleware — The Key to Effective Integration and Project ROI_**](https://gqsdirect.com/blog/middleware-integration-project-roi) _The layer between your storefront and your ERP is where integration budgets go to die, or where they pay for themselves._ > **About the author** > > **Gary Q. Smith** — President, GQS Direct, LLC > > Gary pairs four decades of P&L-accountable direct marketing leadership with practitioner-level command of today’s commerce stack. He is currently an AI evangelist and practitioner and has led four e-commerce website implementations and integrations, built marketing databases, was architect and project leader for a custom ERP system, and managed omnichannel marketing strategy and communications for several businesses. > > _Most consultants have studied these problems. GQS Direct has owned them._ > > [**Start with a complimentary consultation →**](https://gqsdirect.com/booking.html) **The E-Commerce Implementation Series** - [Module 01 — Re-Platforming Is a Systems Project, Not a Web Design Project](https://gqsdirect.com/blog/replatforming-is-a-systems-project) - Module 02 — Your Redirect Map Is Worth More Than Your Theme _(you are here)_ - [Module 03 — Middleware: The Key to Effective Integration and Project ROI](https://gqsdirect.com/blog/middleware-integration-project-roi) - [Module 04 — The Open-Source Paradox: Hidden Maintenance and DevOps Costs](https://gqsdirect.com/blog/open-source-paradox-three-year-tco) - [Module 05 — Your B2B Buyer Is Not a Consumer With a Bigger Cart](https://gqsdirect.com/blog/b2b-ecommerce-platform-decision) - [Module 06 — Your Next Customer Won’t Have Eyes](https://gqsdirect.com/blog/agentic-commerce-machine-readable-catalog) - [Module 07 — The Re-Platforming Payoff](https://gqsdirect.com/blog/replatforming-payoff) --- This blog is powered by Superblog. Visit https://superblog.ai to know more. --- ## Re-Platforming Is a Systems Project, Not a Web Design Project Author: Gary Q. Smith Author URL: https://gqsdirect.com/blog/author/gary-q-smith Published: 2026-10-30 Category: E-Commerce Category URL: https://gqsdirect.com/blog/category/e-commerce Meta Title: Re-Platforming Is a Systems Project | GQS Direct Meta Description: The storefront is 30% of a re-platform. Data migration, integrations and SEO preservation are what sink launches — and what to budget for. URL: https://gqsdirect.com/blog/replatforming-is-a-systems-project **I’ve sat in that kickoff meeting.** Twelve people around a table. Two hours on the calendar. And ninety minutes of it spent on the homepage hero image. Nobody asked which system owns the customer record. Nobody asked what happens to the order when it leaves checkout. Nobody asked who fixes it at 11 p.m. on a Tuesday when inventory and the warehouse disagree. We talked about the picture. Here’s what the picture is worth. On a mid-market re-platform, the customer-facing storefront is **20% to 30%** of the real effort. The rest — **70% to 80%** — goes into how your store talks to your ERP, your product data, your warehouse, your point of sale, and your search equity. **Re-platforming isn’t a redesign. It’s a systems project wearing a nice theme.** I’ve led four e-commerce implementations and served as architect and project leader on a custom ERP replacement. I’ve run the Shopify side of this — the migrations, the integrations, the launch weekends. So take this from someone who has actually run the business: > “If you can’t name the system of record for every field on the order, you are not ready to pick a theme.” The stakes keep rising. US e-commerce passed **$1.23 trillion** in 2025 — about **16.4%** of all retail sales, and still growing faster than stores. More of your revenue runs through this plumbing every year. And PwC’s research found that **32% of customers will walk away from a brand they love after a single bad experience.** One bad launch week can cost you buyers you spent a decade earning. Four lessons from the trenches. They won’t make the project easy. They’ll make it survivable. ## Lesson One: The storefront is the tip of the iceberg Design is visible, so design gets the meeting. Integration is invisible, so integration gets discovered — usually in week nine, usually by your controller. Price the whole stack before you approve anything. Nine phases, and only two of them are about how the site looks: **The migration cost stack** Phase What it actually covers **01 · Discovery & scope mapping** System inventory, data flow mapping, and deciding who owns which field. **02 · Data migration & cleaning** Extracting, profiling, and reformatting product, customer, and order records. **03 · UX & front-end design** Layout, interaction, theme sections. The part everyone pictures. **04 · Storefront development** Turning those designs into responsive, native templates. **05 · Backend systems integration** Secure pipelines between checkout and your ERP, WMS, and POS. **06 · SEO migration & redirects** 1:1 redirect maps and preserved structured data — your rankings live or die here. **07 · QA & scenario testing** End-to-end dry runs: real orders, refunds, exchanges, taxes, load. **08 · Team training** Support, finance, and operations learning the new workflow before launch, not after. **09 · Post-launch stabilization** Hour-by-hour watch on checkout, payments, and error logs. Skullcandy’s move off BigCommerce is the clearest public example I’ve seen of where the work really sits. Their team went live in **90 days**, with test orders flowing into NetSuite within the first **30**. They retired a standalone product information system and ran catalog data natively. And the headline result wasn’t a prettier page — it was a **five-step checkout collapsed into one**, product pages loading in half the time, and **45% holiday revenue growth** year over year. None of that came from the theme. It came from the plumbing. > **A rule I live by** > > Before anyone opens a design file, run a **field-level data audit.** List every system that holds customer, inventory, or order data. For each field, name the system of record — one system, not two. > > Where two systems both claim a field, you’ve found your launch-week outage. In advance, for free. ## Lesson Two: Fragmentation is a tax you pay every month The average consumer brand runs **hundreds of applications**, and only a fraction of them talk to each other. I believe it. I’ve watched a team keep three systems honest with a spreadsheet and a prayer. That fragmentation never shows up as a line item. It shows up as hours. Somebody reconciles inventory by hand every morning. Somebody cancels an order because the website sold what the store already sold. Somebody rebuilds the loyalty report because two systems disagree about who the customer is. Starlight Knitting Society ran a Squarespace site, a Lightspeed register, two inventory systems, and spreadsheets for loyalty. Consolidating onto one Shopify platform gave them back **two hours a day** on inventory alone and more than **20 hours a week** overall — plus organic search that went from nothing to **4,400+ daily impressions**. Slam Jam, coming off Salesforce Commerce Cloud, unified stores and online on Shopify POS. Daily orders rose **15%** once the omnichannel plan switched on, and their setup and running costs fell **50%**. Unified commerce isn’t a buzzword. It’s one database your storefront, your registers, your B2B channel, and your back office all agree on. Everything else is reconciliation dressed up as strategy. ## Lesson Three: Your developers are doing janitorial work Walk into a company running a legacy or self-hosted store and look at what the technical team actually did last week. Patches. Hosting. Memory leaks. A certificate that expired. Plumbing. Not one hour of it moved a single customer closer to buying. The real argument for a managed platform isn’t the license cost. It’s where your people spend Monday morning. Skullcandy’s CEO framed the decision plainly: keep investing in development there, or free up the resources for product and marketing. After the move, launching a product across four regions went from a full day to about **an hour**. Same team. Same payroll. Different work. That’s the number I’d put in the board deck — not the platform fee. > **A word to the next generation** > > The most valuable person on a re-platform isn’t the best developer in the room. It’s whoever can read a P&L and an API schema in the same afternoon. > > Learn both. You’ll run the project, and eventually the company. ## Lesson Four: Pick the window, then hold it DNS propagates on its own schedule. Caches hold old redirects. Data cutovers find the one order status nobody mapped. All of that is survivable in March. None of it is survivable the week before Black Friday. So set a **hard deployment freeze at least 60 days** before Black Friday, Cyber Monday, or whatever your peak is. Put it in writing at kickoff, before anyone is behind schedule and tempted to negotiate. Tie the schedule to your systems, not to a marketing wish. J.Lindeberg replatformed to Shopify Plus in **16 weeks**, timed to follow their new ERP going live — the sequence that makes sense, because the ERP decides what the store can promise. Sales came in **70%** higher afterward, with conversion and average order value both up **7%**. Launch early and quiet. Spend the saved weeks stabilizing. Nobody ever regretted a boring launch. ## If I Sat on Your Board: The question I’d ask A re-platform fails for the same reason most technology projects fail. Not because the software was wrong — because the organization priced the visible 30% and hoped the rest would sort itself out. So before you approve the budget, ask your team one question: > What percentage of this project is below the waterline — and who owns each of those systems? If the answer is a shrug, you don’t have a plan. You have a theme. > **Your next step** > > Before anyone talks redesign, run the field-level data audit. Every system holding customer, inventory, or order data. Every field. One named system of record for each. > > It takes a week. It has saved every project I’ve run six figures and a very long launch weekend. Get that right and the rest is scheduling. Get it wrong and the prettiest homepage in your category won’t save you. **Fix the plumbing first. Then pick the picture.** _Next in this series —_ [**_Module 02: Your Redirect Map Is Worth More Than Your Theme_**](https://gqsdirect.com/blog/redirect-map-seo-preservation-migration) _A re-platform can give back the organic traffic you spent a decade earning — the redirect map is where it happens._ > **About the author** > > **Gary Q. Smith** — President, GQS Direct, LLC > > Gary pairs four decades of P&L-accountable direct marketing leadership with practitioner-level command of today’s commerce stack. He is currently an AI evangelist and practitioner and has led four e-commerce website implementations and integrations, built marketing databases, was architect and project leader for a custom ERP system, and managed omnichannel marketing strategy and communications for several businesses. > > _Most consultants have studied these problems. GQS Direct has owned them._ > > [**Start with a complimentary consultation →**](https://gqsdirect.com/booking.html) **The E-Commerce Implementation Series** - Module 01 — Re-Platforming Is a Systems Project, Not a Web Design Project _(you are here)_ - [Module 02 — Your Redirect Map Is Worth More Than Your Theme](https://gqsdirect.com/blog/redirect-map-seo-preservation-migration) - [Module 03 — Middleware: The Key to Effective Integration and Project ROI](https://gqsdirect.com/blog/middleware-integration-project-roi) - [Module 04 — The Open-Source Paradox: Hidden Maintenance and DevOps Costs](https://gqsdirect.com/blog/open-source-paradox-three-year-tco) - [Module 05 — Your B2B Buyer Is Not a Consumer With a Bigger Cart](https://gqsdirect.com/blog/b2b-ecommerce-platform-decision) - [Module 06 — Your Next Customer Won’t Have Eyes](https://gqsdirect.com/blog/agentic-commerce-machine-readable-catalog) - [Module 07 — The Re-Platforming Payoff](https://gqsdirect.com/blog/replatforming-payoff) --- This blog is powered by Superblog. Visit https://superblog.ai to know more. --- ## The $2.4 Million Blind Spot Author: Gary Q. Smith Author URL: https://gqsdirect.com/blog/author/gary-q-smith Published: 2026-10-26 Category: AI Strategy Category URL: https://gqsdirect.com/blog/category/ai-strategy Meta Title: AI Readiness: The $2.4 Million Blind Spot | GQS Direct Meta Description: Seven in ten AI projects never make it past the pilot, at an average loss of $2.4 million. Four lessons from four decades of technology waves on data readiness, talent, budget and starting small. URL: https://gqsdirect.com/blog/ai-readiness-24-million-blind-spot **I’ve seen this movie before.** In the 1990s, it was the marketing data warehouse. Then e-commerce platforms. Then marketing automation. Every wave came with the same promise and the same panic: _buy now or be left behind._ The merchants who won were never the ones who bought first. They were the ones who did the math first. AI is the same movie with a much bigger budget. And the early reviews are brutal. **73% of businesses that attempt AI without a proper readiness assessment fail within the first year.** McKinsey’s data puts the average loss at **$2.4 million** per failed project. Seven out of ten never make it past the pilot. Those aren’t bad breaks. They’re unforced errors. I’ve planned circulation for 40 million catalogs a year, grown e-commerce to nearly half of company sales, led Shopify and Klaviyo implementations, and served as lead architect on a custom ERP. I’ve sat on both sides of the vendor table. So take this from someone who has actually run the business: > “Diving into AI without an assessment isn’t innovation. It’s a gamble with house money — and the house is your business.” Don’t measure this decision by next quarter. Ask where your business stands **three to five years from now** if you get it wrong — and where it stands if you get it right. Here are the four lessons I’d want you to carry into that decision — the ones that put you among the 27% who succeed. ## Lesson One: Clean data before clever models Catalogers learned this decades ago. A mailing list full of duplicates and dead addresses doesn’t get better because you print a prettier catalog. AI works the same way. It isn’t a magic wand. It’s a math problem, fed by your company’s history. If the history is a mess, the AI becomes an expensive mess. Right now, **67% of small and mid-size companies** have their data scattered across silos with no centralized strategy. A quick glance at your customer file won’t tell you whether you’re ready. The Standard will: **The AI Data Readiness Audit** The Mistake The Standard **Scattered data.** Data sits in silos with no centralized strategy. **Centralized infrastructure.** Data is clean, accessible, and verified against ground-truth sources. **Surface audits.** Assuming you “have data” because you have customer records. **High volume.** A minimum of 10,000 data points to train a model properly. **Recency bias.** Looking only at the last few weeks of performance. **Historical depth.** 6–12 months of operational history to capture seasonal patterns. **Lazy labeling.** Relying on unverified, messy datasets. **95% labeling accuracy.** Precision-labeled data for supervised learning. > **A rule I live by: the One Week Rule** > > If it takes your team more than one week to extract and analyze a specific dataset, you aren’t ready for AI yet. > > That’s not a failing grade. It’s a to-do list. **Fix the plumbing before you buy the expensive faucet.** ## Lesson Two: Grow your own people first The instinct is to hire a $300,000 data scientist on day one. Resist it. Hiring too early is one of the most reliable shortcuts to failure I know. The AWS Playbook describes four ways to source talent: - **Build** — upskill your existing team. For most businesses, the only sustainable path, because your people already hold the institutional knowledge: how your customers buy, why your operations work the way they do. No outsider walks in with that. - **Buy** — hire fresh talent with foundational AI skills. Fast, expensive, and a real cultural risk. - **Borrow** — contractors for specialized, short-term needs. - **Bot** — let AI handle the task itself, essentially “hiring” the software. The most valuable person in your building right now isn’t a coder. It’s the **AI Translator** — someone who understands the business problem and can turn it into technical requirements. Without that person, your technical team will build a $2.4 million paperweight. > **A word to the next generation** > > If you’re early in your career, this is your moment. In every technology wave I’ve worked through, the people who could speak both languages — business and technology — ended up leading. > > Learn the customer. Learn the P&L. Learn the tools. **Become the translator.** The industry needs you more than it knows. ## Lesson Three: Budget for the humans Every system implementation teaches the same lesson: the software goes live on a date. The people go live when they trust it. That trust has a price. Alice Labs’ data shows **change management consumes 20–30% of the total AI project budget** — the Change Management Tax. Skip it, and you get adoption failure. The traditional ROI model won’t warn you. The **Levelized Cost of AI (LCOAI)** framework, published in _ScienceDirect_ in 2026, finds that standard models undercount costs by **40–60%**, because they treat AI as a one-time purchase rather than a living operational system. Before I review any AI budget, I want the **Technology Readiness Level (TRL)** assessed to set the contingency buffer: - **TRL 1–3 (prototype):** carry a ±50% contingency buffer. - **TRL 7–9 (production-ready):** you still need ±20%. And I want to see these recurring costs in the plan — not discovered in year two: - **Recurring inference fees** — the cost of every “ask” you make of the model. - **Model drift monitoring** — models get “stupider” over time, and keeping them sharp costs money. - **Retraining cycles** — AI is not set-it-and-forget-it. ## Lesson Four: Start small. Stay honest. You don’t need a 50-page strategy deck. You need to test the water before you jump in. Here’s the process I recommend: 1. **Pick one painful, repetitive task.** Start with a low-risk annoyance — summarizing meetings or drafting follow-ups. 2. **Choose one mainstream tool.** Don’t shop for months. Pick a major general-purpose assistant and pay the $20–$30 a month for the paid version. Professionals pay for privacy — and for better data handling and quality. 3. **Run a two-week test.** That tool. That task. Period. 4. **Write a one-page internal note.** Document the how-to. That page is the seed of your official AI policy. 5. **Expand slowly.** Once it’s stable, add one more task or one more person. And watch for the **90-Day Killers** — the habits that sink good teams early: - Evaluating ten tools at once. - Starting with high-stakes, customer-facing work. - **Shadow AI** — keeping AI use secret from the team or leadership. Trust is your core asset. Don’t spend it here. - Treating one bad output as proof the technology “doesn’t work.” ## If I Sat on Your Board: The question I’d ask AI success is **20% technology and 80% people, process, and culture.** That was true of every wave before this one, too. The technology is the easy part. Execution is what separates the survivors. So as you look at next quarter’s budget, and the three to five years beyond it, ask yourself: > What is our Automation Readiness Score? If you can’t answer that with hard numbers, you’re flying blind — and that’s a $2.4 million mistake waiting to happen. But here’s the good news. You’re not behind. Not yet. The merchants who win this wave will be the ones who pause, measure, and move with purpose. **Be one of them. I’m glad to help you get there.** > **About the author** > > **Gary Q. Smith** — President, GQS Direct, LLC > > Gary pairs four decades of P&L-accountable direct marketing leadership with practitioner-level command of today’s commerce stack. He is currently an AI evangelist and practitioner and has led four e-commerce website implementations and integrations, built marketing databases, was architect and project leader for a custom ERP system, and managed omnichannel marketing strategy and communications for several businesses. > > _Most consultants have studied these problems. GQS Direct has owned them._ > > [**Start with a complimentary consultation →**](https://gqsdirect.com/booking.html) --- This blog is powered by Superblog. Visit https://superblog.ai to know more. --- ## About Atrium Author: Gary Q. Smith Author URL: https://gqsdirect.com/blog/author/gary-q-smith Published: 2026-10-23 URL: https://gqsdirect.com/blog/about Atrium is an inn at the head of the valley, built against the cliff in 1908 and run by the same family since. Eleven rooms, one kitchen, a stove in the lounge, and the lake walk from the door. ## The rooms Eleven, no two alike. Every one but two faces the valley; the two that do not face the peak, and they are the ones people ask for. Room 4 has no television, and the others are following. ## The kitchen Jonas cooks what the valley grows: bread from the village oven, butter from the farm below, trout from the river on Tuesdays, and whatever the orchard has done that year. Breakfast until eleven, because of the lake. ## This journal Where we write about the walks, the rooms, the kitchen and the seasons as they actually arrive. One post a week, on Mondays, when the rooms are quiet. ## Getting here The bus from the station stops in the village square four times a day. Tell us which one you are on and we will meet it. --- This blog is powered by Superblog. Visit https://superblog.ai to know more. ---