Before "e-commerce SEO" was a conference track, it was a survival skill. We learned it the retail way: at MacMall, eCost, and PC Mall, national e-commerce operations where the catalog ran deeper than any team could hand-tune and a lost ranking showed up in the next morning's sales report. Nobody debated attribution models. The site either sold hardware that day or it didn't.
Twenty years later we run our own products instead of other people's catalogs, but the lessons from that era transfer almost untouched. These are the ones that still decide who wins.
The category page is the business
Retail teaches you fast that product pages are not where the leverage lives. Products go out of stock, get replaced, get renamed by the manufacturer mid-quarter. Category pages endure. The shopper searching for a class of product is worth more than the shopper searching for a part number, because the category shopper hasn't decided yet, and because the category page keeps earning long after any single SKU dies.
So we treated category pages like flagship stores: real copy written by someone who understood the products, deliberate internal linking, faceted navigation that helped shoppers without spawning an infinity of crawlable duplicates. Most e-commerce sites still get this backwards: polished product templates, category pages left as auto-generated grids under a lonely heading. Flip that priority and you've done the highest-value thing in e-commerce SEO.
The related habit worth stealing: merchandise your categories around search demand, not your org chart. Shoppers search the way they think, not the way your buyers organize vendor lines. We built category structures around query language, and when demand shifted, the taxonomy shifted with it. Information architecture is a living SEO asset, not a one-time setup task.
At catalog scale, structure beats heroics
When a catalog is deep, you cannot optimize pages. You can only optimize systems. One template improvement touches every URL that shares it; one template mistake does too. The retail era beat the romance of hand-crafted pages out of us early. The winning move was always the fix that shipped everywhere at once.
Crawl budget was real for us before Google gave it a name. Sort parameters, session IDs, print views, expired promotions. A big retail site manufactures duplicate URLs all day long, and every one of them spends crawler attention the money pages needed. The discipline we built: decide which URLs deserve to exist, make everything else uncrawlable or canonical, and read the server logs, not the rank tracker, to learn what Google actually fetches.
That discipline is the direct ancestor of how we build today. Our broadband directories run thousands of pages on the same logic: templates that earn their keep, internal links that route authority deliberately, nothing crawlable that doesn't deserve to rank. We wrote the method up in our guide to programmatic directory sites.
Products die. Handle the funerals well
Retail catalogs churn constantly, and what separates professionals is what happens to a URL when its product dies. Delete the page and you throw away every link and every signal it ever earned. Leave it live and you send shoppers to a dead end. The retail-era answer still holds: redirect discontinued products to their closest successor or their parent category, keep seasonal pages alive year-round, and never let a money URL return a 404 because someone tidied the catalog.
Out-of-stock is its own art. Temporarily unavailable products should stay indexable, with a clear restock message and alternatives on the page. Permanently gone means a prompt redirect. The worst pattern is the silent soft-404, a live page with an empty buy box, which wastes the shopper's click and teaches Google your site is full of dead ends.
Redesigns are catalog funerals at scale, and they deserve the same care. Every replatform we watched go wrong went wrong the same way: URLs changed, redirects were an afterthought, and rankings built over years evaporated in a weekend.
What still holds in 2026
Search has changed shape. AI answers sit above the blue links now, and shoppers ask assistants for recommendations before they ever open a category page. But look at what those systems cite: fast, structured, unambiguous pages that state plainly what they sell and to whom. Which is to say: the same properties retail SEO has rewarded for twenty years.
Speed still converts. Structured data still earns the rich result. Clean architecture still decides how much of your catalog search engines even see. None of it is glamorous, which is exactly why it keeps working: competitors chase the new thing while the fundamentals sit there compounding for whoever shows up and does them.
The buyers changed too. The retail era was mail-order catalogs moving online; today the same shopper researches on a phone, price-checks in a store aisle, and expects the answer in one scroll. Every era rewards the site that respects the shopper's time. That part has never changed.
If there's one meta-lesson from those years, it's that SEO belongs inside the revenue conversation, not in the marketing-report appendix. We measured our work in orders, and it made us better at everything upstream of the order. That's still how we run SEO engagements today, and it's how we'd suggest you judge anyone you hire. When you're ready to hold the work to that standard, start with how we measure SEO ROI.