A catalogue that undersold itself
The homepage told visitors the store carried around 800 items. The database held 1,515, with 1,513 of them in stock. The shop was almost twice the size its own front page claimed.
Case study · Resale e-commerce
1,515 items. Eleven pages doing the selling.
Jojin’s Kitty Thrift was not short of stock, and not short of search visibility — 8,360 impressions in ninety days. It was short of routes. Of the 199 clicks in that window, 193 landed on eleven pages. Every product page in the catalogue, between them, produced six.
Start a projectThe diagnosis
Read the data before touching the site.
Every change below came from a number rather than an assumption. Two of those numbers contradicted what the shop was saying about itself.
107 categories, one carrying a third of the store
Stock was spread across 107 distinct categories. The largest, labeled simply Toys, held 450 items — close to a third of everything in the shop filed under one generic word.
Clicks landing on eleven pages
193 of 199 clicks went to eleven non-product pages. The product pages — the actual inventory — returned six clicks between them, and only 395 of 1,515 were indexed at all.
The work
Build the routes, not more stock.
Shipped 17 September 2026. The figures above are the measured baseline for the ninety days before that date, 19 June to 14 September 2026. No outcome is claimed on this page yet — there has not been time for one, and a number invented today would be worth nothing to you.
-
01
Correct what the shop claimed about itself
The “around 800” figure was replaced with the real count everywhere it appeared. A storefront that understates its own inventory hands the visitor a reason to leave.
-
02
Give the large categories their own pages
Six new category pages went in — wall art, movies, vintage toys, jewelry, watches, home decor — each one a browsable route into stock that previously existed only as rows in a database.
-
03
Fix the filter that was hiding stock
The retro video games page matched against a term list that missed three of its own labels, hiding products from the page built to show them. Corrected, and the six new pages were added to the sitemap so they can be found at all — see AI for small business.
Questions
What this did and did not prove.
What was actually wrong with the site? +
Nothing was broken. The store had 1,515 items and real search visibility — 8,360 impressions over ninety days. The problem was distribution: 193 of the 199 clicks in that window landed on eleven pages, while every product page between them produced six clicks in total. The catalogue was large and the routes into it were narrow.
Why not just add more products? +
Because volume was never the constraint. The catalogue already held 1,515 items across 107 categories, and the homepage was telling visitors it held around 800. Adding stock to a catalogue nobody can browse only makes the same problem bigger.
How long before this shows up in search results? +
Unknown, and anyone giving you a number is guessing. New category pages have to be crawled, then indexed, then earn position. That is weeks, not days. The baseline on this page is dated so the next measurement is a real comparison rather than a story told backwards.
Why does this page not show a result? +
Because the work shipped on 17 September 2026 and no result exists yet. Publishing one now would mean inventing it. This page shows the diagnosis and the change; the measured outcome gets added when there is one.
Does this only apply to resale and thrift catalogues? +
No. It applies to any catalogue where individual items vastly outnumber the pages people can browse — resale, parts, inventory, property listings. The pattern is identical: the items exist, the routes to them do not.