We fill in your Product Type field. You send a store address, we send the finished file.

Shopify never forces you to fill that field in, so on most stores it is blank on the majority of the catalogue and nobody notices until something that depends on it stops working. We read your whole catalogue, take the categories from the collection names you already use, and hand you a CSV.

Send your store addressThe first few are free while we build up case studies. Say no afterwards and that is a complete answer.

Check your own store first — it takes ten seconds and needs nothing from us

Open yourstore.com/products.json?limit=250 in a browser and search the page for "product_type":"". Every match is a product with the field empty. If you barely find any, you do not need this page and we would rather you closed it.

What the empty field actually breaks

Three things, all of which read as a bug in something else.

Automated collections: A smart collection with a “Product type is equal to” condition matches that exact string. Blank, and the product silently never enters the collection — no error, it just is not there.
Storefront filters: Search & Discovery only offers a filter value if products carry it. With the field empty the filter either does not appear or appears with two options, and it reads as a broken app rather than empty data.
Google Shopping and Performance Max: Listing groups subdivide on product_type. With it blank your whole catalogue lands in “Everything else” and cannot be bid on, excluded, or reported by category. (It does not get you disapproved — that is google_product_category, which Google assigns for you.)

How it works

1
You send the store address. Nothing else.
No app to install, no collaborator access, no login. Your catalogue and your collections are already public at /products.json and /collections.json — the same endpoints your theme uses.
2
We read the whole catalogue, not a sample.
Walked page by page to the end of the feed. You get the real total, not the first 500 products with a percentage extrapolated off it.
3
Every blank product gets the narrowest of your own collections it sits in.
The values are words you already use. Shelves are thrown out, not used as types — New, Sale, Best Sellers, In Store Pickup, Under $50, a seasonal drop code, a theme slider. Those describe where a product sits, not what it is.
4
Anything we cannot place from your own collections is left out.
Deliberately. A wrong type is worse than an empty one, because a wrong one silently files a product in the wrong place and nobody goes back to check.
5
You get a three-column CSV — Handle, Title, Type — or we apply it for you.
Export your products before importing anything from anyone: Shopify’s importer matches on Handle and can touch variants, and we will not tell you a stranger’s spreadsheet is risk-free on a catalogue we do not own. Or send a collaborator invite and we apply it on a draft basis so you can look before anything is published.

What it has actually produced

38 catalogues read so far. 4 of them already had a Product Type on every single product — nothing to do, and we said so rather than invent a problem. Of the 34 with real blanks, 20 produced a file we were willing to hand over and 14 did not — the store's own collections were too coarse to type most of the catalogue, so no file was sent. Every other number on this page — the totals, the median, the floor — is measured over those 20 only. So read it as 20 out of 34, not as a hit rate. The 14 are named here instead of averaged in, because averaging a refusal into a median is how a page quietly turns it into a result.

5,650: blank fields filled
7,215: products read end to end
95%: of the gap closed, median catalogue
1,204: products deliberately left untyped

The weakest file we finished still covered 45% of that store’s blank products — that is the floor, not the average. The largest catalogue finished was 2,299 products, and the most distinct types produced for a single store was 111.

The 20 of 34 catalogues with a real gap whose file was good enough to hand over — the other 14 are counted above, not listed here. No store is named: these are prospects, not customers, and none has agreed to appear on a page selling to somebody else.
Catalogue size	Field blank on	We filled	Left out	Types used
over 2,000 products	2,276 of 2,299	2,090	186	104
500-2,000 products	1,664 of 1,664	944	720	70
100-500 products	499 of 499	488	11	111
100-500 products	324 of 330	324	0	43
100-500 products	290 of 308	211	79	31
100-500 products	289 of 375	283	6	30
100-500 products	275 of 356	273	2	14
100-500 products	186 of 188	183	3	14
100-500 products	169 of 195	169	0	26
100-500 products	134 of 150	94	40	10
100-500 products	116 of 116	84	32	8
100-500 products	114 of 146	73	41	28
under 100 products	97 of 97	83	14	6
under 100 products	91 of 97	89	2	7
under 100 products	91 of 91	41	50	6
under 100 products	59 of 62	45	14	23
100-500 products	55 of 112	55	0	23
under 100 products	45 of 45	44	1	4
under 100 products	43 of 47	43	0	7
under 100 products	37 of 38	34	3	14

What it costs

Right now, nothing, for a small number of stores. We need the case studies more than the money and would rather have a merchant who has seen the file than one who has read this page. After that it is a flat one-off per catalogue, quoted once we have read yours and know how much of it we can honestly place — not a subscription, and not a price quoted before anyone has looked.

There is no revenue claim anywhere on this page. We have never measured a sale caused by this fix and will not imply one. What is measured is above: how many fields were empty, how many we filled, and how many we refused to guess at.

Send your store address Or scan your store yourself first — separate thing, no email needed

Figures on this page were last recomputed on 2026-09-30 from the finished files themselves. Nothing here is typed in by hand, so it cannot drift away from the work.