Why a clean catalog doesn't stay clean

A big catalog cleanup feels finished on the day it ends. The duplicates are merged, the product types agree, the missing attributes are filled in. Then the catalog keeps living. New products arrive, a supplier sends a revised file, someone edits a batch of listings in a hurry. Six months later you open the catalog, spot a handful of familiar-looking problems, and wonder whether the cleanup was worth it or whether you need another one.

Both questions are hard to answer from a glance, because a glance mixes up three different things: problems the cleanup missed, problems that arrived afterwards, and changes someone made on purpose. Each one points to a different response. The review below separates them, and you can run it with a spreadsheet and the records you already keep.

Start from what the cleanup actually covered

Before you judge today's catalog, dig out the record of the original job: when it finished, which products and fields it covered, the house rules it applied, how it was checked and any exceptions it left open. Keep that scope in front of you. If the cleanup was about product types, an image problem today says nothing about whether it worked.

Write down what the job cost as well: preparing the data, reviewing candidates, editing or importing, and checking the result. Note whose time it was and whether the figure comes from a record or from memory. If you don't know, write unknown. A blank you're honest about is more useful than a number you made up.

List what changed since

Next, build a timeline of what happened since, from whatever records you have: new products, imports, manual edits, new supplier files and changes to your house vocabulary. Note which products and fields each one touched, and be clear about where your history has gaps. A current export on its own can't tell you what changed.

Take a tidy mug catalog whose owner adds a supplier range in month two and imports a revised file in month four. Each event raises a question rather than answering one. The new file may repeat products you already sell, or use a different word for the same product type, and a new bundle may be meant to stay separate. Only the records can tell you which.

Check the changed records first

Now compare today's data with the original scope and today's house rules. Start with the records the timeline flags, revisit the exceptions the cleanup left open, then look at a set of products nobody meant to change. You're looking for repeated listings, inconsistent attribute values and missing values in the fields the original job covered. Leave legitimate variants and intentionally separate listings alone.

Give each thing you find a reason, so the work it causes is charged to the right place:

Classify what the review finds
What you foundWhat it meansCount it as
A new issueArrived with a later import or editLater correction
An old missWas there all along, inside the original scopePart of the original job
An intentional changeSomeone meant it, such as a new bundleNo correction
A false alarmLooked wrong but isn'tChecking time only
UnresolvedYou can't tell when it arose, or whether it's wrongAn open question

Count the time honestly

  1. Log each episode once

    One row per piece of work: the date, what triggered it, the products and fields covered, and what it found. Keep checking, correcting and recovering in separate buckets.

  2. Use person-hours

    Add up each person's actual effort. Waiting for access or a decision is elapsed time, not labor; tool fees and invoices are money, not hours.

  3. Leave gaps as gaps

    If you have durations for only some episodes, call the result a recorded subtotal. A recollection can inform your next decision, but label it as one.

Choose follow-up that fits the catalog

Barely changes

No real imports or edits since the cleanup. A light periodic check may be all it needs, and that's a good result, not a reason to manufacture work.

Changes in batches

A new range or a supplier file now and then. Review each import when it lands, against the same rules.

Changes every week

Frequent edits across the fields you care about. This is where regular or continuous checking is worth considering.

Not sure yet

Run the six-month review once. What it finds, and what it costs, tells you which of the three you are.

How Pristine keeps a catalog clean after the cleanup

If the review shows your catalog changes every week, checking by hand becomes a job of its own, and that's the job Pristine was built for. It covers the three steps this review keeps coming back to. On every platform it finds what needs attention, with each finding showing the products involved and the reason. On Shopify, finding is free; Starter and above fix what you approve, saving the old values first so you can restore a change, a batch or a day for at least 90 days while your store is connected; and on Growth and above, Shield checks each product for duplicates as it's saved and rescans your catalog weekly or monthly once you turn that on. On BigCommerce, fixes with a restore point are free during the open beta. Uploaded and exported catalogs are read-only: you get the findings and make the changes yourself. The platform comparison in these docs lists what each one includes.

Those are candidates to review, not a guarantee that no new issue gets through, and not a measure of time saved. You still decide what's an issue and what's intentional. Product analytics record which features are used, never your product content, so keep your worksheet in your own records.

Want to see what your catalog has picked up?

Send us your store's address and we'll run Pristine's checks on a sample of your catalog, then send you a private report within a day. On Shopify we read your public storefront; other platforms send a product export.

Get a free catalog demo