I stopped adding image tools and rebuilt around one ecommerce workflow

Nice pivot. The platform-specific point is the key: I’d pick one marketplace and instrument the shortest path from first scan to successful export, then segment repeat usage by seller rather than raw scans. A useful activation metric could be “fixed at least one flagged image and exported a clean catalog within 24h,” paired with the top failure category. I’m running a $0 experiment with digital-seller tools, and the same lesson is showing up for me: catalog breadth gets attention, but one measurable job gives distribution a chance. Curious whether repeat sellers or first-time audits currently dominate.

The pivot from "collection of utilities" to "one workflow for one persona" is exactly right. Generic tools attract curious visitors; workflow tools attract returning users with a job to finish.

The on-device processing is a genuine competitive moat for marketplace sellers who deal with proprietary product images — they won't paste those into a cloud tool if they can avoid it.

One thing worth testing: the failure grouping step. If sellers see the same failure type repeating across 200 photos, the impulse is to fix in batch. If your export flow surfaces the top 3 failure patterns with a "fix all like this" option, that becomes a reason to upgrade or pay vs. just check and close the tab.

Also curious — are you seeing repeat use from the same sellers, or is it mostly one-and-done catalog checks before launch?

Going from a toolbox to one job is the right call, but the harder version of that decision is still ahead: Amazon, Shopee, Lazada and TikTok Shop are four different products with four different rejection rule sets, so supporting all four rebuilds the same vagueness one level down. I would pick the platform where a rejected listing costs the seller the most money and own that one completely. And marketplace sellers are not on launch platforms, they are in seller forums and inside the agencies managing catalogs for them, which is where I would spend the next month.

The sharper workflow is a big improvement: you’ve turned ‘image tools’ into a marketplace problem with a clear outcome.

I’d make the first promise even more concrete: ‘check your catalog before it gets rejected,’ then show one platform-specific failure, the fix, and a before/after report. That gives sellers enough proof to act before committing a whole folder.

This is a really honest lesson to share — "traffic from generic launch audiences does not equal seller demand" is something a lot of us learn the hard way after a launch. Going from "collection of utilities" to "one specific job for one specific user" seems like it would also make your marketing way easier, since you're not trying to explain five different use cases at once anymore.

On your question: platform-specific grouping (Amazon vs Shopee vs TikTok Shop) is smart since their actual rejection reasons differ. Does the QA report tell sellers exactly which rule was violated (e.g. "background not pure white" vs just "image failed"), or is it more of a general pass/fail per photo right now?

The shift from generic traffic to own-photo checks and exports is a much stronger signal. Once sellers actually run their catalogs through it, do you see repeat usage or completed exports from the same sellers, or is that the next piece of evidence you’re waiting on?

The move from a bag of image utilities to one seller workflow feels much clearer. Measuring completed photo checks and exports instead of page views is the right correction too. For the report, I’d want each failed image to show the exact marketplace rule and a before/after example, because “wrong aspect ratio” still leaves a seller guessing what to change.