About
A consultancy that ended up writing software.
A boutique, data-driven digital marketing consultancy based in Istanbul, running since 2022. The product on this site exists because the same problem kept appearing on every account we touched.
Since 2022
The consultancy
Running ad accounts, CRM lifecycles and the measurement underneath them.
Then
The product
Built because the product feed kept turning out to be what held those accounts back.
How it happened
The bottleneck was never the campaign.
Verintra started in 2022 doing what small consultancies do: taking on the accounts that were too small for an agency retainer and too important to leave on autopilot. Google Ads, Meta, Criteo. Campaign structure, bidding, creative rotation. The usual, done carefully.
Doing that repeatedly across different stores produces a pattern you cannot unsee. On most accounts the largest available improvement was not in the campaign at all. It was one layer down, in the product data the campaign was built on. Titles that did not say the size. Half a catalog with no category. Missing identifiers, so the platform could not tell what the product even was. No amount of bid tuning fixes a title.
So the work became: fix the feed first, then run the campaign. That worked, and it did not scale. Every client meant another spreadsheet, and every spreadsheet went stale inside a week because catalogs change daily.
The tool started as a way to stop doing the same work by hand for the fourth time.
It grew from there. Scoring, so the size of the problem was visible before anyone argued about it. Then AI content generation, but constrained, because a model that invents a material is worse than an empty field. Then export to the channels, then Merchant Center and Meta, then the second score, for whether an assistant can understand a product at all rather than merely accept it.
That second score is why the product is not just a feed tool. Shopping traffic is no longer only people typing into a search box. Assistants answer product questions now, and they answer from structured data rather than from your homepage copy. A catalog that reads well to a crawler and says nothing to a model is a catalog that quietly stops being recommended.
How we decide things
Four rules we do not bend.
These are not values on a wall. Each one closed off an easier option in the product or on this website, which is the only test of whether a principle is real.
Nothing is invented
When the AI writes a title, it may only use attributes already on the product. When it picks a category, a local filter narrows the official taxonomy first and the model may only choose from that shortlist. A category that was never offered cannot come back. This is architecture rather than a prompt asking it politely.
Read before write
The free scan has no write path to your store. Not as a policy someone could change, but because it was never built. You see the grade before granting anything, which is the reverse of how most feed tools ask to be trusted.
Two scores, kept apart
Channel compliance and machine comprehension are different questions, and averaging them into one number destroys the only interesting information: the gap between them. So there are two, and the product shows which dimensions explain the difference.
No number we cannot defend
You will not find client percentages, logo walls or testimonials on this site. We are pre-launch with a handful of real customers, and inventing social proof would contradict the first rule on this list. Every figure here traces to something in the product you can check yourself.
Two halves, one problem
Why both are on the same site.
People sometimes ask which one is the real business. Both are, and they feed each other in a way that would be hard to fake.
Software, for people who want to do it themselves
Self serve. Grade the feed, work down the ranked fixes, publish to the channels. If your feed is the whole problem, this is all you need and you will never have to talk to us.
What it does →People, for when the feed is not the whole problem
Advertising management, growth and CRM, measurement and dashboards. Usually the feed is one symptom among several, and the others need someone in the account rather than a tool.
How we work →The product keeps the consultancy honest, because it measures the thing we used to assert. The consultancy keeps the product useful, because it is built against real accounts rather than a roadmap.
Working together
Remote, and in two languages.
The company is Turkish and the clients are not all in Turkey, so everything runs remotely. That has been true since 2022 and it has never been the constraint anyone expected it to be.
This site is in English with a Turkish translation. Write in whichever you prefer. Nothing gets lost, and nobody is waiting on a translator.
Start with the grade, not with a meeting.
The fastest way to find out whether we are useful to you is to let the product tell you where your feed stands.
Grade my feed, free
About two minutes, read only, and you may not need us at all.
→Talk to us
For everything the feed alone cannot explain.
→Verintra Digital, a marketing consultancy since 2022