Ad Intelligence
Winners and creatives
The five tabs
| Tab | What is on it, and when to use it |
|---|---|
| Scaling now | Advertisers, not products. Every competitor with active ads in this market, judged on how many ads run at once, how far they reach, how long the run has lasted and how many are new. It needs no product match, which is why it always has something to say. This is the default tab. |
| Daily 30 | Thirty new products a day, highest score first, never the same product twice. For finding things. |
| Proven | Products scoring 60 or more out of 100. For following what you have already found. |
| New and rising | Products first seen in the last month that are already being pushed hard, before they can have proven anything. Earlier, and with more false starts. |
| Not ranked, and why | Everything that was measured and did not make a list, with the reason in plain words. Check an empty list against this before believing it. |
Above the tabs is a row of countries. Ads are collected per market and nothing is ever added across them — an ad asked for in the Netherlands and in Germany is two observations, not one bigger one.
How the score works
Four kinds of evidence, twenty-five points each, a hundred in total. Each answers a different question about whether somebody is really spending.
Longevity — how long it keeps running
Nobody keeps paying for an ad that loses money. A product still running after thirty days is the single cleanest signal there is.
Volume — how many creatives at once
One ad is a test. Fifteen running simultaneously is a budget.
Iteration — they keep making new ones
New creatives launching while the old ones still run means somebody is actively working on this product, not letting it coast.
Expansion — how far it has spread
More countries, more shops selling the same product, more angles being tried. A competitor copying it is a positive signal here, not a saturation warning.
What the total means
| Under 60 | Skip. |
| 60 – 69 | Only worth it if you have a unique angle on it. |
| 70 – 79 | Test-worthy. |
| 80 – 89 | High priority. |
| 90 + | Test immediately. |
One override: a product scoring at least 15 out of 25 on all four parts is promoted to high priority whatever the total says. The card marks it All four present. Broad evidence beats one tall spike.
Two floors, no matter what the parts add up to
A product needs at least three creatives, and at least five thousand people reached where a source reported reach. Fail either and the card reads Skip and lands on Not ranked, and why with the reason named — for example “Only 1 creative — one ad running long is not a scale-up.”
What is deliberately not scored
Estimated revenue, views, likes, reviews and discounts. None of them survive contact with reality: they are either invented, gameable, or say nothing about whether a product sells.
Estimated spend and reach do appear on the cards, as context. Reach is used as a floor on the list; it is never a term in the score.
The disclaimer, and why it is there
The score measures visible commercial signals — how many advertisers, how many ads, how long, how wide. It is not evidence of revenue, profit or return. Nobody outside a shop can see those, and a tool that pretends otherwise is lying to you.
If Winners is empty
It will not just say “no results”. It shows a five-step checklist with your real numbers and a tick or a circle against each. Fix the first unticked one.
- Shops watched — the list is the strategy. Nothing here means nothing anywhere.
- How many have a Meta page id — without it Shopdeck sees what a shop sells, not what it advertises.
- Products read from their catalogues — a shop that hides its catalogue can only be watched for ads.
- Ads linked to a product — the hard one. Meta publishes no landing page, so linking is by matching words unless TrendTrack supplied the page.
- Active ads collected for this market — outside the EU, Meta only shows a slice. Try an EU market first.
Whatever the state, Scaling now usually has something, because it works at advertiser level and needs no product match at all.
What to do with a winner
- Open its creatives. Read the copy, watch the video, note the hook in the first three seconds and what the ad promises.
- Look at the launch timeline on the card. A steady drip of new creatives over weeks is a very different story from twelve launched on one day.
- Press Get the pictures when Meta stripped the images and you want to see the creative. It spends TrendTrack credits and tells you how many afterwards.
- Then go and build something. This screen is research; the campaign is built in the Campaign Builder and judged in the Profit Dashboard.
One thing the Admin screen does not do
There is a Scoring tab under Admin with weights you can edit. It does not tune the winner model on this page — it belongs to an older score still shown on product detail pages. Changing it will not move anything on Winners.