Two hundred advertisers, no order.
Search any category and you get a list. A few names you recognize, and a long tail you don't. Somewhere in that tail is a company outspending you this quarter, and next to it is a founder testing a hundred dollars a day. Nothing on the page tells you which is which.
That's the state of ad data by default. Platforms show you that an ad exists. They don't tell you how much weight is behind it. So teams fall back on the proxy they have, brand recognition, and end up studying the competitors they already knew about while the actual movement happens elsewhere.
Scale grades exist to put an order on that list.
Three scales, not one score.
Every advertiser we surface carries three grades, each independent of the others, each running from D at the bottom to A+ at the top.
How much weight is behind the advertising itself, across platforms, right now.
How much attention the business actually pulls, paid and organic together.
How much commercial volume sits behind the operation.
We keep them separate on purpose. Collapsing three signals into one number feels tidy and destroys the most useful information in the data. A company that spends heavily with little sales volume behind it is a different animal from one with heavy sales and modest spend, and a single score would render them identical.
How a grade is made.
Each scale is produced by our own model, fed by every signal we can responsibly gather about an advertiser: the ads they run and the volume behind them, the platforms they run on, the properties they own, the traffic those properties draw, and the commercial footprint around the business.
We enrich that with third-party data where it makes the picture sharper, then reconcile everything against a single resolved entity. That resolution step matters more than any individual source. An advertiser running under four names, three domains, and a regional variant is one company, and grading the fragments separately would rank all four as small.
The result is a graded advertiser rather than a graded ad account. It's an estimate, and we treat it as one; the grades are built to be right about relative position, not to reproduce a number on someone's internal dashboard.
Why letters and not numbers.
A number invites false precision. Told an advertiser spends 4.2 million a year, you will reason about the 0.2, and neither we nor anyone outside their finance team can stand behind it. A grade says what's actually knowable: this advertiser sits near the top of its market, or near the bottom.
Letters also survive the thing numbers do badly, which is comparison across very different categories. And they are compact. A model can hold a whole category of graded advertisers in context and sort them without a single extra call, which is the difference between asking about one competitor and asking about a market.
The combinations tell the story.
The grades are most useful read together, because the gaps between them describe what a company is doing.
A+ spend with a low sales grade usually means a land grab or a heavy test phase; worth watching closely, since they're buying learning fast. Strong sales with a modest spend grade points to a business that isn't buying its growth, so their creative is worth studying for a reason other than budget. High traffic with low spend suggests the demand is already there and the ads are topping it up.
Sudden movement is its own signal. An advertiser climbing from C to A on spend inside a quarter is committing to something, and the ads they launched on the way up are the ones to look at.
What the grades unlock.
With grades attached to every advertiser, the questions get sharper. Show me only A and above in this category. Rank the free-trial offers by the spend behind them. Find advertisers whose spend grade rose two steps this quarter and show me what they launched.
Your AI can do all of that reasoning already. The grades are what let it tell a serious operator from a hundred dollar test, and get to the answer without you naming the competitors first.