Public ad libraries hide almost every performance number. The one they leave in plain sight is time.
What a public library shows
Open any of the big libraries and you get the creative, the copy, the advertiser and some dates. Meta shows when each ad started running. Google's Ads Transparency Center shows the last date an ad ran. TikTok's Commercial Content Library, which covers ads shown in the EEA, Switzerland and the UK, lists the dates an ad ran. LinkedIn shows run dates for ads targeted at the EU.
In some regions you get a little more. For ads delivered in the EU, Meta adds an estimated reach figure and the targeting the advertiser chose. TikTok's library includes targeting details and the number of people an ad was served to. LinkedIn adds impressions and targeting for EU-targeted ads.
What you will not see, for ordinary commercial ads, is spend, clicks, sales or return. The numbers the brand actually uses to make decisions stay private.
So you have a choice. Guess at the private numbers, or read the public ones well. This post is about the second option.
Why the other numbers mislead
Each number you can see tells you something. Most of them tell you less than they seem to.
| Number | What it seems to say | Why it misleads |
|---|---|---|
| Likes and comments | People love this ad | They measure reaction, not sales. Arguments and giveaways score well. |
| Number of live ads | This brand is winning | Volume shows activity. Thirty new tests prove nothing yet. |
| Reach, where shown | Big money behind it | Reach shows delivery, not return. A brand can buy reach for an ad that does not sell. |
| Spend estimates from tools | What the brand spends | Modelled, not reported. Useful as a rough range at best. |
| TikTok Top Ads | These are the winners | Good for ideas, but it is a curated set that advertisers agreed to share. It will not show you one competitor's whole account. |
None of these are useless. They just answer a different question from the one you care about, which is: is this ad making the brand money?
Why time is harder to fake
Every day an ad runs, someone pays for it. Keeping it live is a decision the brand makes again and again, based on numbers you cannot see: cost per purchase, return on spend, margin.
That is what makes days running useful. If an ad has been live for 90 days, the people with the real numbers have looked at them many times and kept going. Few brands keep funding an ad for three months to impress anyone else.
Days running is the brand's private dashboard, leaked as a date.
There is a second reason it works. Likes capture how people felt in the first few hours. Days running captures how an ad held up over weeks of changing audiences, costs and competition, as judged by the people who can see the sales.
It also has practical advantages over everything else on the page:
- It is everywhere. Every major library shows some kind of date.
- It is comparable. A day is a day, whatever the format, platform or category.
- It is expensive to fake. Engagement can be inflated. Longevity costs real money every single day.
Where days running goes wrong
Days running is a proxy, not proof of profit. It fails in predictable ways, and it helps to know them before you lean on it.
- Small budgets. An ad can run for months on a few dollars a day. Long life, little spend.
- Forgotten ads. Sometimes nobody switched it off.
- Always-on campaigns. Retargeting and brand ads are often left running and judged on different terms.
- Seasons. A Christmas ad stops in January because Christmas ended, not because it failed.
- Start dates. The date is when delivery was set to begin. It does not prove the ad ran every day since.
- Relaunches. When a brand brings an old ad back as a new one, the clock resets. A proven idea can look a week old.
- Young ads. An ad launched last Tuesday is not a loser. It is just new.
The fix for most of these is to compare within a brand, not across brands. Sixty days means something different at a brand that rotates ads every week than at one that leaves everything on. Look for the ads that have outlived the brand's own normal.
Reading it with the other signals
Days running gets stronger when other signals agree with it. Two matter most. Variants: how many near-identical versions of an ad are running at once. Relaunches: whether an ad went away and came back.
A hypothetical example. Say a supplement brand has 40 live ads. Thirty-one started in the last fortnight, which tells you it is testing hard. Six have been running for more than 60 days. Two of those six have five or more versions live. Those two ads are where to start: long-lived, and being scaled.
You can run the same check on any brand by hand. Work out days running for every live ad and take the ten oldest. Then check each one for versions and for any sign it has run before. That short list shows you what the brand has kept backing for longest.
That weighting is what our Signal score does. It is a 0 to 100 score built from days running, variants and relaunches, and it is labelled "Estimated, not spend". It will not tell you what an ad earns. It will tell you where to look first, which is most of the battle when a brand has dozens of ads live.
For the ten-minute version of this method, read how to read a competitor's ad library. For the signal most people miss entirely, see Relaunches: the cheapest research in paid social.
