A long-running ad is a profitable ad

No platform publishes ad spend, so every competitor budget figure you meet is speculation. The best substitute is how long an ad has been alive: nobody pays monthly for an ad that does not return. And that figure is public.

Why the inference works

Paid ads get switched off quickly when they do not perform — that is how everybody runs ads. An ad still live after three months has survived several budget reviews.

It follows that in a brand's list of ads, the longest-lived one says what the market wants to hear. That is worth more than knowing how much they spend.

How to read it

Multi-platform ad radar — 20 credits, 60 ads, runs in the background — returns ads running in Vietnam on Facebook/Instagram, Google and LinkedIn, longest-running first, with variants of the same creative merged.

Read in three tiers:

Tier 1 — alive over 90 days. This is their core message. If you read only one thing, read this.

Tier 2 — 30 to 90 days. Working, but not yet proven. Usually variants of tier 1.

Tier 3 — under 30 days. Tests. Read them to see which direction they are trying, but conclude nothing about performance.

The number of variants of one creative is also a signal: many variants means they are testing that direction seriously rather than posting for the sake of it.

Four cases where the inference fails

Brand advertising, not performance advertising. Large companies run long-term awareness ads without measuring on orders.

Forgotten ads. Real, especially in large accounts with many campaigns.

A seasonal ad coming back. A Tet ad running each December looks "new" but has actually been validated over years.

The library only holds a data window. Google Ads Transparency Center has a data window of roughly 30 days, so "last shown" there does not mean the ad just started (ad library comparison, accessed 21/09/2026).

That last case is why you should read Meta and Google differently: Meta gives a start date, Google gives a last-shown date.

Something to do right away

Run the radar for three competitors, pull out the ads alive over 90 days, and write on one page: what problem they name and what result they promise. If all three name the same problem, that is the problem the market cares about most — and if you are saying something else, that is worth revisiting.

See also see the ads a competitor is running.

Frequently asked questions

How many days counts as long-running? There is no standard threshold. The three tiers above are a usable split; adjust for your industry.

Can I infer budget from run time? No. Run time speaks to effectiveness, not to money.

Does the radar cover TikTok? No. See the reason in the TikTok ad library.