Using Run Duration to Tell Which Google Ads Are Actually Profitable

In public advertising data there is exactly one field that comes close to "is this ad making money": run duration. Nobody keeps paying, out of pocket, to run a losing search ad for three months straight. This post covers how the number is derived, what each of the four filter tiers answers, and four cases where a long run means the opposite of what you think.

1. What the number is

The "stability" column in the creative list shows the number of days between first seen and last seen, inclusive. It is not impressions, not spend, and not our estimate — it is the difference between two verifiable dates taken from that creative's appearance record in the public ad archive.

The page bands it into five tiers, each with its own color and label:

Above the creative list on an advertiser detail page (reachable from the advertiser directory or search) there are two filters: run duration with tiers of ≥30 days / ≥90 days / ≥180 days / ≥1 year, and last active ≥ which cuts by date. Use them together — either one alone produces wrong conclusions.

2. Why a long run approximates profit

Search advertising is pure paid traffic: no organic tail, no sunk cost to amortize, every click billed in real time. An ad that runs for a year has only two explanations — either its ROI is positive, or someone has been paying for a mistake for a year. The second happens, but not at scale, and not across many creatives in the same account.

That is the whole chain of reasoning, and also its limit: it proves willingness to keep paying, not how much was earned. Any tool claiming to hand you competitor spend or keyword bids is giving you modeled estimates — those fields do not exist in the public archive, for anyone.

3. Four combinations of the two filters

Row one is where attackers should start: long-running and still live means the landing page and the angle behind it have been validated with real money. Row four is the one most people skip — by the time an ad reads "Very Stable" you are a year late; the 30-to-90-day band is where the timing advantage actually is.

Defenders use it inverted. When a brand sees a dozen advertisers pointing at its domain, the instinct is to chase all of them, but most stopped after three days. Filter to ≥ 90 days first; what remains is sustained interception, and "ran for N days" is the single most persuasive line in a notice — see A Brand Bidding Evidence Checklist.

4. Four cases where the field lies to you

1. Evergreen brand-defense ads.
A large brand can run its own brand-term ads for years, because that is defensive spend, not an ROI decision. Tell them apart by landing-domain composition: Booking.com, Stubhub, Inc. and Tripadvisor put essentially everything into 2 owned domains. That is a brand advertising itself, and there is nothing to copy.

2. Agency-managed long tails.
Dealer Spike (1,700+ landing domains), Dealer.com, Incorporated (1,600+) and DealerOn, Inc. (1,600+) buy for hundreds of end clients. Their ads run continuously because the client keeps renewing, not because that individual ad won. Performance across domains inside one such entity varies enormously, so read them per landing domain — that is what the domain filter on the advertiser page is for.

3. Seasonality and geo rotation.
A creative can appear at two points far apart with nothing in between. Duration is a first-to-last difference; it cannot see the gap. The check is to push "last active" forward: a creative labeled 400 days with no activity in the past 30 is more likely seasonal reuse than continuous delivery.

4. Creative rotation on a stable landing page.
Mature buyers replace assets constantly, and every replacement is a new creative ID with the clock reset. So when judging a landing page, do not look at the single longest creative — look at how many creatives keep rolling on that domain over time. Conversely, one creative running unchanged for three years usually means the position is not competitive enough to bother refreshing.

5. A workflow you can copy

Taking a known competitor or agency as the starting point:

1. Open its advertiser detail page and, before filtering, read the landing-domain count in the summary. That decides whether you are looking at a brand advertising itself or an agency, and the two are read completely differently.
2. Set run duration to ≥ 180 days and last active to the last 7 days. What remains is their current core.
3. In the results, watch the landing-domain column and note the domains that repeat. The most repeated ones are their most profitable directions.
4. Open those domain pages and check who else runs the same domain. Only them? It is their exclusive offer. A crowd? It is a public red ocean.
5. Add the advertisers and domains worth following to favorites with a note, and next week look only at what changed.

Steps 3 and 4 expand into a full method in Mapping an Advertiser's Entire Landing-Page Network, which also covers when to stop.

6. How far the free view gets you

Anonymous access returns a sample: the totals are real — the page states how many rows matched and how many are hidden — but the list shows only some of the rows, and there is a freshness lag on top. That is enough to classify an advertiser's shape (domain count, rough stability distribution) but not enough to enumerate every creative that is both ≥180 days and live this week — which is exactly the output this method is for. When you need full enumeration, that is what a subscription is for.

FAQ

How is run duration calculated?
Last seen minus first seen, inclusive of both days. Both dates come from that creative's appearance record in the public ad archive; neither is estimated.

If an ad ran a long time, is it definitely profitable?
Not necessarily. Brand-defense spend, agency long tails renewed by clients, and seasonal reuse all produce long durations unrelated to a single ad's ROI. Classify the advertiser by its landing-domain composition first.

Why do creatives on the same landing page have such different durations?
Because mature buyers rotate assets, and each replacement starts a new creative with a new clock. Judge the domain by how creatives roll over time, not by one creative.

Can I sort by duration?
The page offers tiered filters (≥30 / ≥90 / ≥180 / ≥1 year) used together with the last-active date. Narrow with the filters first, then read the date columns.

Does this field show how much they spent?
No. Duration shows willingness to keep paying; it cannot be converted into an amount. The public archive carries no spend or bid fields.

Landing-domain counts cited here are lower bounds on the current page values; each detail page shows its own data cutoff and is authoritative. Stability tiers verified 2026-08-17.