Is a month a long time to run the same SaaS ad?

A competitor has been running the same image ad for a month. Is that unusually long?

On Meta, it would be well above the median observed run. On Google, it would be well below it.

We analyzed 348,955 SaaS ad creatives in Upbase as of September 14, 2026. Meta image ads had a median observed run of 10 days. Google image ads had a median of 150.6 days, or about five months.

That puts a 30-day-old image ad at three times Meta’s median, but only a fifth of Google’s.

So before reading much into how long an ad has been around, check what you’re comparing it with. The platform makes a substantial difference, and the format matters too.

What counts as a long run?

Here is the breakdown by platform and format. We measured each creative from its platform-reported launch date to its recorded end date, or its most recent sighting if it was still running.

Every Meta format had a median below three weeks. Google image ads were measured in months.

The image comparison is particularly useful because both groups contain substantial samples:208,202 Google creatives and 18,814 Meta creatives. Their median observed runs were 15.1 times apart.

That does not tell us why they differed. Advertisers, placements, objectives, and other factors were not held constant. It does show how misleading it would be to judge their ages against the same benchmark.

The Meta text result deserves less weight. Its eight-day median comes from just 149 creatives, making it a much thinner comparison than the image or video groups. Google’s “Unknown format” category contains ads whose format we could not identify.

Even on the same platform, the format changes the picture

Suppose you’re reviewing a Google video ad that has been running for two months.

Compared with Google image ads, that might look like a relatively short run. Compared with other Google video ads, it is close to the 57.8-day median.

Google image ads had 2.6 times the median observed run of Google video ads. A benchmark that groups both together would hide that difference.

Meta has its own variation. Carousels recorded a median of 19 days, compared with 10 days for image ads. That is nearly double, even though both sit well below Google’s medians.

These differences are worth keeping in mind when reviewing a creative library. A two-week-old Meta image ad and a two-week-old Meta carousel have the same age, but different positions relative to their category medians.

None of this establishes that a carousel performs better than an image, or that switching formats will make an ad last longer. It gives you a more relevant comparison when asking whether a particular creative has had a short or long run.

A monthly competitor check can miss short runs

Imagine a competitor launches an ad on the fifth of the month and stops it on the fifteenth.

You check their active ads on the first of each month. That creative never appears in either snapshot.

With a 10-day median observed run for Meta image ads, this is a gap worth considering in your research. A monthly view of currently active ads may leave out activity that started and ended between visits.

For a competitor watchlist, a weekly review gives you more opportunities to catch those shorter runs. Better still, include ended ads in your research rather than relying entirely on what is live when you happen to look.

This matters when you’re trying to understand changes in messaging. The ad a competitor kept running for months may be easy to find. The offer they tried briefly, then dropped, could be absent from your next active-ad check.

The duration data does not tell us whether that short run was a failed test, a scheduled promotion, or something else. But you need to see the creative before you can investigate it.

The ads that last more than a year deserve a closer look

At the other end of the sample, the 90th percentile was 432.2 days, or roughly 14 months. Approximately one in ten recorded runs had reached that duration or longer.

That is a substantial stretch of time for a single creative. It also puts a year-old ad in perspective: well above the overall median, but still below the cutoff for the longest-running tenth of this sample.

For competitor research, these ads are worth studying alongside the newer ones.

Look at the promise in the headline, the offer, the audience being addressed, and the proof used to support the message. Then compare those elements with the advertiser’s more recent creatives. Are they repeating the same benefit in different ways? Has the offer changed while the central message stayed consistent?

Those are questions the long-running ad can help you investigate. Its age alone cannot answer them.

It also cannot prove profitability. An ad’s recorded duration does not reveal its spend, conversions, or whether it delivered continuously throughout that period. Treat persistence as a reason to look more closely, not as proof that you should copy the creative.

Why the 80-day headline is the least useful planning number

Across the full sample, the median observed run was 79.9 days.

The problem with using that as your benchmark is what sits behind it. Google image ads account for 59.7% of all creatives in the dataset. Nearly six out of every ten observations come from the category with the longest median.

The overall figure accurately describes the middle of this particular sample. It does not describe a typical Meta image ad, and it gives a different impression from the Google video median too.

For your own reporting, keep the comparison specific. A Meta image ad belongs next to other Meta image ads. A Google video ad belongs next to other Google video ads.

Then use age as context when reviewing performance. An ad that has run much longer than its category median deserves a look, but that is not, by itself, a reason to replace it. An ad that stops sooner is not automatically a failure either.

The useful question is what its results tell you to do next.

About the data

This analysis covers 348,955 creatives from 2,395 companies with qualifying dated ad records, observed on September 14, 2026. We included only creatives with a platform-reported launch date within the preceding two years. Records without that date were excluded rather than assigned the date we first discovered them.

For ended ads, duration runs from launch to the recorded end date. For ongoing ads, it runs to the most recent sighting.39.7% of creatives were still running, so the figures combine completed runs with time observed so far. They are not estimates of the median lifespan of completed ads.

Each creative contributes one observation. Ads are attributed to companies rather than individual products, and companies with more qualifying creatives contribute more observations. The results describe this sample at the observation date; they do not establish changes in lifespan over time.