How Long to See Results with Creative Ads: The Formula

Allow 7 to 14 days to get a first usable signal on a creative, and 2 to 3 months to land a genuinely winning creative. The reason is arithmetic, not magic: only 5 to 7% of tested creatives become winners according to Motion (Creative Benchmarks 2026) and Admiral Media (June 2026). You therefore need about 38 tested creatives to have a 90% chance of finding one. In other words, the question “how long to see results with creative ads” has no answer in weeks: it has an answer in volume divided by your monthly production throughput. According to the SRI and UDECAM digital advertising observatory (l'Observatoire de l'e-pub du SRI et de l'UDECAM), relayed by E-marketing.fr in July 2026, the Social lever of digital advertising in France grew by 16% in the first half of 2026, reaching €2.22 billion.

How long does it really take to see results with creative ads?

There is not one lead time, there are four, and they are constantly confused. When a marketing director asks us “how long before we see something”, they are usually thinking of one thing only: the revenue curve going up. Except that between launching a video and the measurable effect on overall acquisition cost, there are four clocks running at different speeds.

The first clock is technical: the platform learns to deliver. The second is creative: the video appeals or it doesn't. The third is statistical: among your tests, one creative eventually stands out and absorbs the spend. The fourth is financial: your blended CAC moves, or it doesn't.

Mixing up the four is the number one cause of contract breakdowns after six weeks. The client looks at the ROAS of a video that has 40 conversions and concludes that “it doesn't work”. The agency answers “we need to give it time”. Nobody is wrong, everyone is talking about a different clock.

We put this table on the table at the first meeting. Not to save time: to prevent a week-1 CPA from becoming a board-level topic in week 2.

Why is the real unit not time, but the number of creatives tested?

Lead time is a quotient: the number of creatives needed to land a winner, divided by your monthly production throughput. Until you have laid out both terms of the division, no date makes sense.

Let's start with the numerator. In April 2026 Motion published its Creative Benchmarks 2026 report, built on 578,750 creatives, 6,015 advertiser accounts and $1.29 billion of Meta spend. Verdict: around 5% of creatives reach winner status, defined as spend greater than or equal to 10 times the account's median spend and $500. Admiral Media, on more than 10,000 ads analysed (June 2026), gives a range of 6 to 7%. Motion also details a win rate by spend tier, from about 3.8% (Micro) to 8.2% (Enterprise). Opascope puts it at 5 to 7% (July 2026).

The available sources therefore converge on the same order of magnitude. We can work with 6%.

The rest is elementary probability, which we publish here as our own calculation and not as data published by a third party. At a 6% win rate, the probability of having at least one winner is 50% at 12 creatives tested, 80% at 27, and 90% at 38. At 5%, you need 45 creatives to reach 90%. At 7%, 32 are enough.

38 creatives tested for a 90% chance of having at least one winner. At 4 videos a month, that is 9 months. At 12 a month, 3 months. At 20 a month, less than 2 months. The calendar does not depend on your patience, it depends on your cadence.

Foxwell Digital, in July 2026, noted in its reading of the Motion report a point we see every day in accounts: half of ads receive almost no spend, and around 6% concentrate most of the budget. The factor that separates accounts that find winners from those that go round in circles is neither targeting nor bid: it is the intersection of test cadence × volume.

What exactly counts in “38 creatives”

A legitimate question, and a classic trap. Three hook variants on the same edit: is that three creatives or one? Our operational answer: a hook variant on the same video body counts as an iteration, not a concept. A new concept is a new angle, a new format or a new creator. In the 38, we count the units actually delivered and judged separately, what we call a creative iteration, with its precise definition and its purchase unit.

Opascope recommends 5 to 10 concepts per week spun into 2 to 3 variants, with a budget split 80% scale / 20% test. On our French e-commerce accounts, this is sustainable above €20,000 of monthly media. Below that, this cadence is a fantasy: you do not have enough conversions to read 20 tests a week.

One last thing on the numerator, because it is the real subject behind the lead-time question: 38 creatives come at a price. If you want to translate it into euros before translating it into weeks, we have costed the real cost of a winning creative in 2026.

What can you judge after 7 days, and what can't you judge?

At D7, you can judge whether the video captures attention. You cannot judge whether it sells. These are two measurements that do not need the same volume of data.

The 3-second hook rate stabilises quickly: a few thousand impressions are enough to know that a hook doesn't hold up. The hold rate follows the same logic. Outbound CTR needs a bit more room, but remains readable within a week on a decent budget. CPA, on the other hand, needs conversions, and that is where everything changes.

The ad set / creative confusion nobody corrects

The famous threshold of 50 conversion events over a rolling 7-day window to exit Meta's learning phase is counted per ad set, not per creative. If you have 5 ads in one ad set, they share the same pool of 50 conversions (adlibrary.com, 2026). Direct consequence, rarely stated: exiting learning absolutely does not mean you know which of your 5 videos is good. At TikTok, the TikTok for Business help centre (updated April 2026) recommends aiming for more than 20 conversions in the first week for the learning phase of its Search Ads campaigns.

Two clocks, again. The algorithm settles at ad set level. Your creative reading needs volume at ad level. For a creative to be judgeable on a conversion criterion, we internally aim for 25 to 30 conversions on the creative itself, not on the ad set. It is a field threshold, not a published constant: below it, we have seen too many false positives to commit to a verdict.

Which brings us to a firm position: a test that kills an angle in 10 days is a result. In 2025 we stopped presenting performance reviews that only showed the videos that are going up. A “everyday practicality” angle cleanly buried after 12 days is media budget that will not be burnt over the following three months. On a body-care account we have supported since January 2026, four angles out of six were eliminated before the end of the first month. That sorting is what made month 3 profitable.

At 30 days, then, you can demonstrate a lot of things, and not at all what you think. We have detailed elsewhere what a 30-day audit really demonstrates, and above all what it cannot prove.

What minimum media budget is needed for time to matter?

Below a certain daily budget, waiting does not lengthen the lead time: it cancels it. You are not waiting for results, you are waiting for nothing.

The calculation is straightforward. To reach 50 conversions in 7 days in an ad set, you need to spend 50 × your target CPA over the week, divided by 7. Here is what that gives.

These amounts are a floor per ad set, not a campaign budget. This is exactly the media floor below which waiting is pointless, and we would rather say it before signature than at the third monthly review.

Cyberlicious documented as early as August 2024 the three causes of a learning phase that drags on: budget too low, cost cap too restrictive, audience too narrow. Nothing has changed. What we add is that the first cause is by far the most frequent among French SMEs. An advertiser with €2,500 of monthly media spread over three ad sets and eight videos is not doing creative testing. They are playing the lottery with one ticket a month.

Two honest ways out in this case. Either we optimise on an event higher up the funnel, add to cart, initiate checkout, to regain volume and therefore readability. Or we concentrate the entire budget on a single ad set and a single angle, accepting that the lead time moves to six or nine months. What we refuse to do is sell 12 videos a month to someone who does not have the media to read them.

How long does it take even before launching?

Between signature and day 1 of delivery, there are 2 to 4 weeks that almost nobody counts in the timeline. Anglo-Saxon articles on creative testing reason as if the videos already existed. At a French advertiser with legal sign-off and an art director to get validated, they don't exist.

The most expensive line is the legal sign-off. On a food supplement account in 2026, it added thirteen days to the first wave, thirteen days during which media budget kept running on old, fatigued creatives. We now systematically ask for the list of authorised claims before the first shoot. It is settled in a one-hour meeting and saves two weeks. The European Union's Your Europe portal also points out that food businesses operating in the EU may only use authorised health claims if they meet the specific and general requirements that apply to them.

So when an e-commerce director tells us “I have to present results to the executive committee in 6 weeks”: it is feasible, but what they will present are creative verdicts and validated angles. Not a falling blended CAC.

How long before a winning creative wears out?

A winning creative is not a permanent asset: its useful life is counted in weeks, not in quarters. That is what makes the lead-time question cyclical and not linear. Médiamétrie's Observatoire des réseaux sociaux 2026 also notes that the first reason for unfollowing remains content judged uninteresting or repetitive.

Opascope, in July 2026, identifies three measurable fatigue signals: frequency above 3.0, CPM up 20%, CTR down 15%. Their recommendation: rotate the top creatives every 7 to 14 days. On our high-budget accounts we observe the same, with one significant nuance: fatigue depends mainly on the budget / audience size ratio. A creative that holds for six weeks on a broad FR audience can die in nine days on a retargeting pool of 200,000 people.

What this implies for your calendar: the month you find your first winner is not the month you stop producing. It is the month production becomes profitable, and therefore the month you increase throughput, not the reverse. We regularly see advertisers cut production as soon as a video performs. Six weeks later, CPA climbs back up, there is nothing in the pipeline, and the brand starts again from zero with three weeks of production lead time ahead of it.

What we cannot predict

Let's put it in black and white: nobody can say in advance which video will be the winner. Neither can we. Out of thirty concepts, the one the whole team was sure about regularly ends up at the bottom of the ranking, and a product shot filmed in a kitchen captures 40% of the spend. The only thing experience improves is the rate: going from a 4% win rate to 7%, which brings the number of tests needed down from 45 to 32 for the same confidence. That is huge in budget terms. It is not clairvoyance.

What is a realistic 6-month timeline depending on your production throughput?

With media budget throughput held equal, lead time is played out entirely on the number of creatives you put live each month. Here is the calculation translated into scenarios, still on the basis of a 6% win rate.

Read the first line and the last. Same business, same platform, same win rate. A factor of 8 on lead time. That is precisely what you are buying when you buy a monthly throughput: not video files, a date.

The 12 creatives/month scenario, month by month

• Month 0brief, casting, batched shoot of 10 to 12 concepts, editing. Zero data. That is normal, and it is the moment when the client worries the most.
• Month 112 creatives live, learning, first hook rates. Two or three angles eliminated. Overall CPA doesn't move yet, sometimes it degrades slightly: the share of budget in testing pulls the average down.
• Month 224 cumulative creatives, 77% chance of having hit a winner. In practice, a creative often appears that captures 20 to 30% of the spend. We iterate on it: new hooks, new creator, new edit on the same angle.
• Month 336 creatives, the portfolio takes shape: 2 to 4 creatives carry most of the budget. Media CPA starts to hold below target in a repeatable way.
• Months 4 to 6production serves to replace fatigued creatives and open new angles. This is the window where blended CAC becomes an honest indicator, because there is finally enough volume and enough hindsight.

The tipping point we observe most often falls between week 7 and week 11. Not week 4. Any commercial promise of consolidated results at 30 days deserves a simple question: on how many tested creatives?

When should you conclude it won't work?

Beyond 30 tested creatives with no winner at all, the problem is no longer creative: it lies in the offer, the price or the landing page. It is the most useful stopping criterion we have, and also the one that costs us contracts.

The logic is that of the inverted win rate. At 6%, the probability of finding nothing across 30 properly funded tests is about 16%. Across 45 tests, it drops below 7%. Once you get there, continuing to produce videos amounts to paying for an already refuted hypothesis. What is blocking is elsewhere: an average basket too low against the French CPM, a promise the market doesn't believe, a product page that loses 70% of visitors, delivery in 8 days when the competitor delivers in 48 hours.

Admanage noted in July 2026 a documented hit rate of 6.6% and attributed 56% of performance results to creative quality, more than bid or targeting. We agree. But 56% is not 100%. The remaining 44% cannot be fixed with a camera.

Your time-based decision grid

• Cut a creative: 3s hook rate half your account's median after 8,000 to 10,000 impressions. No need to wait 14 days for that.
• Kill an angle: 3 creatives on the same angle, each with sufficient volume, all 40% above target CPA. Typical lead time: 10 to 14 days.
• Iterate: good hook rate, poor outbound CTR, or good CTR and poor conversion rate. The video is right about something, keep it.
• Scale: a creative exceeds 10 times the account's median spend and holds its CPA over 7 consecutive days. Increase in steps of 20 to 30%.
• Question the offer: 30 creatives tested, at least 5 distinct angles, media floor respected, no winner.

That leaves the question asked at the start: how long to see results with creative ads? The only defensible answer fits in a division, not a number. Take 38 creatives, divide by your real monthly throughput, add three weeks of production upstream, and check that each ad set has 50 weekly conversions to make the tests readable. If the result gives you nine months, you have two levers: increase throughput, or reduce ambition. Choosing to “wait” is not one of them.

Frequently asked questions

How long before you can cut a creative ads campaign?

An individual creative is cut from 8,000 to 10,000 impressions if its 3-second hook rate is far below the account median. A whole angle is cut after 3 creatives tested with sufficient volume, i.e. 10 to 14 days. A whole campaign is never cut before checking that the media floor per ad set was respected.

Why does my learning phase stay stuck?

Meta asks for 50 conversion events over a rolling 7-day window per ad set. The three documented causes of a learning phase that drags on are a daily budget that is too low, a cost cap that is too restrictive and an audience that is too narrow (Cyberlicious, August 2024). Among French advertisers, it is the budget in more than eight cases out of ten.

How many videos do you need to test before getting a winning creative?

With a 6% win rate, consistent with the Motion and Admiral Media 2026 benchmarks, you need 12 creatives for a 50% chance of having one, 27 for 80% and 38 for 90%. At a 5% win rate, the 90% threshold rises to 45 creatives. These calculations are ours, applied to the published rates.

How long does ROAS take to stabilise?

On an individual creative, ROAS is not readable before 25 to 30 conversions attributed to that creative, a threshold drawn from our campaigns, not from a publication. At account level, a repeatable ROAS tends to appear between week 7 and week 11, when two to four creatives carry most of the spend.

I have to present results to my executive committee in 6 weeks, is that doable?

Yes, provided you present the right results. In six weeks you will have solid creative verdicts: validated angles, eliminated angles, compared hook rates, CPA by angle. You will not have a demonstrable fall in blended CAC. Announce that scope upfront, otherwise the meeting will be played out on the wrong metric.

Does a small budget simply lengthen the lead time?

No, it cancels it. Below roughly €214/day per ad set for a target CPA of €30, you do not reach the 50 weekly conversions that make a test readable. You accumulate inconclusive data. Better a single properly funded angle than six underfunded angles for six months.