Why Adgent
Product All features It reads Account memory Daily verdict Ground truth Trust Gate Creative intelligence Change Ledger It builds Campaign build Media plan Live dashboards Agents & reports Interactive demo
Solutions By team For agencies For in-house By industry Ecommerce & DTC Lead generation Travel & hospitality Marketplaces Local & multi-location Mobile apps & games
Pricing About Blog Get a free audit

Free tool

Is there enough signal in the account to optimize on?

Both platforms publish a minimum conversion volume, and both keep spending happily below it. Put in what the account actually produced and see how far it is from the floor — before you read anything into the cost per result.

No signup · nothing leaves your browser
What the account produced

Count only the ad sets or campaigns optimizing toward the event you care about. Volume on a different event does not help the one being learned.

Share of the published floor

23%

11.3 conversions per ad set per 7 days, against Meta’s stated minimum of 50.

  • Per unit, per window11.3
  • Published floor50
  • Shortfall per unit38.7
  • Units that would clear it if consolidated0

Below the floor. Consolidating will not reach it at this volume, so optimize toward an event higher in the funnel that clears 50 — and treat the current cost per result as unstable rather than as a verdict.

The floors, as published

Both platforms wrote the number down. Neither enforces it.

Every figure here is quoted from the platform’s own documentation, linked in the last column. Nothing in this table is our estimate.

Platform & strategyPublished minimumLevel it applies toSource
Meta — any optimization event50 events / 7 daysper ad setMeta for Developers
Google Search & Shopping — Target ROAS15 / 30 daysconversion-tracking levelGoogle Ads Help
Google Display — Target ROAS15 / 30 daysacross all campaignsGoogle Ads Help
Google Demand Gen — Target ROAS50 / 35 daysper campaign, 10 in the last 7Google Ads Help
Google Video Action — Target ROAS30 / 30 daysper campaignGoogle Ads Help
Google — any Smart Bidding, for reading results30 conversions (50 for Target ROAS)over the period you evaluate — “a month or longer”Google Ads Help

Read the last row carefully: Google frames 30 (and 50 for Target ROAS) as the volume you need to evaluate performance, not as an eligibility gate. That is the honest version — the campaign will run below it, and the number it reports back will move for reasons that have nothing to do with your changes.

What it means in practice

Tidy structure is how accounts fall under the floor.

Nothing warns you. A campaign below the threshold spends exactly like one above it: delivery stays exploratory, the cost per result swings on small numbers, and every weekly report reads as a performance story rather than a sample-size story. The platforms publish the floor and then let you run underneath it indefinitely.

The most common way in is organisation. Split one healthy campaign carrying 90 conversions a month into six tidy product campaigns and you get six units of 15 — all under Google’s evaluation floor, all thrashing, all worse than the one you started with. That mechanic, and the structure that avoids it, is in Google Ads account structure in 2026.

On Meta the same thing happens through scaling: a large budget change re-opens the learning phase, and the ad set needs the 50 events again from a standing start. How to scale a winner without breaking it covers the step size that avoids the reset.

And when the volume genuinely is not there, the useful answer is a refusal rather than a target. Adgent’s Trust Gate is that refusal in the product: it declines to build on an event the account cannot feed, using the account’s own numbers as the reason — a pixel with four purchases in thirty days is running at about one a week against a floor of fifty in seven, and no bid strategy fixes that.

  • Volume is not correctness. Clearing the floor says the optimizer has enough examples to learn from. It says nothing about whether the conversions are counted once, valued correctly, or attributed to the right click.
  • Day-one numbers are not final. Meta’s 7-day click window keeps rewriting yesterday, so a thin account looks worse in the morning than it will by Thursday. Why day-one data reads pessimistic by construction.
  • The floor is per event. Optimizing three ad sets toward three different events splits the signal three ways even when the account total looks healthy. Pick the deepest event that clears the floor, and judge the rest on the business outcome.
Frequently asked

Conversion volume, answered.

How many conversions does Meta need to exit the learning phase?
Meta's own developer documentation says to structure ad sets for a minimum of 50 events over a 7-day period, and that not meeting the minimum events threshold can increase your costs per result. The threshold is per ad set, per week, on the event you optimize for — not per campaign and not per account.
How many conversions does Google Smart Bidding need?
It depends on the strategy and the campaign type. Google states 15 conversions in the past 30 days at the conversion-tracking level as a Target ROAS requirement for Search and Shopping, 50 in the past 35 days for Demand Gen, and 30 in the past 30 days for Video Action campaigns. Separately, Google recommends evaluating performance over a period with at least 30 conversions, and 50 for Target ROAS.
What should I do if the account is below the floor?
Consolidate rather than split: three ad sets at 12 conversions a week each are all below the threshold, while one at 36 is closer to clearing it. The other two moves are optimizing toward an event higher in the funnel that does clear the floor, knowing it carries less value, and fixing whatever is dropping conversions before adding budget.
Does clearing the floor mean the numbers can be trusted?
No. Volume and correctness are different problems. An account can clear every published threshold and still report a ROAS that is wrong because the conversion value is set to gross revenue, the pixel is double-counting, or yesterday's figure has not finished maturing inside the attribution window. This check answers the volume question only.

Method. The check divides the conversions you enter by the number of ad sets or campaigns sharing them and compares the result against the published figure for that platform and strategy — the table above is the whole model. No benchmark data, no estimate, nothing stored: it computes in your browser and the page makes no network request while you use it. Every threshold was re-read at source on and is quoted, not paraphrased. Written by Nurettin Demiral. If a platform changes a number, tell us and we will date the correction on this page.

Free account audit

See which of your ad sets are actually under the floor.

This page works on one number you type. We will run the check across every campaign and ad set on your live accounts — read-only — and show you the ones optimizing on too little to learn from.

  • Read-only — nothing is changed
  • Per ad set and per campaign, not an account average
  • Findings you can check line by line in Ads Manager
What comes back

The brief Adgent would write on day one, the leaks it found and what they are worth — and, where the data cannot carry a verdict, what it will not call yet and why.

We usually reply within a couple of working days.

Thanks — we’ve got it.

We’ll be in touch within a couple of working days.