Every or One: The Conversion Counting Setting That Inflates Your Lead Numbers

Every conversion action in Google Ads has a Count setting with two options: Every, or One. It sits in a part of the interface most people configure once and never revisit, and it quietly decides whether a single person filling in your form five times appears in your reporting as five leads or as one. For a lead generation account, picking the wrong one is the difference between a number you can defend in a client meeting and a number that falls apart the moment anyone opens the CRM.

This is not a broken tag. Nothing is misfiring, nothing is misconfigured in your tag manager, and a tracking audit that only looks at whether tags fire will pass this account without comment. The tag is doing exactly what it was told. The question is what you told it to count.

What the setting actually controls

When somebody clicks your ad and then converts, Google attributes that conversion to the click. The Count setting decides what happens if the same person converts again, still attributed to that same original click.

• Every counts all of them. One click leading to four purchases records four conversions.
• One counts the first and ignores the rest. One click leading to four form submissions records one conversion.

Google’s broad guidance is that sales and ecommerce transactions generally want Every, because a customer buying twice genuinely is worth twice as much, while leads generally want One, because the same person submitting the same enquiry form three times is still one lead. The default depends on the conversion source, so the only reliable move is to open each conversion action and read what it currently says rather than assuming.

What goes wrong in lead generation

The expensive direction is a lead form conversion action set to Every. Here is why it inflates so reliably, and why the inflation is invisible.

People submit forms more than once

Not occasionally. Constantly. Somebody fills in a contact form, is not sure it worked because the confirmation was subtle, and submits it again. Somebody enquires about one service on Monday and a different one on Thursday. Somebody fills in the form, gets no reply within a day, and tries again. Every one of those is one human being with one requirement, and with Every selected, every one of them is a separate conversion in your reporting.

Thank you pages get revisited

If your conversion fires on a thank you page load, then every time that URL loads again it fires again. People refresh. People hit back and then forward. People bookmark the page because it has a reference number on it. Browsers restore tabs. With Every selected, each of those reloads is another conversion credited to the original ad click.

Multi-step forms fire more than once

A form split across several steps, each with its own URL or its own event, can fire a conversion at more than one stage if the trigger was written loosely. One genuine enquiry becomes three recorded conversions, and the pattern is consistent enough that it looks like real volume rather than a bug.

The part that costs money

Inflated reporting is embarrassing. What actually costs money is what the inflation teaches the bidding algorithm.

Smart Bidding optimises toward recorded conversions. If duplicate submissions are being recorded as separate conversions, then the audiences, placements, keywords and times of day that produce duplicate submitters look like your best performing segments. They are not better. They are simply more likely to submit twice, which often correlates with confusion, uncertainty, or a form experience that leaves people unsure whether it worked.

So the algorithm spends more to find more people like them. Your cost per conversion improves on paper while your cost per actual customer gets worse, and the gap between the two widens for as long as the setting stays wrong. This is the same shape of failure as a conversion tag on an All Pages trigger, just smaller and far harder to notice, because the inflated number is still broadly proportional to real demand.

The opposite mistake

Ecommerce accounts sometimes have the reverse problem: a purchase conversion action set to One. Now a customer who buys thrice in a month after a single ad click counts once, and your reported revenue from that click is a third of the truth.

This direction is less common and usually less damaging, but it has a nasty side effect. It systematically undervalues exactly the campaigns that bring in repeat buyers, which are usually your most valuable campaigns. You end up cutting budget on the thing that was working because its reported return looks mediocre.

How to tell whether yours is wrong

1. Open the setting and read it. Go through your conversion actions one at a time and check what Count says on each. This takes a few minutes and resolves the question directly. Do it per action, because a single account commonly contains some that should be Every and some that should be One.
2. Look at the repeat rate. Google Ads reports a repeat rate metric that shows how many conversions are coming per converting click. For a lead form, a repeat rate meaningfully above one is telling you that the same clicks are converting over and over, and that is the inflation measured for you in your own account.
3. Compare against the CRM. This is the check that settles arguments. Count the genuine, de-duplicated enquiries your client received last month and compare that to the conversions Google Ads reported. If Google says 180 and the CRM contains 120 real people, you have found roughly where the difference lives.
4. Watch for conversions exceeding clicks. If a campaign ever reports more conversions than it had clicks, counting is almost always part of the explanation.

Fixing it properly

Changing the dropdown is the immediate fix, and it applies going forward rather than restating history. Expect your reported conversion volume to drop the day you change it, and expect that to be uncomfortable if nobody has been warned. Tell the client first, in advance, with the explanation. A conversion count that falls by a third without warning is a difficult conversation. The same drop, pre-announced as a correction with the CRM comparison attached, is a credibility win.

Also expect a bidding adjustment period. The algorithm has been learning from the inflated signal, and it needs time with the corrected one. Resist changing targets in the same week, because you will not be able to tell which change caused what.

The better fix is upstream

The Count setting controls how Google treats repeats. It does not stop your site generating them. Where you can, prevent the duplicates at the source:

• Fire conversions on a confirmed submission event rather than on a thank you page load, so a refresh cannot re-trigger it.
• Pass a unique transaction or order identifier with ecommerce conversions, so the platform can recognise a repeated report of the same event.
• Make sure multi-step forms fire exactly once, at the final step, and verify it rather than assuming it.
• Give people an unmistakable confirmation. A visible, obvious success message removes the most common reason anyone submits a form twice.

Counting is not attribution, and the difference matters

A common misreading is that One conversion means Google will only ever count one conversion per person. It does not. The Count setting operates per click, not per human being.

If somebody clicks your ad on Monday and submits the form twice that afternoon, One records a single conversion. If the same person clicks a different ad of yours on Thursday and submits again, that is a new click, and it gets its own conversion. Same person, two conversions, with the setting working exactly as designed.

This matters for two reasons. It explains why switching to One does not collapse your numbers as far as people sometimes fear, which is useful when you are preparing a client for the change. And it means a genuinely repeat enquirer who keeps coming back through your ads will still be counted more than once, which is usually the behaviour you want, because each of those clicks cost you money and each one produced something.

The conversion window interacts here too. A conversion only counts if it happens inside the window configured on that action, so a long sales cycle with a short window will undercount regardless of how counting is set. If your numbers look low and counting is already correct, the window is the next thing to check.

Why this hides so well

Every diagnostic you would normally run comes back clean. The tag fires. The conversion action is active and recording. Tag Assistant is happy. The numbers are not absurd, just somewhat better than reality, and better numbers do not generate support tickets. The only thing that exposes it is comparing the ad platform against a source of truth that counts humans instead of events, and most accounts never run that comparison.

If you want to check the mechanical side of a site, our free tracking scan reads the public source and the published tag manager container in about thirty seconds and tells you what fires on what, which is where duplicate-prone triggers show up. It will not tell you what your Count setting says, because that lives inside your Google Ads account rather than on the website, and the account check above is how you confirm that half.

Settings also drift. Conversion actions get rebuilt, duplicated, and recreated by whoever is in the account that month, and a new action arrives with whatever default it arrives with. Taglert rechecks every connected account every thirty minutes and emails you when conversion tracking changes, so a new action configured the wrong way does not spend six months inflating a client report.