It all depends on the denominator
A conversion rate only means something with its denominator, and each tool picks its own. According to Google's documentation, Google Ads divides conversions by interactions with the ad, which means clicks for a search campaign. Web analytics tools divide them by sessions or by visitors.
The figures then describe the same reality without looking alike. A visitor who comes three times before buying makes one conversion for three sessions, but one conversion for one visitor: the same journey gives 33% or 100%, depending on the tool.
The numerator varies too. A platform set to count every purchase can record several conversions for a single click, and show a rate above 100%. That is a counting setting, not performance.
What a falling rate can hide
Conversion rate is a ratio: it falls when conversions fall, but also when traffic grows faster than they do. A campaign that broadens its delivery brings in visitors who are less ready to buy; the rate drops, while the number of sales can rise. Judging a ratio without looking at volumes leads you to cut what is growing.
The traffic mix matters just as much. A month in which the share of searches for your brand rises shows a better conversion rate without anything having changed on the site: people searching for your name have already made up their minds.
Finally, a sudden overnight drop is rarely commercial. A tracking tag that no longer fires, a modified confirmation page, a misconfigured consent banner all produce exactly this symptom. The first check is technical.
The link between the click and the result
A site's average conversion rate mixes visitors who have nothing in common: loyal customers, curious visitors from a social network, buyers ready to order after a branded search. Comparing it with the rate of an acquisition campaign makes no sense.
The useful reading is by source and by landing page, like for like: the same campaign, the same page, from one period to the next. That is where a gap really means something: a slower page, a less clear offer, an ad promise the page no longer keeps.
It is also the link between the cost of a click and the cost of a result: CPA equals CPC divided by the conversion rate. At an unchanged CPC, going from a 2% to a 3% conversion rate cuts CPA by a third, without touching the bids.
Frequently asked questions
How do you calculate conversion rate?
Divide the number of conversions by the number of visits, visitors, or clicks, then multiply by one hundred. 30 orders from 1,500 visits give a conversion rate of 2%.
What is a good conversion rate?
No average serves as a reference: the rate depends on the sector, the price, the traffic source, and above all the denominator chosen. The only reliable benchmark is your own history, by source and like for like.
Why don't Google Ads and my analytics tool show the same rate?
Because they do not divide by the same thing: Google Ads divides conversions by clicks, the analytics tool by sessions or visitors. They also do not attribute conversions by the same rules, and do not necessarily receive the same number of them.
Is a falling conversion rate always a bad sign?
No. If it falls because traffic is growing faster than conversions, the number of sales can rise. It should be read alongside volumes and CPA, never on its own.