Google enhanced conversions send a hashed email address alongside your conversion, so Google can match it against a signed-in account and recover conversions the click identifier alone would have lost. The mechanism is sound and the uplift figure everybody quotes deserves more scepticism than it gets.
The short version: it recovers conversions you were always getting, the recovery is real, and the published percentage comes from marketing rather than from research.
This is for you if somebody has recommended this and quoted you a number to justify it.
What Google enhanced conversions actually do
They give Google a second way to recognise a conversion.
Normally a conversion is tied to a click through an identifier stored in the URL and then in a cookie. If that identifier is lost, and it often is, the conversion happens and Google never connects it to the advert that caused it.
With enhanced conversions, your page also sends a hashed version of the customer’s email or phone. Google compares that against its own signed-in users and matches the conversion that way instead.
The hashing happens in the browser, before anything is transmitted. Google receives a scrambled string rather than the address itself, which is the technical answer to the obvious question.
What this does not do is create conversions. It finds ones that already happened, which is why the honest description is recovery rather than uplift.
Where the uplift number comes from
Google says early adopters saw a median conversion rate increase of 5% on Search and 17% on YouTube.
Read what is absent. No sample size. There is no methodology, no date range, and no definition of what counted as an early adopter. Google’s own disclaimer notes that results vary.
Then notice where the figure lives. It appears in marketing collateral. Google’s official documentation for enhanced conversions for leads contains no uplift figure at all.
That gap between the marketing and the documentation is the useful signal. When a company will put a number in an advert and not in its own help pages, the number is doing a different job in each place.
A second figure that is lower and more useful
There is one first-party observation worth setting beside Google’s claim, and it is more modest.
The agency LGG Media reported that enhanced conversions for leads recover an incremental 8 to 12 percent, against Google’s marketed 10 to 15 percent which they describe as the theoretical ceiling rather than the expectation. They also put baseline click identifier capture at roughly 60% before the change.
That is an agency’s own observation, with no sample size and no methodology, so it carries its own limits. It is useful precisely because it is lower than the vendor figure and comes from somebody with no product to sell you.
Two weak sources pointing the same direction is better evidence than one confident one. Both suggest a real but modest recovery, in the high single digits to low teens, rather than anything transformative.
Why the recovery is worth having anyway
A modest gain on the input side compounds, which is the part the percentage does not convey.
Recovered conversions do two jobs. They correct your reporting, so your cost per lead reflects what you actually got. And they feed the bidding, so the algorithm learns from more of what really happened.
The second matters more than the first. A smart bidding strategy working from 60% of your conversions is making decisions on partial evidence, and every conversion you restore improves the quality of every subsequent decision rather than just the current month’s report.
That is the honest case for Google enhanced conversions. Not a step change, an input improvement that keeps paying.
What it does not fix
It improves matching for the conversions you already count. It has no opinion about whether those are the right conversions.
If your conversion action is a form fill, enhanced conversions will help Google match more form fills, then optimise harder towards people who fill in forms. That is not progress if those people never buy.
The order that matters is unchanged. Decide what you count, then improve how well it is matched, then improve how reliably it arrives. Doing the second and third while the first is wrong is the most common expensive mistake in this whole topic.
Offline conversion import is the tool for the first problem. These two are complementary, and the sequence matters.
The practical blocker is usually the page
Setting this up is rarely a tagging problem. It is a question of whether the data exists where the tag can reach it.
Plenty of confirmation pages show a thank you message and nothing else. The email address the customer just typed is not present on the page, so there is nothing for the tag to hash.
Fixing that means a developer exposing the value at the right moment, which is a small job that has to be scheduled, tested and not broken by the next redesign. On a large account that is obviously worth doing. On a small one, weigh it against what else that developer time could achieve.
Where the data is already on the page, this is close to free and you should just do it.
What Google enhanced conversions change about consent
There is a governance question here that the technical write-ups mostly skip, and somebody in your business needs to have answered it.
Hashing is not anonymisation. A hashed email is a scrambled string that still identifies one person to anybody holding the same address, which is precisely why the matching works at all. Treating it as anonymous because it looks like gibberish is a mistake worth avoiding in writing.
So the question is not whether the transmission is secure. It is whether your privacy notice tells people this happens, and whether you have a lawful basis for it in the markets you operate in.
That is a conversation for whoever owns privacy in your business rather than for your advertising manager. It usually takes one short exchange and it is far better had before switching the feature on than after somebody asks.
None of this is a reason to avoid it. Plenty of businesses run it correctly and disclose it plainly. It is a reason to spend ten minutes on the paperwork rather than discovering the gap during a review.
What to change this week
Three steps.
Check whether your conversion confirmation page has the customer’s email available to the tag. That single question decides whether this is a ten minute job or a development ticket.
Then, if you already have Google enhanced conversions running, find out what share of your conversions are being matched through them. That figure is more useful to you than any published uplift claim.
Then stop quoting the 5% to anybody, or quote it with its limits attached, because repeating an undocumented vendor figure as fact is how this industry builds its worst habits.
What conversion numbers actually measure covers the counting underneath, sending closed-won data back is the fix for the more important problem, and platform reported conversions explains why recovery changes your reporting without changing your results. Getting the email onto the confirmation page is usually web development work. If you want to know what your account is currently matching, book a teardown.