Enhanced conversions in a shop: better attribution, no extra revenue

Enhanced conversions improve the measuring instrument, not the business

Enhanced conversions are a measurement correction. According to Google, they improve the accuracy of conversion measurement and supplement the existing conversion data. Purchases that previously could not be attributed to an ad click show up in the account afterwards. Nothing changes in the shop backend as a result: the orders existed before, they were just not credited to Google Ads.

This gives you the way to read every report after setup. A rise in conversions or ROAS in the following weeks is first of all a measurement effect. As an average, Google cited in 2023 a 5% higher conversion rate in Search for advertisers who adopt enhanced conversions. That is a manufacturer claim about measured conversion rates, not a statement about additional revenue for the retailer.

Where a real performance effect can arise: Smart Bidding gets more observed purchases as a learning signal, and Google says it also uses them to model conversions of users who are not signed in. Whether this turns into more revenue is shown only by the backend, not by the Google Ads report.

How matching with hashing works

The process is simple. On the order confirmation page, the tag reads customer data such as email, name, address or phone number, converts it into a hash value using the one-way hash function SHA256 and sends only that. Google compares this value with the hashed data of signed-in Google accounts. If there is a match and it can be attributed to an ad click, a conversion is counted. According to Google, hash values without a match are deleted.

Two details matter for implementation. Country, region, city and postcode are not hashed, only the fields that directly identify a person. And Google commits to using the data only to provide the service and not to share it with other advertisers.

Hashing is a protective measure for transport, not anonymisation. Hashing the same email address always gives the same value. The matching relies on exactly that, and for exactly that reason the legal question stays open (more on this below).

Why the jump in the report shifts your target ROAS

Target ROAS is a number in the measurement basis of the account. If the measurement basis changes, the same number means something different. A worked example with invented, round values:

Suppose a shop spends €10,000 a month, and the account reports a conversion value of €40,000, so ROAS 4. Target ROAS is set at 4, derived from the margin. After setup, the account shows 10% more conversion value for the same backend revenue.

If the target stays at 4, Smart Bidding has room after the change and bids more aggressively until the measured ROAS is back at 4. In the old measurement basis that is only 3.64. If your break-even ROAS is 3.5, you are suddenly working just above the limit without having deliberately changed a single setting.

The rule: new target ROAS = old target ROAS × (new capture rate ÷ old capture rate). The capture rate is the Google Ads conversion value divided by a fixed comparison value from the shop, such as the revenue of orders with Google Ads origin or total revenue with a stable channel mix. In the example: 4 × 1.1 = 4.4.

The rule does not necessarily apply. If purchases were systematically under-recorded before, Google Ads was in truth more profitable than measured, and the old target was stricter than necessary. It can then be right to leave the target as it is and buy more volume. That is a budget decision you take deliberately, not a side effect of a tracking change. How large and how frequent adjustments should be is covered in changing target ROAS.

How to check the effect against the backend

I would record the capture rate before setup and measure it again afterwards. The process:

1. Four weeks before the change: note the conversion value in Google Ads and the comparison revenue in the shop for the same period, using the same definition (gross or net, with or without cancellations).
2. Set up enhanced conversions and note the date.
3. Four weeks afterwards: run the same calculation. Leave out the last few days, because conversions arrive late.
4. If the capture rate rises clearly, adjust target ROAS accordingly or deliberately leave it.
5. Do not report before-and-after comparisons of ROAS and conversion numbers across the change date as performance development.

If the rate only moves within normal fluctuation, there is nothing to adjust. The basic reconciliation between account and shop system is described in checking conversion data.

Which setup makes sense for a shop

Google names three routes: automatically through the Google tag or Google Tag Manager, manually through the tag with defined fields, or via the API from the server. With the API, the retailer hashes the data themselves. The email address alone is enough, and further fields such as phone number and address raise the match rate according to Google.

My assessment for most shops: the manual route through the tag, with email as a required field and phone and address as far as the checkout collects them anyway. Automatic detection is convenient, but you do not know exactly which field it picks up. The API pays off if you already measure server-side or the shop does not output the data cleanly on the order confirmation page. I consider new required fields in the checkout purely for measurement a mistake, because they can cost more abandonments than the better attribution is worth.

Before the tag collects data, you have to accept the terms for customer-provided data in the account. And the tag only captures the data if the required consent parameters are set to granted. The basics of clean purchase tracking are in conversion tracking in an online shop.

Legal basis: hashing does not answer the GDPR question

First, hashed email addresses remain personal data. Under Recital 26 of the GDPR, pseudonymised data that can be attributed to a person with the use of additional information continues to count as data about an identifiable person. A match that establishes exactly this attribution falls under it.

Second, Google classifies the feature as data processing under the Google Ads Data Processing Terms. You, as the retailer, are the controller. Google leaves it to you to decide whether you need consent or rely on another legal basis. You must disclose in your privacy policy that you pass customer data to third parties for ad measurement and, where legally required, obtain consent.

Third, in practice: a German insurer bases enhanced conversions in its privacy notice on consent under Art. 6(1)(a) GDPR. That is an example, not a requirement.

For the effect in the EEA this means: the consent mode parameter ad_user_data controls whether user data may go to Google for advertising purposes. Enhanced conversions only apply to buyers who have agreed to this. How high that share is depends on your banner, and the choice between the modes is described in Consent Mode v2. I therefore expect a smaller measurement effect in a DACH shop (Germany, Austria and Switzerland) than Google's averages, because only buyers with consent count.

When the effort is worth it

For a shop with Smart Bidding, yes, if three conditions are met: the privacy notice and legal basis are settled, consent through ad_user_data is connected cleanly, and you adjust target ROAS to the new capture rate or decide deliberately against it. Without that last step, you buy more expensive revenue with better data and take it for growth. Status of the feature descriptions: Google's help pages as of November 2026.