Enhanced conversions vs server-side tracking

Enhanced conversions are a Google Ads feature that sends hashed customer data with a conversion your tag already reports, so Google can match it to an ad. Google Ads server-side tracking decides something earlier: whether the conversion reaches Google at all.

An order can be lost at either step. We tested Google’s requests in real browsers with six blocker setups to see where extra matching data helps and where the request itself is stopped.

What are enhanced conversions in Google Ads?

Enhanced conversions send hashed first-party customer data to Google with a conversion that your tag already reports. According to Google Ads Help (“About enhanced conversions”), they supplement existing conversion tags with data such as email, name, home address and phone number.

The data is normalised and hashed with SHA-256 before it is sent, and Google says this improves measurement accuracy. With the tag routes, the extra data travels on the existing conversion request. It doesn’t change how the tag loads or where that request goes.

You need the customer fields available at purchase and a conversion tag that already fires. For a visitor whose tag doesn’t fire, there’s no conversion for the extra data to supplement.

How do you set up enhanced conversions?

You turn them on in Google Ads, choose how the customer data reaches Google, and pass at least one customer field. The steps below follow Google’s own guides for the Google tag and Google Tag Manager.

Before you start, Google asks you to have:

• The exact URL where the conversion tag fires, and its conversion event trigger.
• First-party customer data on that page: an email address, full name and home address, and/or a phone number.
• A developer for any code change.
• Confirmation that you can comply with Google’s customer data policies.

Google offers three routes:

Whichever route you choose, the tag collects the data in one of three ways: automatic detection, selectors or variables you point to, or a snippet that sends it directly.

Email is preferred, and an address needs a first name, last name, postal code and country.

Google’s pages differ on one label. The setup guides still name the panel “Enhanced conversions”, while the updates page calls it “Customer data use”. Follow the label your account shows.

How do you check that enhanced conversions are working?

Google documents two checks, and a third shows up once the feature has had time to train.

Google’s DevTools check also lists google.com/pagead/1p-conversion/ for “some browsers”. If you see only “tv.1~em” with no hash after it, Google says the parameter is being sent empty, which can happen when the customer data isn’t available at the time of conversion.

Diagnostics timing depends on the route

Google’s pages give different timings, so check the guidance for the route you used.

What do the diagnostic statuses and match rates mean?

Google grades a set-up as Excellent, Good, Needs attention or No recent data, and labels each match rate Insufficient volume, No matches, Low, High or Unknown.

Google says alerts are not shown when you have fewer than 20 conversions in the past 7 days.

It also says enhanced conversions are most helpful at 20 or more conversions a week. Coverage is the number of enhanced conversion events with the parameter divided by the number of conversion events.

The errors Google lists include accepting the customer data terms (data sent before that isn’t processed), missing or empty fields, and badly formatted email or phone numbers.

They also include missing address fields (country, first name, last name and postal code are all needed) and pages not served over HTTPS. If the report flags a problem, fix the implementation before you judge the results.

What did our ad blocker lab show?

In our lab, Brave Shields, uBlock Origin Lite and AdGuard with Tracking Protection stopped the standard Google Ads conversion requests. They also blocked, redirected or replaced the Google Tag Manager, GA4 and Google Ads tags. We ran each setup in a real browser.

Show the dataHide the data

How many of 4 standard Google requests did each blocker setup stop?

Blocker setup	Requests stopped or replaced (of 4)
Brave Shields	4
uBlock Origin Lite	4
AdGuard + Tracking Protection	4
AdBlock	0

The lab loaded each tag the way a shop does, in real browsers, each in a fresh, throwaway profile on a public origin (example.com).

A request counted as blocked when the browser reported net::ERR_BLOCKED_BY_CLIENT or the blocker swapped the response for a stand-in. The results were the same in Brave and in Chrome for Testing for every extension tested in both.

Not every blocker stops the Google Ads tag. AdGuard on its default filters, AdBlock and AdGuard DNS did not block it. On its default filters, AdGuard blocked only the view-through conversion request; the conversion hit and the tags got through.

AdBlock blocked the Meta Pixel instead.

Show the dataHide the data

In how many of 6 blocker setups did each request get through?

Request tested	Setups where it got through (of 6)
Standard Google Ads tag	3
Full Signal-style first-party endpoint	6

Per setup, the lab recorded this for the Google Ads requests and the Full Signal-style endpoint:

Do blockers stop the request that carries the hashed email?

In three of six setups, yes: the blockers that stopped the standard conversion requests also stopped these two. On 7 October 2026 we sent the two conversion requests named for this check, googleadservices.com/pagead/conversion/ and google.com/pagead/1p-conversion/, with a test hashed email in the em parameter.

Those same three setups blocked both. The other three setups, and a browser with no blocker, let both reach Google.

The em parameter changed no result: each request with em got the same outcome as the same request without it. So if you run Google’s DevTools check in a profile with one of the first three blockers, expect the request not to appear.

The doubleclick.net view-through request with em was blocked in every setup except AdBlock in Brave. The 7 October run added the requests that carry the hashed email, and they followed the standard conversion request in every setup.

Where that request is stopped, the data that rides on it is stopped with it. That is why how the conversion is delivered matters before how it is matched.

Does Safari block the tags?

No. Safari works differently: WebKit tracking prevention restricts cookies instead of blocking loads, so tags still load.

Show the dataHide the data

How many days does a cookie last in Safari?

How the cookie is set	Cookie lifetime in days
Set by a script	7
Full Signal first-party cookie	30

Shoppers often buy days after the click, so a cookie that expires sooner makes a returning visitor more likely to arrive without the data that links them back to the ad.

Our ad blocker conversion tracking test shows Safari as a sixth blocker for that reason.

How does server-side tracking differ from enhanced conversions?

Enhanced conversions improve how Google matches a conversion that reaches it. Server-side tracking decides whether the conversion is sent at all, so each fixes a different gap.

An order can be lost at either step.

Show the dataHide the data

100 shop orders: how many reach Google Ads, and how many are credited to an ad?

Step on the way to Google Ads	Orders still counted (of 100)
Orders in the shop	100
Conversion reaches Google	85
Credited to an ad	75

In this illustrative example, 15 of 100 orders never reach Google. Another 10 arrive but cannot be matched to an ad. Server-side delivery addresses the first gap; enhanced conversions address the second. You can use both.

Do enhanced conversions work with server-side tracking?

Google documents one server route, the Google Ads API, and it supplements a conversion that a tag has already recorded, so it does not replace the tag.

• A tag still records the conversion. Google’s API guide says you must record conversions using a regular conversion tracking tag, then send hashed first-party data over an API connection up to 24 hours after the conversion.
• The server holds the customer data. Google’s developer documentation says the tag sends the click identifier (such as a GCLID) and an order ID at conversion. You import the order ID with the hashed data later, within 24 hours of the original conversion.
• Duplicates and hashing. The API guide says to use a transaction ID so Google can match tag conversions with API conversions, to send over HTTPS, and to normalise and hash the data with SHA-256.

The Google pages we read do not mention server-side tagging containers or ad blockers. Product Metrics’ own route is on the Full Signal Tracking page and in the next section.

To compare Google tag gateway, server-side tagging and Full Signal Tracking, see Google tag gateway vs server-side tagging. Server-side tracking tools for ecommerce sets ten providers side by side.

How does Full Signal Tracking work?

Server-side tracking sends the conversion to Google from a server rather than from the visitor’s browser, so a purchase can still count when something in the browser would have stopped the standard tag. Full Signal Tracking is Product Metrics’ version.

It gets through every adblocker: in our 30 September 2026 lab, a Full Signal-style first-party endpoint was reached in all 6 blocker setups tested (Brave Shields, uBlock Origin Lite, AdGuard with Tracking Protection, AdGuard default filters, AdBlock and AdGuard DNS).

Full Signal Tracking works in four parts. The example codes come from our server-side tracking for Google Ads and Meta page, which also shows the setup per platform:

1. A unique endpoint on your own domain. An opaque path on your shop domain, such as yourshop.com/x7Kq2mVd…, with no “collect”, “track” or “pixel” in it, so adblockers let it through.
2. Auto-rotation keeps it reachable. If a blocker ever catches your endpoint, it rotates to a fresh one on your domain (/x7Kq2mVd… → /Qm7vXt2L…). No tag edits, no tickets.
3. Click IDs kept first-party. The gclid (gclid=Cj0KQ…x7Rw) is kept first-party and server-side, so a purchase days after the click still carries it.
4. Server-side delivery to Google Ads. Purchases go from our server into your conversion actions, with enhanced conversions and Consent Mode v2. Refunds adjust later.

Every purchase then reaches Google Ads and Smart Bidding from your own domain.

To check it yourself, run the ad blocker tracking test in your own browser to see what each blocker stops and whether Full Signal gets through. Paste a request into the SDK debugger to decode it and see whether the collector would accept it.

Full Signal handles both the route and the matching data. It sends hashed customer data with the purchase, so it works with enhanced conversions; the fields depend on the customer data your checkout provides.

First-party is not consent-free. Full Signal reads the visitor’s consent choice before sending anything, as the Full Signal Tracking page describes:

• It respects Consent Mode v2: ad_storage, ad_user_data and ad_personalization.
• It works with Shopify Customer Privacy, Cookiebot, OneTrust, Usercentrics and Osano.
• Customer data is hashed before it leaves our server.

Auto-rotation changes the path, never the consent rules: declined consent stays declined.

Should you set up server-side tracking or enhanced conversions first?

Compare the two counts over the same dates. Missing requests point to a delivery problem. Requests that arrive without enough customer data point to a matching problem.

Show the dataHide the data

Orders in the shop against conversions in Google Ads, week by week

Week	Orders in the shop	Conversions in Google Ads
Week 1	100	85
Week 2	120	102
Week 3	110	94
Week 4	130	110
Week 5	125	106
Week 6	115	98

Not every order comes from an ad, so compare like with like, for example orders that carry a Google click ID. Then work through the steps below, changing one thing at a time.

Before you start, ask your developer which customer fields checkout exposes, whether consent is captured before that data is sent, and which route the conversion takes today.

If the count has a gap, set up server-side tracking first. Full Signal Tracking is included in the Pro subscription and not billed separately; see pricing.

Where the count is close but you pass little customer data to Google, enhanced conversions are the next step, and you can use both.

How much does one missing conversion change a product’s numbers?

For one product in our example, the browser tag alone reported 41 of 55 purchases, which moves its ROAS from 3.85 to 2.87.

Smart Bidding and your product decisions learn from conversions, so products that lose them look worse than they are.

A ROAS decision, or a POAS decision, can be wrong as a result. What is a good ROAS explains why the verdict depends on the product’s own break-even.

The worked example on our Full Signal Tracking page shows how far this can go. One product priced at €139.99 had 55 orders in 30 days from Google Ads clicks.

With the browser tag alone, adblockers and browser privacy kept 14 of them out of Google Ads.

Show the dataHide the data

55 orders from Google Ads clicks: how many reach Google Ads?

Where the orders are counted	Orders counted (of 55)
Orders in your shop from Google Ads clicks	55
What Google Ads received from the browser tag	41
What Google Ads receives with Full Signal Tracking	55

On the same €2,000 of ad spend, the 14 missing purchases change the product’s numbers in Google Ads and its segment in Product Segmentation:

Illustrative data. Profit is €50.91 per order after product cost, 36.4% of the €139.99 price, and POAS is profit divided by ad spend (ROAS times 36.4%).

Both POAS values are above break-even (1), so the product earns a small profit either way. Product Segmentation judges profitability against your target, not against break-even.

With a POAS target of 2, 1.04 is low profitability on a high-potential product, which is a Dog, and 1.40 is medium profitability, which is a Question Mark.

In Google Ads, the 41 purchases the browser tag also saw are worth €5,739.59. The 14 that only Full Signal Tracking sent, blocked in the browser, add €1,959.86, so €7,699.45 is credited to your campaigns.

The order ID is sent as the transaction ID, so Google Ads counts each order once. Google Ads does the attributing.

With the browser tag alone, the product looks like one for lower priority. With every purchase counted, it is a product to test on price and feed instead.

Judge each product against its own break-even, not an account average, whether you measure ROAS or POAS (POAS vs ROAS). Break-even ROAS per product shows the maths.

The break-even ROAS calculator does it for you. Product Segmentation then segments your products with machine learning on profitability (ROAS straight away, POAS once you connect margins), stock, return rate, potential and competitor prices, not just clicks.

Final thoughts

The count is the test: a shop that compares its orders with its conversions knows which problem it has. This week, compare your orders with your conversions, then run the ad blocker test on your own shop.

Once Full Signal is installed, confirm the requests with the SDK debugger. For numbers per product, Product Segmentation shows where the return sits once your conversions are complete, and you can start with the free analysis.

Look at this comparison again after a month; shops, consent setups and blockers change.