Google Ads documents {search_term_matchtype} to separate AI Max expansion clicks

Google Ads now documents a ValueTrack parameter called {search_term_matchtype} that marks clicks AI Max matched beyond your keywords. Google's "About ValueTrack parameters" help page (answer 2375447), read on 12 November 2026, has a section titled "Tracking AI Max matches" that now includes the parameter. It tells advertisers who opt into AI Max to combine {matchtype}, {targetid}, {keyword} and {search_term_matchtype} in their tracking URLs so that standard keyword traffic can be separated from AI Max traffic in Google Analytics, a CRM or any other system that reads landing-page URLs.

The help page carries no publication or update date, and Google has not put out a dated announcement about the parameter that this article could find. So the honest statement is narrow: the help page now documents the parameter and the logic for using it. When the parameter started populating on live clicks is not stated. Copies of the page saved by the Internet Archive show the same section listing only three parameters, {matchtype}, {targetid} and {keyword}, as late as 28 October 2026, so the fourth parameter and the filtering rules were added to the page after that date.

What Google's help page now says

The section is short. Its first instruction is that "if you opt into AI Max", you should use the four parameters together, because together they "allow you to segment standard keyword traffic from AI Max keywordless matches." The page then gives a table of four traffic classes and the value each parameter returns for each class. After the table it gives two filtering rules for reporting tools.

Elsewhere on the same page, Google already defined {matchtype} as the match type of the keyword that triggered the ad: "b" for broad, "p" for phrase, "e" for exact, and "a" for AI Max keywordless matches. The page also restates the general rule for all ValueTrack parameters: if the system cannot insert a value, Google Ads replaces the parameter with an empty space.

The four traffic classes

Google's table sets out what each parameter returns for each kind of click. The rows below are copied from the help page, with the {keyword} column summarised.

A note under the table explains the {targetid} values. A keywordless match outputs a value starting with "kwl-". The default automated expansion has the ID kwl-3500001. Advertiser URL rules carry an ID that reflects the specific URL rule criterion.

The two filtering rules

Google gives this logic for anyone who needs to "isolate and analyze your broad search traffic":

• Isolating AI Max expansion traffic: filter for clicks where {matchtype} equals e, p or b, but {search_term_matchtype} equals a.
• Isolating standard keyword traffic: filter for clicks where {matchtype} equals e, p or b and {search_term_matchtype} does not equal a.

Keywordless traffic needs no new parameter to find. A {matchtype} value of "a" already identifies it, and the page says so directly.

Why {search_term_matchtype} matters for AI Max tracking

Read the table one column at a time and the reason for the new parameter is plain. Compare the first two rows, standard keyword match and AI Max broad expansions. {matchtype} returns the keyword's own match type in both. {targetid} returns the same kwd- keyword ID in both. {keyword} returns the matched keyword in both. Without the fourth column, the two rows look identical in a URL.

In practice, that means a click AI Max found by expanding past one of your keywords could be logged against that keyword in an analytics tool, carrying the keyword's ID and match type, with nothing in the URL to show the query came through an AI Max expansion. {search_term_matchtype} is the only column of the four where those two rows differ: e, p or b for a standard match, "a" for the AI Max expansion.

Keywordless traffic was already easier to separate. Because {matchtype} returns "a" for keywordless matches and {targetid} returns a kwl- value instead of a kwd- value, those clicks could be split out without the new parameter. The gap the new parameter closes is the expansion traffic that sits on top of keywords you already bid on.

What this lets an advertiser measure

What URL parameters add is click-level data that leaves Google's interface. A lead form, a CRM record, an order in a store backend or a GA4 session can carry the values with it. Once the AI Max expansion flag is on each click, you can compare the downstream quality of those clicks against standard keyword clicks on measures that may never reach Google Ads: qualified-lead rate, sales accepted by a team, refunds, repeat purchase, or revenue recorded offline.

This is the kind of question AI Max raises and conversion counts alone do not answer. A conversion action that fires on a form submission tells you volume. It does not tell you whether the leads from expansion queries close at the same rate as leads from your own keywords. With {search_term_matchtype} in the URL, that comparison becomes a filter in your own data rather than a guess.

How ValueTrack parameters work

For readers who have not set up ValueTrack before, the same help page explains the mechanics. ValueTrack parameters are URL parameters written inside braces, such as {matchtype}. They can be part of an ad's final URL, its tracking template or a custom parameter. When someone clicks the ad, Google Ads replaces each parameter with a value describing that click.

Google's own example uses a shoe store at www.example.com that wants to know which devices people click from. The store adds the {lpurl} and {device} parameters to its tracking template as {lpurl}?device={device}. A click from a desktop or laptop then lands on www.example.com/?device=c. Google recommends setting tracking templates at the account level.

The page lists several practical notes worth knowing before you edit anything:

• ValueTrack IDs are 64-bit, so any system storing them must not truncate long numbers.
• Changes to tracking templates take 24 to 48 hours to be reflected in ad serving after they are made.
• Parallel tracking is now the only click measurement method for all accounts, and Google advises checking with your click measurement provider that it is compatible.
• ValueTrack is compatible with Search, Display, AI Max for Search, Shopping, Video and App campaigns, with extra requirements for Dynamic Search Ads (the tracking template must include a parameter such as {lpurl}) and App campaigns (a third-party analytics solution that uses Confirmed Installs).

What to check in your own account

If you run AI Max for Search, the help page points to a short set of steps. None of them change bidding or targeting. They only change what your URLs record.

1. Check where tracking lives today. Look at the account, campaign and ad group levels for existing tracking templates, final URL suffixes and custom parameters.
2. Add all four parameters, not just the new one. Google's instruction is to use {matchtype}, {targetid}, {keyword} and {search_term_matchtype} in combination. The new one alone tells you a click was an AI Max match. The other three tell you which keyword, rule or default expansion it came from.
3. Name the URL keys clearly. Google defines the parameter values, not the query-string names you attach them to. Pick names your analytics team will recognise and keep them identical across accounts.
4. Test before relying on it. Google links a "Test your landing page" guide from the ValueTrack page. Confirm the landing page loads with the extra query string and that no redirect strips it.
5. Capture the values where decisions are made. The help page says ValueTrack details can be viewed in Google Analytics or in third-party or site-analytics software, so decide in advance which of those systems will store the four values for each click.
6. Allow for the delay. Because template changes take 24 to 48 hours to reach ad serving, do not judge the data on the first day.

Advertisers who already pass {matchtype} and {keyword} into a CRM are most of the way there. The work is adding one more parameter and one more column, then building the filter Google describes.

What the help page does not say

The documentation is brief, and several things a planner would want to know are absent:

• There is no date for when {search_term_matchtype} started returning values, and no statement on whether older clicks can be reclassified. Treat it as forward-looking from the day your templates update.
• The page does not say whether the parameter behaves the same in a final URL suffix as in a tracking template. It names final URL, tracking template and custom parameter as places where ValueTrack can go in general.
• It does not describe any new reporting column, bidding control or targeting setting tied to the parameter. It is a measurement field.
• It does not define {targetid} or {keyword} outside the AI Max section, so their wider behaviour has to be checked against Google's separate parameter list.
• The opening line of the section talks about separating keyword traffic from "AI Max keywordless matches", while the filtering rule for {search_term_matchtype} is about expansions on existing keywords. Both are in the table, but the wording is worth reading closely before building reports.

What this means for Thai marketers

The help page does not limit the parameter by country or language, and AI Max for Search is listed among the compatible campaign types, so nothing on the page excludes Thai-language campaigns. For Thai lead-generation advertisers who score leads in a CRM or over LINE, the practical gain is being able to tag each lead with whether it came from an AI Max expansion and then compare close rates. Ecommerce advertisers can do the same with order value. The setup is a tracking task rather than a campaign change, which makes it a reasonable item for a Google Ads management review, and the GA4 side fits naturally into a GA4 setup and migration project.

FAQ

What is {search_term_matchtype} in Google Ads?

It is a ValueTrack parameter that returns "a" when a click came from an AI Max match and e, p or b when the search term matched a keyword in the standard way. Google documents it in the "Tracking AI Max matches" section of its About ValueTrack parameters help page, answer 2375447.

How do I separate AI Max expansion clicks from normal keyword clicks?

Filter for clicks where {matchtype} is e, p or b and {search_term_matchtype} is a. Those are AI Max broad expansions on your keywords. Clicks where {matchtype} is e, p or b and {search_term_matchtype} is not a are standard keyword traffic, according to Google's help page.

When did Google launch the parameter?

The source does not give a date. The help page has no publication or update date, and it only documents how the parameter works. Assume it applies to clicks after your tracking templates are updated, and verify with your own test clicks.

Do I have to do anything if I use AI Max?

The help page does not describe the parameter as required for ads to serve, but without it your URLs cannot show which keyword clicks were AI Max expansions. Google's guidance is to add {matchtype}, {targetid}, {keyword} and {search_term_matchtype} together, and template changes take 24 to 48 hours to reach ad serving.

Does it work for campaigns in Thailand?

The help page sets no country or language limit, and it lists AI Max for Search among compatible campaigns. It does not mention Thailand specifically, so confirm with a test click on a Thai-language campaign before relying on the data.

The bottom line

The change is small in code and useful in practice. One more parameter in a tracking template lets an advertiser see, click by click, which traffic AI Max added on top of the keywords they chose, and judge that traffic on business results rather than platform conversions. The documentation is thin on dates and edge cases, so test it in your own account first. If you want help wiring it into GA4 or a CRM, Relevant Audience's paid search team can review your tracking setup.