Question from Reddit user:
Hi all
Has anyone ever successfully used a conversion action as a negative identifier for Google Ads? Let’s say for example the dreaded job seeker/spammer has clicked one of your Ads, and thus wasted your click/budget (and you know its a job seeker/spammer because you have setup a specific conversion action to identify them), yet you’ve been clever and setup a specific conversion action to track it; is it possible to somehow inform Google Ads that this is in fact a negative conversion and an example of traffic that you don’t want? For example, could an audience be created from these conversion types, and then negatively bid upon? Perhaps there’s some other work around?
Cheers
Ste
Answer from Nabil:
The short answer is:
While you cannot directly flag a conversion action as a “negative conversion” within the Google Ads interface, your proposed idea of creating an audience from users who complete this negative conversion and then applying that audience as an exclusion is the correct and primary workaround for filtering out unwanted traffic like job seekers or spammers.
Furthermore, the most robust and scalable way to prevent these clicks entirely, or to quickly correct the bidding algorithm, is to use server-side data processing via the Google Ads API and a server-side tagging solution to send this negative signal immediately and accurately.
The long answer is:
Your approach of setting up a specific conversion action to identify unwanted traffic is very clever and is exactly the first step toward a solution.
The issue is that Google Ads’ automated bidding systems are optimized to drive conversions, meaning that even a “bad” conversion, like a job application on a lead form, might still be seen by the system as a successful conversion event, leading to continued spend on that user or demographic.
To actively inform Google Ads that this is traffic you do not want, you have two primary methods: Audience Exclusion and Server-Side Automation.
The first method is the Audience Exclusion you suggested.
Once a user triggers your special “negative” conversion, they are added to an audience segment based on that conversion event in Google Analytics 4 (GA4) or Google Ads.
You then navigate to the Audiences section in Google Ads and apply this new audience as an exclusion to your relevant campaigns and ad groups.
This effectively tells Google, “Do not show my ads to anyone in this audience.”
This method is effective but works retrospectively and has a time delay, as the audience needs to build and update.
The second, and more advanced, method involves Server-Side Automation using the Google Ads API and a server-side container hosted on a platform like Stape or Google Cloud Platform.
The workflow here is designed for rapid response and control:
When a user completes your spam conversion, you use Google Tag Manager to capture the Google Click ID (gclid
) associated with that session and send it immediately to your server-side environment.
From there, you use the Google Ads API to perform two key actions.
First, you can use the API to instantly add the associated user ID or customer match data to a Customer List within Google Ads, which you can then set as an exclusion audience, achieving the same result as the manual audience method but often faster.
Second, and this is highly effective for immediate feedback, you can use the API to implement a concept similar to a negative conversion value.
While not a native feature, if you are bidding on conversion value (using a strategy like Target ROAS), you can send an offline conversion with a value of zero or a very small, negative-impacting value back to the Google Ads system for this specific gclid
.
While Google’s algorithm may not treat it as a true negative, sending zero value tells the algorithm this conversion is worth nothing, dampening its enthusiasm for similar traffic much faster than waiting for an audience to populate.
This server-side solution gives you granular, programmatic control over how you signal to Google which clicks are valuable and which are not.