Performance Max Audience Signals: Steer the Algorithm

Performance Max is largely a black box. You supply assets, a budget and a goal, and Google decides for itself where, when and to whom it shows your ads. For B2B advertisers that feels uncomfortable: you hand over much of the control you did have with classic search campaigns. Audience signals are the most important lever you have left to push that algorithm in the right direction. Yet in most accounts they are misunderstood or half filled in. In this article you will learn what audience signals really do, why they are crucial for your performance, and how to build them so Performance Max optimizes for pipeline instead of the cheapest clicks.

What are audience signals exactly?

An audience signal is not an audience on which you enforce hard targeting. That distinction is the whole point, and exactly where most advertisers go wrong. With a classic search campaign or old display targeting you pick an audience and your ad stays within it. With Performance Max it works differently: your signal is a suggestion, a starting point. You tell the algorithm: “These are the people I think will convert, start searching here.” The system then goes looking for profiles that resemble that signal, including outside the group you supplied.

That means your signal speeds up the learning phase, but does not protect you against the wrong traffic. Give it a broad, vague signal and the algorithm starts from a broad, vague point and takes longer to find the right people. Give it a sharp signal of your best customers and the system has a solid starting base right away. For anyone who wants to understand the broader logic of paid search, our pillar on SEA explains how Performance Max fits into the full Google Ads landscape.

Why audience signals make or break your performance

Performance Max optimizes for the goal you give it. Set that goal to “as many conversions as possible at low cost” and feed it weak signals, and the algorithm will faithfully find the cheapest conversions: newsletter sign-ups, light downloads, forms from people who will never become customers. The system does exactly what you ask. The problem is that you are asking it the wrong question.

This is where the difference lies between a campaign that buys clicks and one that buys pipeline. A strong audience signal, combined with the right conversion value, steers the algorithm toward profiles that truly resemble your paying customers. Weak signals plus a shallow conversion goal steer it toward volume without value. You do not notice that difference in your dashboard, where the conversions look fine, but you do in your sales conversations, where the leads do not qualify. That is why we do not treat audience signals as a checkbox but as a strategic decision about who you want to reach.

For a B2B company with a long sales cycle this weighs twice as heavily. You have too few conversions per month to let the algorithm learn purely on data volume. The signal you provide therefore carries relatively much weight. The sharper you make it, the less noise the system has to filter out.

Which signals work, and which to avoid

Not all signals are equal. Google lets you choose from various sources, and the order of quality is fairly clear.

• First-party customer lists. Upload your list of existing customers or qualified leads from your CRM. This is your strongest signal, because they are real people who really bought from you. The algorithm searches for profiles that resemble them. Read in our guide on first-party data how to keep those lists clean.
• Website visitors and converters. People who visited your important pages or filled in a form form a strong behavioral signal. They already showed intent. Segment where possible: visitors to your pricing page say more than visitors to your blog.
• Custom segments based on search behavior. You can build a segment around people who googled specific terms or visited competitor websites. This works reasonably well in B2B because it approximates purchase intent, but it remains an approximation.
• Broad interests and demographics. This is the weakest category. A signal like “interested in marketing, 25-54 years old” is so broad that it gives the algorithm barely any direction. Use it at most as a supplement, never as your main signal.

The practical rule: build your signals from the inside out. Start with who is already a customer, expand to who behaved like a prospect, and only then to broader approaches. If you fill an asset group with only broad interests, you are asking the algorithm to guess from scratch.

That order of quality lines up neatly side by side: the closer a signal sits to your real customers, the more sharply it steers the algorithm.

How to build your audience signals per asset group

In Performance Max, audience signals belong to the asset group, not to the campaign as a whole. That gives you structure: you can create different asset groups for different types of customers, each with its own signal and its own creatives.

1. Give each asset group one clear signal. Do not mix your customer list, your website visitors and three broad interests into one group. You will not know afterward what worked. Build one asset group around your first-party customer data, another around website behavior, and keep them separate so you can compare.
2. Leave the learning phase alone. Every time you change a signal, you partly restart the learning phase. Give a new signal enough time and conversions before you judge. Constantly adjusting feels productive but resets your progress and burns budget.

3. Combine the signal with strong assets that fit that group. An asset group aimed at existing-customer lookalikes deserves different messaging than a group aimed at cold searchers. The signal and the creative must speak to the same person.
4. Exclude the wrong traffic where you can. Use account-level exclusions and, where available, brand exclusions so your budget does not leak toward your own brand name or irrelevant placements. The signal pulls in, the exclusions push away.

Want to embed this setup in a full campaign structure? Read how to build a Google Ads campaign that steers toward the right goals from the start.

The signal is worthless without the right conversion value

Here comes the part most articles skip. A perfect audience signal steers the algorithm toward the right people, but if your conversion goal is weak, the system still optimizes for the wrong result. Audience signal and conversion value go hand in hand.

In B2B this means: feed offline conversion data back to Google. When a lead qualifies in your CRM, receives a quote or closes a deal, send that value back to your campaign via offline conversion tracking. That way Performance Max learns not just “this person filled in a form”, but “this person became a deal of a certain value”. The algorithm then goes looking for profiles that resemble your real customers, not everyone who ever filled in a form.

This is exactly the wedge with which we look at paid search. SEA is for us the fast-acquisition layer of one orchestrated growth engine, and that layer is only worth as much as the pipeline it produces. A Performance Max campaign that optimizes for cheap clicks or a nice ROAS in the dashboard looks healthy and yet delivers nothing. One that optimizes for lead-to-deal attribution buys revenue. The audience signal determines where the algorithm starts; the conversion value determines what it optimizes toward.

For honest advice: not every B2B company has enough conversion volume to let Performance Max learn optimally. If you only have a handful of qualified leads per month, a tighter search campaign can sometimes perform better until your data volume grows. We are a growth partner, not a channel seller, so we first look at whether this format fits your stage. A thorough setup for lead generation through Google Ads always starts with that question.

Frequently asked questions about audience signals

Is an audience signal the same as targeting?

No. With targeting your ad stays within the chosen audience. An audience signal is a starting point: it tells Performance Max who your best customer probably is, after which the algorithm also looks for similar profiles beyond that group. Your signal speeds up learning, but does not protect you against the wrong traffic.

Which signal should I build first?

Start with your first-party data: a customer list from your CRM and your website converters. Those are real people who really bought from you or showed clear intent. Use broad interests and demographics at most as a supplement, never as your main signal.

How long should I let a signal run before I adjust?

Give a new signal enough time and conversions to come out of the learning phase before you judge. Every change partly resets that learning phase. Adjusting frequently feels active but burns budget and slows your results.

Why does my Performance Max deliver many conversions but little revenue?

Almost always because the algorithm optimizes for a weak conversion goal. Feed offline conversion value from your CRM back to the campaign, so the system optimizes for qualified leads and deals instead of cheap forms.

Ready to make your Performance Max steer toward pipeline?

Audience signals are the most important lever you still operate yourself in Performance Max. Fill them with first-party data, link them to real conversion value from your CRM, and you push the algorithm toward profiles that resemble your paying customers instead of the cheapest clicks. If you do not, the black box optimizes for volume without value, and you only see it in your sales conversations.

Want an experienced google ads specialist to review your Performance Max setup and link your audience signals to lead-to-deal attribution? Book your free intake