The single most common Performance Max complaint we hear from people using Google Ads is the same sentence, almost word for word: “Why is half my budget going to people who already bought from me?” Google’s answer is buried inside the bidding setup as a sub-option called new-customer acquisition. Almost every PMax account we audit either has it switched off entirely or has it dialed too aggressively, and both mistakes leave money on the table. After two years of running this exact configuration across the PMax campaigns in our portfolio, we’ve landed on a recommendation that holds up across e-commerce, lead gen, and local service: “Bid higher for new customers” wins for nearly every business. “Bid only for new customers” is a trap most advertisers should avoid.
Here is what each setting actually does, why one of them outperforms the other, and how we set the numbers when we deploy this for clients.
The PMax problem this setting is designed to solve
Performance Max is built to spend wherever Google thinks a conversion is cheapest. That sounds great until you realize “cheapest conversion” almost always means existing customer. People who have already heard of you, visited your site, or bought from you are vastly easier to convert than strangers, so Google’s algorithm naturally pours budget at them. Within a few weeks of launch, most PMax campaigns settle into a pattern where remarketing-style placements eat 40–70% of spend, and acquisition stalls.
Note: Exceptions for healthcare private practices where they are not allowed to run remarketing on Google Ads due to Google’s policy. PMax’s default then works well as a workaround to use as remarketing. For industries that aren’t allow to remarket, PMax default is a good way to go around that.
This is exactly what most business owners are reacting to when they tell us their PMax campaign “feels stuck.” PMax is just doing the math Google asked it to do, and the math says warm audiences are more efficient and you’ll get cheaper conversions when they show the ad to previous visiting customers.
The new customer acquisition setting is how you tell Google to weight that math differently.
The two settings, plainly explained
Inside any Performance Max campaign, under the bidding section, there is an option called “Customer acquisition.” When you turn it on, Google gives you two choices:
- Bid higher for new customers. Google continues to serve ads to both new and existing customers, but it treats conversions from new customers as if they were worth more. You decide how much more by entering an incremental conversion value.
- Bid only for new customers. Google effectively stops bidding into auctions where the user is on your existing-customer list. The campaign turns into a pure acquisition machine.
The second option sounds appealing the first time you read it. Why not just bid only for new customers? In our experience, that’s the wrong instinct, and Google punishes you hard for choosing it by making cost per conversion extremely expensive in a way that almost kills the campaign.
If your Performance Max campaign feels stuck on the same warm audience, we can audit the bidding setup and fix it before the next billing cycle.
Why “Bid Higher” is frequently the right call
The reason has nothing to do with Google’s interface and everything to do with how warm-audience traffic actually performs. When we look at PMax data across our client base, conversions from existing customers do three things at once:
They convert at much higher rates than new-customer conversions. They cost less per conversion. And they often carry repeat-purchase or referral value that doesn’t show up in the immediate ROAS number.
Telling Google to stop bidding on those auctions doesn’t free up budget for acquisition and it just leaves money on the table. The same dollar that would have driven a $200 reorder from an existing customer now goes into a colder auction where the conversion is statistically less likely to happen at all. The campaign becomes more “acquisition-flavored” in its reporting, but total profit usually drops.
What you actually want is rebalancing, not complete exclusion. “Bid higher for new customers” gives Google permission to keep the cheap, profitable existing-customer conversions while raising the bar for new-customer ones, so the algorithm willingly overpays in colder auctions when it has to. That’s how you grow without giving up the easy revenue.
In the majority of accounts where we’ve migrated a client from “Bid only for new customers” to “Bid higher,” total conversion value has increased within the first 60 days. The split between new and existing customers shifts toward new (which was the goal), but the absolute dollar volume of new-customer conversions also goes up. This is because Google now has the budget room to pursue them harder when warm-audience auctions are unavailable.
How to actually set the incremental value
This is where most accounts get the configuration wrong even after they pick the right option. When you choose “Bid higher for new customers,” Google asks you to enter an incremental conversion value. Google’s asking you: “How much more is a new customer worth to you than an existing one?”.
If you leave it at the default or pick a token number like $5, Google won’t change its behavior much. If you pick a wildly aggressive number, you’ll over-pay in cold auctions and ROAS will tank.
The right number is grounded in your actual customer economics. Here is the calculation we do for our clients after they give us their numbers:
- Segment your existing customers by lifetime value. Pull the top 10% (sometimes top 20% for higher-volume businesses) of your customer base by total spend or repeat order count. These are the customers worth replicating. If you don’t have exact number, estimate the LTV.
- Calculate the spread. Subtract the average value of the bottom 80% from the average value of the top group. That delta is the upper bound of how much more a high-quality new customer is worth to your business.
- Discount for uncertainty. Google can’t perfectly predict which new customers will turn into top-tier ones. We typically set the incremental value at 40–60% of that spread for the first 60 days, then adjust based on data.
Concretely: if your top 10% of customers are worth $400 over their lifetime and your average customer is worth $150, the spread is $250. We’d start with an incremental value of around $120.
Meaning Google will bid as if a new-customer conversion is worth $120 more than an existing-customer one. After a couple of months we look at the actual cohort behavior and either nudge it up or down.
This is also where industry knowledge that Google can’t see makes a real difference. We’ve worked with custom fitting lingerie store clients where customers in certain ZIP codes were 3x more likely to book and show up and average AOV is 2-3x higher. Google has no way to know that on its own. By feeding it segmented customer match lists /customer data from call tracking and letting the new-customer bid amplify those segments, we’ve turned PMax from a one-size-fits-all blast into something that’s actually targeted.
What we’ve seen with our clients
Across our PMax portfolio, the pattern has been consistent enough that we’ve made it a default configuration step in our onboarding process.
We worked with an e-commerce client in the specialty consumables space. They had small basket size, but strong repeat-purchase economics. Their previous agency had switched on “Bid only for new customers” because the founder kept asking why so much budget was going to existing buyers. Within 60 days of that change, total revenue had dropped 18% and the founder didn’t realize the campaign was the cause. When we audited the account, switching back to “Bid higher” with an incremental value calibrated to their LTV spread restored the existing-customer revenue stream and added 22% more new-customer acquisitions on top.
We’ve seen the inverse pattern in lead-gen accounts. A B2B services client running PMax with no customer-acquisition setting at all or conversion tracking wasn’t happy with the leads. The campaign had drifted into chasing the cheapest possible form fills, bot audiences. Turning on “Bid higher for new customers”: the lead count went down but closed deals went up.
For healthcare businesses like Sage Dental implant case study where remarketing isn’t allowed, the calculation is slightly different. We instead layered PMax campaign without bidding preference for new our old customers as an addition to their main Google Ads campaigns. They were able to achieve cheaper cost per click, more clicks, and more conversions at cheaper conversion rate.
If your account has been on autopilot, the customer-acquisition lever is usually the fastest place to find more revenue.
When “Bid Only for New Customers” actually makes sense
There are a small number of legitimate cases for the more aggressive setting, and we want to be honest about them.
The first is a campaign that is genuinely only about acquisition because another channel already covers retention. If you have a separate, well-tuned remarketing or email program handling existing customers, and you’ve built PMax specifically to fill the top of the funnel, “Bid only” can make sense as a way to prevent budget overlap.
The second is a scenario where new customers are worth 10-20x or more over existing customers. Aka in industries where old members do not pay more money (lifetime deals) and do not refer others.
The third is a launch period for a new product line within an established brand, where the existing customer file genuinely isn’t relevant to the new audience. Even then, we’d usually recommend a separate PMax campaign for the new line rather than altering the bidding strategy of the main one.
Outside of those cases, the cost of switching off existing-customer auctions is almost always higher than the apparent benefit of “all my budget goes to acquisition.” It’s the wrong lever for the right problem.
A step-by-step setup we use with clients
When we deploy this configuration on a new account, the order matters. Trying to do step five before step three is how most teams end up with garbage data and a setting that doesn’t move performance:
- Switch the campaign to conversion-value bidding, not just conversions. This is foundational. The new-customer setting only meaningfully changes behavior if Google has differential conversion values to optimize against. We’ve covered the broader reasoning in our Performance Max campaign optimization guide.
- Upload a complete customer match list. Google needs to know who counts as “existing.”
- Segment your customers by lifetime value before turning anything on. Pull the top 10–20% by LTV or by repeat-purchase rate.
- Calculate the incremental value. Use 40–60% of the spread between top-tier and average customer LTV for the first 60 days.
- Turn on “Bid higher for new customers” and enter that incremental value.
- Monitor for 30 days before adjusting. Look at the new-vs-existing split, the absolute new-customer count, and the close rate (for lead gen) or repeat-purchase rate (for e-commerce). Adjust the incremental value up if new-customer volume is too low, down if ROAS is dropping.
The biggest mistake we see at this step is treating the configuration as a one-time setup. PMax learns continuously, and the relationship between bid weights and outcomes drifts as your customer base grows. This is why conversion tracking new data from calls using tools such as WhatConverts, CallRails, or CallTrackingMetrics matters. It adds new customer data back into google so you don’t have to manually import in new customer list every day or end up with data that lags behind.
Frequently asked questions
Do I need a minimum conversion volume before this works?
In our experience, around 30 conversions per month is the practical floor for PMax to use this signal meaningfully. Below that, the algorithm doesn’t have enough data to differentiate new from existing customer auctions reliably, and the bidding setting becomes mostly cosmetic. If you’re under that threshold, focus on getting more conversion volume first, through asset-group cleanup and conversion-value bidding, before fighting with new-customer acquisition.
What if I don’t have a defined customer list yet?
Start building up one, upload your CRM contacts, your Shopify customers, your repeat-booking customer list into a list. Without an uploaded list, the new-customer setting can’t function, because Google has no way to tell who counts as existing.
Should I use “Bid only” temporarily as an experiment?
This is the question we get most. Our short answer: no, not on your main campaign. If you want to test acquisition-only bidding, build a separate PMax campaign that runs alongside the main one with a small budget. Switching the main campaign to “Bid only” is the equivalent of breaking what’s working in order to test something that probably isn’t going to work.
How is this different from regular smart bidding?
Smart bidding handles the auction-by-auction math given your conversion goals. The new-customer acquisition setting modifies the inputs to that math by inflating the perceived value of one conversion type over another. They work together, you can’t use one without the other in PMax.
Where to go from here
If you’re running a Performance Max campaign and you’ve never deliberately configured the customer-acquisition setting, you’re almost certainly losing money to the default behavior. The fix is rarely complicated. It’s usually a one-hour session of customer segmentation, an LTV calculation, and the right incremental value plugged into the right field.
We work through this exact configuration with every client running PMax as part of our Google Ads management service, alongside the asset-group optimization, conversion-tracking validation, and search-term audits that round out a full PMax tune-up. If your account has been on autopilot for more than 90 days, that’s almost always where the next round of growth is hiding.
Let’s look at your customer-acquisition setup together and find the revenue your PMax campaign is leaving on the table.
