AI Max vs Performance Max: Which Google Ads Approach Can Help You Acquire More Profitable Customers?

Google Ads is Google's paid advertising platform. Businesses use it to reach prospective customers through Search, YouTube, Shopping and other Google placements, with campaigns designed around outcomes such as enquiries, purchases or revenue. Its AI capabilities can extend the searches a campaign reaches, choose bids, develop assets and learn from the conversion data a business provides.

This Full Court Press series examines AI advertising through a commercial question: how can a business widen automation while retaining informed judgement about the customers, offers and economics it wants to grow? Google is a useful place to start because its AI now influences a connected chain of decisions: which demand a campaign can pursue, who sees an advert, where that person lands and how much Google bids for the opportunity.

AI Max and Performance Max can both expand Google's latitude. Their contribution depends on the outcome the business asks Google to pursue. A company seeking a high-margin recurring customer needs a different campaign design from one seeking a high volume of discounted first purchases.

---

What is the difference between AI Max and Performance Max?

AI Max is a set of AI capabilities inside Google Search campaigns. It can use broad match and keywordless matching to expand beyond the advertiser's written keyword list. It also includes text customisation and Final URL Expansion, which can select a relevant page from the advertiser's website when that control is enabled.

Performance Max is a goal-based campaign type that can run across Google's channels, including Search, YouTube, Display, Discover, Gmail and Maps. It uses conversion goals, creative assets, feed information, search themes and audience signals supplied by the advertiser.

A business can use both when each campaign has a clear role, a defensible conversion definition and a reporting view that considers overlap.

---

What do Google's AI Max examples show?

Google introduced AI Max for Search campaigns on 6 May 2025 as Search was becoming more exploratory and multimodal. Its platform promise has three parts: Search Term Matching can use existing keywords, creative and URLs to find additional demand; Text Customisation and Final URL Expansion can adapt the message and destination; and locations of interest, brand controls, search-term visibility, asset reporting and URL controls give advertisers ways to govern that expansion.

Google's launch article reports that advertisers activating AI Max typically saw 14% more conversions or conversion value at a similar CPA or ROAS. It reports 27% for campaigns receiving more than 70% of conversions or conversion value from exact- or phrase-match keywords. Google identifies the source as 2025 internal data for non-retail advertisers. These aggregate platform figures leave customer margin, retention and incrementality unanswered.

Google's platform illustration: an apparel campaign using the exact-match keyword “red midi dress” could miss a person searching for “colorful midi dresses for spring and summer”. Google uses the example to show how AI Max can find demand outside the written keyword plan, then adapt advertising content and the destination to the query. The commercial question is whether those additional searches produce customers with acceptable acquisition cost, margin and repeat behaviour.

Google-published advertiser cases: Google says L’Oréal reached a new query — “what is the best cream for facial dark spots?” — and recorded a two-times higher conversion rate with a 31% lower cost per conversion. In the same launch article, Google says Australian utility connection service MyConnect recorded 16% more leads at a 13% lower cost per action, with 30% more conversions from new queries. These cases are Google-reported advertiser results. They offer no forecast for another advertiser or proof of profitable incremental customers.

---

What do Google's Performance Max examples show?

Google's 2021 Performance Max launch framed it as one campaign for buying advertising across Search, YouTube, Display, Discover, Gmail and Maps. The business supplies the outcome, conversion values, creative and useful audience signals; Google pursues that goal across eligible inventory and reports on assets, audiences and demand trends. Google describes intended uses including online sales, lead generation and offline sales. Its current lead guidance allows goals such as qualified lead, converted lead, submitted form and phone call, so the team can give Google a more commercially meaningful instruction than a general engagement action.

Google's launch article reported an average 13% increase in total incremental conversions at a similar cost per action, based on Google data from July to September 2021. That is Google's aggregate study result. An advertiser still needs its own design to establish which additional customers the campaign caused and whether those customers were profitable.

Google-published advertiser cases: Google says Allianz Spain and Jellyfish increased qualified car-insurance leads by 15% at a lower cost per lead compared with generic Search campaigns. Google also says Deezer and Artefact increased subscriptions by 28% on the web path at a 15% lower cost per subscription compared with the app path. The named comparators matter: these cases describe reported results in particular campaign settings. They leave customer margin, retention and causal incrementality beyond those comparisons open.

---

Begin with the customer and buying occasion

Consider a beverage company that sells single cans through retailers and mixed-flavour packs through its own website. It has spare production capacity and wants to grow monthly subscriptions. A first-order promotion can attract trial, while a household that stays at the regular price creates a different commercial outcome from a one-off discounted buyer.

Leadership can define the desired customer in terms the campaign can eventually use: someone within the delivery area who chooses a subscription, reaches an acceptable first-order margin and remains active beyond an agreed renewal point. That choice affects the product format, introductory offer, delivery promise, landing-page message and conversion value.

It also affects the search inputs. “Energy drinks delivery”, “monthly drinks subscription”, “mixed-flavour subscription box” and “office beverage subscription” describe different buying occasions. The keywords, search themes and offer each need testing against purchase behaviour, renewal and contribution margin.

---

How keywords, search themes and audience signals work with AI

A keyword is an advertiser's targeting input. A search term is the phrase a person actually enters into Google. AI Max can use broad match and keywordless matching, informed by existing keywords, advertising assets and landing pages, to match demand beyond the written list. This can uncover valuable demand the team had overlooked. It can also attract searches for a different customer or offer.

Performance Max accepts search themes, which describe what customers are likely to search for. Advertisers can also provide audience signals such as customer lists, website visitors and custom segments. Google uses these inputs to guide discovery; the campaign can explore beyond them.

Existing Search advertisers should also inspect configuration changes. Google announced progressive September 2026 upgrades for eligible Search campaigns using standalone automatically created assets or campaign-level broad match. Campaign-level broad-match migrations turn on Search Term Matching while leaving Text Customisation and Final URL Expansion off. Standalone automatically-created-asset migrations turn on Search Term Matching and Text Customisation while leaving Final URL Expansion off. Existing brand inclusions and exclusions carry across. Google's migration announcement sets out the scope.

AI Max includes practical controls: locations of interest, brand inclusions and exclusions, URL inclusions and exclusions, and reporting on broad or keywordless matching, headlines and selected landing pages. These controls turn a broad instruction such as “find subscription customers in our delivery area” into a more accountable operating decision.

---

Can a recurring-product campaign use Performance Max?

Potentially, although Shopping eligibility has specific rules. Google Merchant Center restricts recurring-billing products and services, with defined product and market exceptions. The beverage company's subscription may use campaign elements beyond Shopping, while Shopping eligibility needs confirmation under Google's subscription policy.

This distinction is commercially material. A customer discovering a drink through YouTube may be ready to try a mixed pack. Someone searching for regular delivery may respond to a subscription page. The advert, destination, product configuration and ongoing billing terms should describe the actual offer consistently.

Performance Max gives the advertiser campaign-level choices around daily budget, brand exclusions, placement exclusions and content-suitability filters. It can generate additional text and video assets from the material supplied. Google then allocates budget across eligible inventory according to the campaign's goals and targets. The business remains responsible for product, price, claim and landing-page alignment.

---

What is Google AI actually bidding towards?

Smart Bidding sets bids at auction time against the conversion goals and values selected in the account. The critical management decision is the commercial meaning assigned to a conversion.

Google provides further guidance on value-based bidding, Target CPA and Target ROAS. Actual campaign results vary.

Google changed the visible labels for some target-based strategies from June 2026. Accounts may now show Target CPA and Target ROAS where earlier documentation refers to “Maximise conversions with a Target CPA” or “Maximise conversion value with a Target ROAS”. The bidding behaviour remains the relevant management issue: does the selected goal and assigned value describe the customer outcome worth acquiring?

Google also changed its target-based bidding systems from 17 August 2026 for campaigns limited by budget. Google says Target CPA and Target ROAS campaigns can see temporary traffic and performance movement during the transition, and it introduced a Bid Target Adjustment Tool in July. This is a live-account review point, particularly when leadership compares performance before and after a budget or bidding change. Google's current Target ROAS guidance provides the qualification.

For the beverage company, first-order revenue captures the beginning of customer value. Renewal, cancellation, margin after discount, product cost and delivery all affect sustainable acquisition spending. Google supports estimated conversion values that incorporate repeat business. An estimate needs a defensible basis. A twelve-month value assigned to every new subscriber may encourage acquisition spending that later customer behaviour fails to support.

Management therefore owns the value definition, renewal assumption and review frequency. Google owns the auction-time bid.

---

Can Google bid differently for new, existing and lapsed customers?

Yes, in eligible Search and Performance Max configurations. Google's customer lifecycle goals include New Customer Value, High Value New Customer and New Customer Only options. Performance Max also supports defined retention goals for lapsed customers and loyalty-programme members in compatible configurations.

For the beverage company, this creates a more precise commercial choice than treating every purchase equally. New Customer Value can bid more competitively for a new subscriber while allowing activity among potential returning customers. High Value New Customer can distinguish a high-value prospect where the company has an appropriate customer segment. New Customer Only can reserve budget for acquisition and requires a separate decision about existing customers.

These features rely on accurate purchase information and, in many configurations, maintained Customer Match lists or conversion tagging. A list of customers who once purchased at a heavy discount may be a weak basis for defining the next valuable subscriber.

---

Where human control sits

Google can process auctions, test asset combinations and adjust bids at a scale unavailable to a manual team. Commercial leaders retain responsibility for the growth objective, acquisition economics, offer, customer experience and controls.

Negative keywords help exclude unwanted searches. AI Max respects negative keywords. Performance Max supports them on Search and Shopping, while other placements have their own suitability controls. Search-term reporting gives marketers a partial view of the queries being reached. Google's guidance covers the relevant AI Max controls, Performance Max controls and search-term reporting.

The campaign team also needs to inspect creative and destinations. AI Max can generate text from existing material, and Final URL Expansion can select website pages. Existing tracking templates need testing with dynamic landing pages because incompatible configurations can create broken URLs. A named owner should have authority to correct, pause or narrow activity where customer quality, claims, tracking or fulfilment capacity create concern.

---

How does management read the results?

Search and Performance Max can reach the same searches. An eligible Search exact-match keyword identical to a query generally receives priority, subject to budget and other eligibility conditions. Google's campaign-prioritisation guidance explains the rule.

A combined commercial view should include branded searches, purchases by existing customers and the possibility that Performance Max captured demand previously received by an existing Search campaign.

Google's Data Strength Uplift metric concerns conversions recovered through first-party data setup. It helps a team inspect measurement inputs. Incremental acquisition requires additional evidence.

Leadership can follow comparable groups of newly acquired customers through renewal, margin and fulfilment cost. Holdout or geographic tests can help where feasible. The design should match the spend, available data and decision being made.

Google's channel-performance report provides campaign, channel and format views, including diagnostics for missing assets, product-feed issues, Final URL Expansion and budget constraints. Asset-group and placement reports add delivery evidence. A complete commercial assessment also needs customer, margin and incrementality evidence.

---

How Full Court Press can help

Full Court Press helps commercial teams establish the decision framework before widening AI-led advertising. The work can include:

Commercial Architecture is the Full Court Press term for the structure connecting commercial work from diagnosis through to running revenue: how the operating design is set, what gets built and who carries it once it runs.

• Defining the priority customer, buying occasion, offer and growth objective.
• Examining keywords, search themes, audience inputs, product feeds, landing-page proposition and campaign structure.
• Agreeing conversion definitions and values that reflect customer quality, margin and retention.
• Reviewing AI Max migration settings, Final URL Expansion, exclusions, suitability controls and named ownership.
• Assessing whether lifecycle goals, new-customer value and Customer Match inputs represent the commercial customer definition.
• Establishing the management view across AI Max, Performance Max and existing Search activity.
• Designing a proportionate measurement plan connecting Google Ads, customer data, channel reporting and commercial review.

The intended result is a clearer basis for deciding where Google AI has latitude, what it learns from and how leadership will judge the commercial effect.

---

Sources and method