Performance Max 2026: Complete Guide and Strategies for Google Ads • Meteora Web Agency

Are you spending on Google Ads but not seeing results? The problem might not be your budget, but how you distribute it. Performance Max campaigns promise automation and results across all Google channels. But without a strategy, you're burning money. At Meteora Web, we've managed hundreds of PMax campaigns for clients across Italy. Here's what really works.

What Performance Max is and why it's not a magic wand

Performance Max is an objective-based campaign type that automatically combines Search, Shopping, Display, YouTube, Discovery and Gmail into a single campaign. The idea is simple: give Google a goal (conversions, conversion value) and a budget, and Google uses machine learning to find the best channel, creative and audience combinations to achieve it. Sounds magical, but it's not.

We see it often in projects that come to us: a client activates PMax without creative assets, without audience signals, without decent conversion tracking. Google then does what it can: shows text-only ads on Display, spends on generic audiences, and the ROAS is below zero. PMax is an amplifier: if you have a solid foundation (data, assets, monitoring), it works. If not, it only amplifies mistakes.

How PMax machine learning works

Performance Max uses an automated bidding model called “Maximise conversions” or “Maximise conversion value”, with optional CPA target or ROAS target. The model trains on your campaign's historical data and Google's aggregate signals. To work well, it needs:

A minimum conversion volume (at least 30–50 per week per campaign).
Clean conversion tracking without double counting.
Audience signals to guide the algorithm towards the right audience.

Don't expect miracles in the first week: a learning period of at least 2 weeks is required. We recommend not touching the campaign for the first 14 days, except to fix blocking errors.

Initial setup: structure that makes a difference

A PMax should not be a single campaign for your entire business. Split by product, category or goal. Practical example: a fashion e-commerce client we work with created three PMax campaigns: one for the spring/summer collection, one for autumn/winter, and one for accessories. Each campaign with dedicated asset groups per category (e.g. shoes, bags, belts).

Golden rule: never put products with very different margins in the same campaign. Google will optimise for conversion volume, not profit margin. A €10 product and a €200 product in the same PMax? The ML will push the cheaper one because it sells more units, but your profit suffers.

Asset groups: not just text

Each asset group contains headlines, descriptions, images, videos and (for Shopping) feed data. Google combines them automatically to create ads on every channel. For good performance, upload:

5–10 images (at least one vertical, one landscape, one square).
1–3 videos (even short – 15–30 seconds. If you have no video, upload an animated image or a video generated with Canva).
4–5 headlines (max 30 characters each).
5 descriptions (max 90 characters each).

Beware: Google can combine headlines and descriptions in unexpected ways. Always test combinations with the “Ad preview” tool before publishing. We once saw a case where a “Free shipping” headline was paired with an image of a very different product, creating confusion.

Audience signals: the secret to not wasting budget

Performance Max has no explicit targeting like traditional Display campaigns. Instead, you use audience signals to suggest which audience might be more interested. You can include:

Custom intent: based on keywords or URLs your ideal target searches for.
Remarketing: site visitors, cart abandoners.
Your data lists: email lists of past customers uploaded via Customer Match.
Demographic segments (age, gender, income).

Real case: a client of ours, a clothing store in Sicily, had a PMax with only images and text, no audience signal. ROAS was 1.2x. After inserting a list of past customers (exported from their POS) as an audience signal and creating a custom intent based on keywords like “men's summer clothing Sicily” the ROAS jumped to 3.8x in three weeks. Audience signals are not mandatory, but they are the most effective way to prevent Google spending on unqualified audiences.

Budget and bidding strategy

Start with a daily budget that allows at least 10 conversions per week. For bidding strategy, we suggest:

If you're at the beginning: use “Maximise conversions” without CPA target for the first 2–3 weeks so Google learns without constraints.
After 3–4 weeks: switch to “ROAS target” or “CPA target” with a realistic value (e.g. ROAS 4x if margin is 25%).

Monitor cost per conversion daily in the first two weeks. If you see an abnormal CPC (e.g. higher than your product's average price), check diagnostics: there might be a conflict with other campaigns (e.g. a standard Shopping campaign covering the same products).

Tracking and measurement: indispensable

A PMax without reliable conversion tracking is like driving blindfolded. Google optimises based on the data you give it: if the data is dirty, optimisations are wrong.

Set up conversion tracking with Google Tag Manager or directly with the gtag. For e-commerce sites, ensure the “purchase” event includes value and currency parameters, plus item details for dynamic feeds. Here's an example for GTM (custom HTML tag):

<script>
  gtag('event', 'purchase', {
    transaction_id: 'T12345',
    value: 49.99,
    currency: 'EUR',
    items: [{
      id: 'SKU001',
      name: 'Sneakers',
      category: 'Footwear',
      quantity: 1,
      price: 49.99
    }]
  });
</script>

Important: avoid double counting. If you already have GA4 with enhanced measurement, disable duplicate ecommerce event measurement in Google Ads, or use a dedicated Google Ads tag.

Optimization: how to improve performance over time

Performance Max is not “set and forget”. Here are the weekly optimisations we do for our clients:

Asset report: check which images, headlines and descriptions have low performance (rated “Low” or “Learning”). Replace them with better versions.
Exclusion lists: PMax does not natively support negative keywords, but you can create a separate brand campaign to exclude branded queries, or use account-level exclusion lists to block irrelevant Display sites.
Product feed updates: for Shopping, update your feed at least once a day. Use Google Merchant Center with automatic scheduling. Verify prices, availability and images are up-to-date.
Conversion lag: in the “Conversions” report of Google Ads, check the average conversion lag time. If you see a long lag (e.g. 7–14 days), consider extending the attribution window (up to 90 days).
Automated rules: create rules to pause the campaign if ROAS falls below a threshold (e.g. 2x) for 3 consecutive days. This prevents burning budget unnecessarily.

Common mistakes and how to avoid them

1) Not enough learning time. Many clients pause the campaign after 3 days because it's not seeing conversions. Be patient – Google needs to collect data. Don't touch it for 14 days.

2) No brand exclusion. PMax can cannibalise your brand campaigns. Create a separate brand campaign or use an exclusion feed in Merchant Center.

3) Invalid product feeds. A feed with format errors or broken images leads to wasted impressions. Check Merchant Center diagnostics weekly.

4) Ignoring diagnostic reports. Google Ads provides detailed reports on “Asset group”, “Audience”, “Locations”. Use them to understand where you're spending and where you're not.

5) Not testing asset variations. Always upload at least 2 versions of each headline and description. Google will automatically test those with better performance.

What to do now

Here's a checklist to start a solid Performance Max campaign:

Verify conversion tracking – ensure every purchase or lead is tracked correctly with value and currency.
Create at least 3 asset groups – per product, category or goal. Include images, videos, headlines and descriptions.
Set audience signals – insert remarketing, custom intent and previous customer lists.
Set a realistic target – ROAS target based on product margin.
Leave the campaign to learn for 2 weeks – don't modify unless there are critical errors.
Monitor weekly – asset report, conversion lag, cost per conversion.

To dive deeper into the entire Google Ads ecosystem, read our pillar guide: Google Ads and PPC: The Definitive Pillar Guide for Profitable Campaigns.

For a personalised consultation, contact us. At Meteora Web, we work every day with businesses to turn ad spend into real revenue.