Performance Max for Ontario E-commerce: 2026 Setup Guide

Performance Max is now where most Ontario e-commerce brands are spending the majority of their Google Ads budget. It runs on AI across Search, Shopping, YouTube, Display, Gmail, and Discover, and when it’s set up well it outperforms a human-managed Shopping campaign by 15–30%. When it’s set up poorly it quietly burns budget on branded traffic you’d have gotten for free.

This post walks through the structure that actually works in 2026 — how to segment campaigns, build asset groups, feed the product feed, and (critically) tell Google where not to spend.

Why pMax matters for Ontario e-commerce

Three structural reasons:

• AI bidding uses signals traditional Shopping campaigns can’t touch. User intent, device history, time-of-day, lookalike patterns. The bidding model is genuinely better than what a human can manage at the SKU level.
• It captures Search + Shopping + YouTube + Gmail in one campaign. For brands with limited budget, this is more efficient than running 4 separate campaigns, each with its own learning period.
• Ontario’s competitive verticals (DTC beauty, home goods, boutique retail) reward early adopters. Brands still using only standard Shopping are giving up 2–4x points of ROAS that pMax can capture.

Campaign structure

The biggest mistake we see is one giant pMax campaign for the entire store. Google tolerates it but doesn’t optimize well across heterogeneous products.

Recommended structure by catalogue size

Small catalogue (under 100 SKUs):

• 1 pMax campaign for the whole store
• Divided into 2–3 asset groups by category

Medium catalogue (100–500 SKUs):

• 1 pMax campaign per margin tier or category cluster
• Example: “pMax — High Margin Core,” “pMax — Seasonal,” “pMax — Clearance”

Large catalogue (500+ SKUs):

• 1 pMax campaign per product category that has meaningfully different target ROAS
• Example: “pMax — Skincare,” “pMax — Haircare,” “pMax — Accessories”

Segmentation lets Google optimize each campaign against its own target ROAS. Lumping everything together forces the algorithm to average — which is almost always worse than segmentation.

Asset groups

Each pMax campaign contains 1–5 asset groups. Think of an asset group as a mini-campaign focused on one product line or audience segment.

What each asset group needs:

• 5–15 headlines (30 chars each). Mix product attributes, benefits, and audience cues.
• 5 long headlines (90 chars each). Longer-form product hooks.
• 5 descriptions (90 chars). Clear, specific, not generic “shop now.”
• 15+ images. Hero product shots, lifestyle shots, social-proof graphics, context-of-use shots. Square + landscape + portrait aspect ratios.
• 5+ videos. If you don’t have video, Google auto-generates one from your images. The auto-generated version is serviceable; custom video typically performs 20–40% better.
• 1 logo — clean, on a transparent or white background.
• 1 product feed (via Google Merchant Center).
• Audience signals — your customer list, high-intent site visitors, and optional in-market audiences.

Rule of thumb: more assets = more variations Google can test. But quality matters — don’t pad the count with duplicates or obviously bad shots. Google’s reporting will tell you which asset strength rating each field receives (Poor / Average / Good / Best). Aim for Good or Best across the board.

The product feed (Google Merchant Center)

The feed is usually where pMax performance is won or lost.

Non-negotiables:

• Title optimization. Titles should be structured: [Brand] [Product name] [Key attribute] [Size/colour] [Gender/category]. Example: Alder & Co. Organic Cotton Tee — Navy — Men's — Medium. Keyword density matters here like it used to in SEO.
• GTIN + MPN on every product that has one. Google uses these to match your products to price comparison and Shopping Actions.
• High-quality images. At least 800×800px, clean backgrounds, zero text overlays.
• Complete attribute coverage. Colour, size, material, gender, age group — every Merchant Center-recognized attribute you can fill. Partial feeds get throttled.
• Availability updates. If Merchant Center sees “in stock” on a product that’s actually sold out, you get disapprovals. Sync daily at minimum; real-time inventory sync (via the Content API or your Shopify app) is much better.

Common feed problems we see:

• Titles that are just product names without attributes.
• Product types that don’t map to Google’s taxonomy.
• Missing GTINs on branded products (Google deprioritizes these).
• Disapproved products nobody caught because the disapproval email went to spam.

Exclusions: where pMax loses money

This is where pMax campaigns are silently hemorrhaging budget in 80% of accounts we audit.

1. Brand exclusions

By default, pMax will bid on your own brand name. On branded terms you’d rank #1 organically anyway, Google will happily charge you $2-5/click to send traffic to your site.

Fix: add negative keywords at the account level for your brand terms. In 2026, pMax supports campaign-level negative keywords directly (previously required opening a Google Ads support ticket).

Sometimes you do want pMax bidding on brand — specifically to block competitors who are bidding on your brand. But if no one else is bidding, save the budget.

2. Competitor brand exclusions (if applicable)

If you sell other brands in your store, be careful about bidding on competitor brand names that you don’t carry. Trademark disputes are real; pMax’s auto-expansion of search terms can push you into them.

3. Poor-performing products

If 20% of your SKUs drive 80% of revenue, the remaining 80% of SKUs are dragging ROAS down. Use custom labels in the feed to segment performers from non-performers, then exclude low-performers from pMax campaigns (or move them to a dedicated “clearance” campaign with lower target ROAS).

4. Placements (partial control only)

pMax gives you limited control over placements. You can exclude specific YouTube channels or placement URLs by adding them to your account-level exclusion list, but you can’t exclude entire networks (Display, YouTube) while keeping Shopping and Search.

Target ROAS and learning period

Set your target ROAS at 10–20% above the breakeven point. Too low and you’ll burn budget. Too high and the algorithm starves for data.

Don’t touch it for at least 4 weeks. pMax needs a learning period of 50+ conversions per campaign. Adjusting target ROAS during learning resets it.

After 4 weeks:

• If actual ROAS exceeds target by 25%+ consistently, you’re leaving volume on the table. Lower target ROAS to drive more spend.
• If actual ROAS falls short of target consistently, raise target ROAS by 10–15% and let it re-stabilize.

Weekly monitoring routine

• Check campaign-level ROAS and spend.
• Review the Insights tab for emerging search themes and audience segments.
• Check Merchant Center for new disapprovals. Fix same-day.
• Review asset strength ratings. Swap out any “Poor” rated headlines or images.
• Check search term report (now available in pMax since late 2024). Add negatives where needed.

Where this fits

pMax is one of three Google Ads tactics we cover in the Google Ads for Ontario Businesses 2026 Playbook. If your conversion tracking isn’t solid, pMax will optimize toward the wrong outcome — so start with the conversion tracking setup post. For the tactical mistakes that often accompany pMax misconfiguration, see 5 Google Ads Mistakes Burning Your Budget.

If you want us to audit or rebuild your pMax campaigns, that’s part of our Google Ads engagement. For DTC brands, we typically pair pMax management with a Klaviyo buildout via our Klaviyo email service — ads drive the first purchase, email drives the second through tenth.