Performance Max Is Hard to Beat for E-commerce. For Lead Gen, Be Careful.

Performance Max, Google’s campaign type that runs across Search, Shopping, YouTube, Gmail and the Display network from a single setup, divides opinion. Some advertisers swear by it. Others call it a black box that spends their money where it likes. Both camps have a point, and the difference usually comes down to two things: what the campaign optimises towards, and what you feed it.

PMax is an AI system. It will find the cheapest possible way to hit the goal you set. Give it a goal like a purchase, and it has to find real buyers. Give it a goal a bot can complete, like a form fill, and it may well find bots.

The e-commerce and lead gen divide

This is the distinction most PMax advice skips, and it should shape your decision before any setup work begins.

A purchase is a clean signal. Someone has to enter card details and spend money, so the algorithm can only hit a revenue or ROAS target by finding real customers. Pointed at e-commerce revenue, PMax learns from signals that are hard to game, and it is difficult to beat.

A lead is a weaker signal. Forms can be filled by bots, click farms and accidental taps, and PMax buys plenty of cheap Display and in-app inventory where invalid traffic thrives. I’ve run PMax for lead generation and watched it produce random, low-quality enquiries and seen an increase in click fraud. The dashboard called it a success, and Google’s account managers push it on every call. The client’s inbox said otherwise. For some clients, I don’t use PMax at all for exactly this reason.

If you run lead gen and still want to try it, close the loop. Import qualified leads from your CRM using offline conversion tracking and bid on those, never on raw form fills. If you can’t do that, standard Search campaigns on Google’s properties give you more control and cleaner traffic.

Be ruthless about what counts as a conversion

This is the single biggest mistake I see. If you tell PMax that newsletter sign-ups, page views and purchases are all “conversions”, the algorithm will chase the easy ones, because when counted like this, ten newsletter sign-ups look better to it than one sale. Set one primary, high-value conversion action: a purchase, a CRM-qualified lead, or a phone call over 60 seconds. Track the softer actions as secondary, but don’t let them drive the bidding. This one decision often separates a profitable PMax campaign from a busy, expensive one.

Your first-party data is your edge

In 2026, the data you own about your customers is the most valuable asset you can bring to Google Ads. Feed it in two ways. Turn on Enhanced Conversions, which passes hashed (scrambled) customer emails back to Google to recover conversions that cookies miss, especially across devices and where tracking is blocked. And upload your customer lists as audience signals, telling the AI what a good customer actually looks like. In my experience, accounts feeding in solid first-party data run at a meaningfully lower cost per acquisition than those leaning on Google’s audience guesses alone. That edge comes from data you already have.

Audience signals guide, they don’t restrict

A common worry is that adding audience signals limits who sees your ads. It doesn’t. PMax treats them as a strong hint about where to start looking, then expands. So give it good hints: your converting customers, your high-value segments, people who’ve bought before. The better the starting signal, the faster it finds more of the right people.

Set bidding targets gently, then tighten

Start with an aggressive target ROAS or target CPA and you’ll restrict the campaign during its learning phase. You could choose to set your initial target around 10 to 20% below your current average, for example, let it stabilise for two to three weeks, then tighten incrementally. Patience pays here. Always consider your business KPIs when you determine how conservative or aggressive you bid.

Creative is now your targeting

This is the mental shift that trips people up. Modern ad systems read your creative, the images, the video, the words, to decide who to show it to. The creative effectively chooses the audience. Thin, generic assets look weak, and they point the targeting at the wrong people. Give each asset group a strong, varied set of headlines, descriptions, high-quality images and, ideally, video. Skip video and Google can auto-generate something basic on your behalf.

Watch where the budget goes

PMax’s fairest criticism is opacity. Use the reporting it does offer, check the search-term, placement and asset-group data, and add negative keywords and brand exclusions where it’s wasting spend on traffic you don’t want or were getting for free anyway.

Watch for the warning signs of invalid traffic too: CTR spikes with no sales behind them, clusters of enquiries with gibberish names, identical messages arriving at odd hours. Google filters some invalid clicks automatically, but it misses plenty, and on lead-gen accounts a third-party click fraud tool could earn its fee. You get less visibility than a Search campaign gives you, and more than the black-box reputation suggests, if you know where to look.

Where to focus your time

The old PPC craft of micro-managing keyword bids alone is largely gone. Your time is well spent on three more things: conversion strategy (teaching it the right goal), audience strategy (feeding it your first-party data), and creative strategy (giving it strong raw material). Those three inputs are where modern paid search is won or lost.

A realistic expectation

PMax won’t rescue a weak offer, a slow website or a confusing landing page. No amount of clever bidding fixes those. What it does, when fed properly and pointed at a goal that’s hard to fake, is find more of the right people more efficiently than manual campaigns usually can. Treat it as a powerful engine that needs quality fuel, never a set-and-forget miracle.

If your Performance Max campaign feels like it’s spending without much to show, check the goal first. Optimising towards leads? Question every enquiry before you trust the numbers. Optimising towards revenue? The fix is almost always upstream: the conversion setup, the data, or the creative. That’s an afternoon’s work to review, and usually the difference between PMax draining budget and driving it.