Performance Max Agency London

A Performance Max agency in London manages Google's automated campaign type that places ads across Search, Shopping, Display, YouTube, Gmail, and Maps from a single campaign, using machine learning bidding to decide where your budget converts best. The agency's job is not to fight that automation but to feed it better inputs, asset groups, audience signals, and product feeds, than Google's algorithm gets left to guess with on its own.

A homeware retailer in Shoreditch handed their Google Ads account to Performance Max in early 2026 expecting the same instant lift their competitors talked about online. Three months in, spend was up, conversions were flat, and nobody on the team could say why one week outperformed another. The account had never been given a real audience signal, never had a video asset, and was running on eleven generic product images pulled straight from the supplier catalogue. The algorithm was not broken. It had simply never been given anything worth learning from.

What Performance Max Actually Is, Once You Strip the Marketing Language Away

Every Google Ads rep describes Performance Max the same way: one campaign, every channel, powered by AI. That description is accurate and almost useless, because it tells you nothing about what the campaign needs from you to work.

This is Automation Bias doing its usual damage: the assumption that because a system is automated, it must already be optimised. An automated dishwasher still needs someone to load the plates correctly, scrape off the worst of the food, and choose the right cycle for what is actually inside it. Performance Max is the dishwasher. Your asset groups, audience signals, and product feed are the plates. Run it empty or badly loaded and it will still hum along convincingly, producing a result nobody actually wants.

What Performance Max genuinely automates well is the bidding and placement decision, working out in real time whether a search query, a YouTube viewer, or a Discover feed scroller is more likely to convert right now. What it cannot automate is deciding what your business actually wants to say to each of those audiences. That part is still entirely a human job, and it is the part most accounts get wrong.

Why 72% of Advertisers Have Already Made the Switch

According to Digital Applied (2026), 72% of Google Ads advertisers now run at least one Performance Max campaign, up 34 percentage points since 2024, and Performance Max now drives 45% of all Google Ads conversions. That is not a niche feature anymore. It is the default lens Google increasingly views advertiser accounts through.

Everett Rogers' Diffusion of Innovation curve explains the shape of that number without needing to guess at motive: the early adopters moved first because Google incentivised the shift, then retired Smart Shopping and Smart Display campaigns outright, and the late majority followed because staying on the old campaign types stopped being an option at all. Imagine a shop that used to let you pay by cheque, cash, or card, then one day only accepts card. Most customers switch not because card was obviously better, but because the alternative quietly disappeared.

The practical takeaway for a London business is that Performance Max adoption numbers say more about Google's product roadmap than about how well any individual account is running it. High adoption is not proof of high performance. It is proof that most advertisers no longer have a choice.

Asset Groups: The Inputs That Actually Control the Machine

An asset group is where a Performance Max campaign's actual creative lives: headlines, descriptions, images, logos, and video, bundled together for a specific audience or product theme. Get this thin and the algorithm has nothing distinctive to test. Get it rich and varied and the algorithm has real signal to learn from.

This is the Input-Output principle at work, more bluntly known as Garbage In, Garbage Out. A machine learning system optimises whatever you feed it, faithfully and without judgement. Imagine teaching a child to bake by only ever showing them one photo of one type of cake. They will get very good at recognising that exact cake, and hopeless at anything else you actually needed them to make. A Performance Max asset group with three headlines and one stock photo is that one photo of one cake, endlessly repeated across every channel Google has.

According to Google's own Performance Max guidance, advertisers who adopt Performance Max see an average increase of 27% more conversions or value at a similar cost per action or return on ad spend, but that figure assumes a properly resourced asset group, not the minimum Google technically allows you to submit.

The Video Gap Most London Accounts Leave on the Table

Ask most London business owners what video assets are in their Performance Max campaign and the honest answer is usually none, or one auto-generated clip Google stitched together from product images without anyone reviewing it.

This is the Picture Superiority Effect in action: people remember and respond to moving, visual information far more reliably than static text, and YouTube inventory inside Performance Max simply cannot be reached effectively without it. According to Google's Performance Max guidance, advertisers who included at least one video asset saw an average 12% increase in total conversions, and advertisers who uploaded video in all three orientations, horizontal, vertical, and square, saw 20% more conversions on YouTube specifically compared with horizontal video alone.

Think of it like a shop with three entrances, a wide street door, a narrow side alley, and a small square courtyard gate. If you only ever put up a sign shaped for the street door, the people walking past the alley and the courtyard never see it clearly, even though the shop behind all three doors is exactly the same. Uploading video in every orientation is putting up the right sign at every entrance you actually have.

Feed Optimisation: Why Ecommerce Performance Max Lives or Dies on Data Quality

For an ecommerce business, Performance Max pulls its Shopping placements directly from your Merchant Center product feed, titles, categories, images, availability, and price. A feed with vague titles and missing categories gives the algorithm almost nothing to match against real search intent.

This is the Pareto Principle showing up in feed structure: a small proportion of well-categorised, well-titled, high-margin products usually drive most of the actual revenue Performance Max generates, while a long tail of thin, poorly labelled listings mostly just dilutes the signal. Imagine a market stall where half the fruit has a clear, correct label, apples, £2 a bag, and the other half is unlabelled fruit in an unmarked crate. Buyers reach for the labelled apples every time, not because the unlabelled fruit is worse, but because nobody can tell what it is quickly enough to decide.

Fixing feed titles, adding structured attributes, and prioritising accurate categorisation on the highest-margin SKUs is unglamorous work that rarely shows up in a campaign screenshot, but it is consistently where the biggest performance gains in ecommerce Performance Max accounts come from.

The Sunk Cost Trap: Why Broken PMax Accounts Keep Running Anyway

The most common mistake we see in London accounts is not a technical error. It is a business simply continuing to fund a Performance Max campaign that was never given a fair setup, because the spend is already committed and switching it off feels like admitting the money already spent was wasted.

This is the Sunk Cost Fallacy at its most expensive: the decision to keep spending should depend only on what future spend will return, not on what has already been spent. Imagine buying cinema tickets, realising twenty minutes in that the film is genuinely bad, and staying anyway because you already paid for the ticket. The ticket money is gone either way. The only real choice left is whether to also give up the next ninety minutes.

The businesses that get the most out of Performance Max are the ones willing to pause a badly structured campaign, rebuild the asset groups and audience signals properly, and relaunch, rather than letting an under-resourced campaign quietly drain budget for months because stopping felt like a loss.

When Performance Max Beats Search, and When It Genuinely Does Not

Performance Max is not a universal upgrade over Search campaigns. It is a different tool for a different job, and treating it as a straight replacement is where a lot of London accounts lose control they actually needed.

This is Clayton Christensen's Jobs to Be Done theory applied to campaign selection: the right tool depends entirely on the job the business needs done. A business chasing broad reach across every stage of the buying journey, with a healthy volume of existing conversion data, is well suited to Performance Max. A business with a small number of extremely high-value, tightly defined search terms, where knowing the exact keyword that triggered each sale matters for the sales process itself, often gets more control and clarity running that specific segment through Search instead. According to Dataslayer (2025), most guidance recommends at least 30 conversions in the trailing 30 days before Performance Max has enough signal to optimise reliably, which rules it out entirely for very low-volume accounts in the short term.

The businesses that get this right usually run both, a tightly controlled Search campaign for their highest-intent terms, and Performance Max for the broader reach across channels a manually built campaign could never realistically cover on its own.

What Performance Max Management Actually Costs in London

Media budget and management fees are two separate numbers, and conflating them is where a lot of first-time buyers misjudge what they are actually paying for. According to IAB UK (2025), UK digital ad spend is forecast to reach £45 billion in 2026, growing at around 10% year on year, with Search remaining the largest single channel at 44% of digital spend, £8.3 billion in the first half of 2025 alone. That scale is exactly why agency time spent on asset quality and feed structure, not just campaign setup, is the part worth paying for.

This is Anchoring Bias working against buyers who fixate on the management fee percentage without asking what work that percentage actually buys. A 10% management fee on a well-resourced account with proper asset production and monthly feed audits is a different purchase entirely from the same 10% fee on an account nobody actively touches after launch. Ask what the fee includes, not just what the number is.

Sources and References

1. Digital Applied. "Google Ads Performance Max 2026: Campaign Guide." 2026. digitalapplied.com
2. Google Ads Help. "Multiply conversions with Performance Max." support.google.com
3. Dataslayer. "Google Ads Performance Max: Complete Guide 2025." 2025. dataslayer.ai
4. IAB UK. "Retail Media and TV+ take centre stage as UK ad spend forecast to hit £45 billion by 2026." 2025. iabuk.com