Why Your Google Ads ROAS Is Dropping in 2026
Google now decides where roughly 80% of a typical Performance Max budget goes before a human ever touches it. The campaign type that powers most accounts in 2026 hands almost every meaningful allocation decision to an algorithm, and that single shift explains more falling ROAS figures than rising CPCs ever will. The advertisers losing ground aren’t being outbid. They’re being out-structured by their own automation.
If your Google Ads returns looked healthier a year ago, you aren’t imagining it. CPCs are higher, Performance Max is quietly absorbing budgets you can’t fully account for, and the account structures that printed money in 2023 and 2024 are coming apart in 2026.
We manage paid budgets across B2B and DTC brands at Social Schnell, and the pattern landing on our desk right now is one of the sharpest performance shifts we’ve seen in years. Same products, same audiences, same offers, and a return on ad spend that’s drifting down month after month for reasons the account owner can’t put a finger on.
Here’s the part that should give you some relief: it’s fixable, and usually faster than people expect. But not by doing more of what stopped working. The accounts recovering are the ones that understood what actually changed underneath them and rebuilt to match it. So let’s start with the obvious question.
What Is ROAS, and Why Is Yours Falling?
ROAS is return on ad spend, the revenue you earn for every pound you put into Google Ads. Spend £1,000, bring back £4,000, and your ROAS is 4x. Simple maths, but it’s the number your entire paid channel lives or dies by, because it ties advertising cost directly to revenue.
When ROAS drops, only two things can happen. Either you’re spending more to earn the same revenue, or revenue is sliding while spend holds. In 2026, across a large share of accounts, both are happening at once, and the causes are structural rather than down to a bad month or a louder competitor.
That distinction matters more than anything else in this article. If the cause were purely market-driven, the fix would be to spend less and wait it out. It isn’t. The fix is to change how your account feeds Google’s machine learning, and that’s entirely within your control.
Why Google Ads ROAS Is Dropping in 2026: The Real Causes
The reflex when ROAS falls is to blame rising CPCs and more aggressive bidders. Both are real, and both are happening. But they’re symptoms, not the disease. The root causes are sitting inside your account structure.
Performance Max changed how your budget gets spent
Performance Max is now the default campaign type across most verticals, and that default is not neutral. PMax pools your budget across Search, Shopping, Display, YouTube, Discover, and Gmail at once, then the algorithm decides where it lands. Give it no segmentation and no audience signals, and it drifts toward the cheapest inventory available, which is almost always Display and Discover. High volume, low intent, weak conversion rates.
Your budget looks fully spent. Your ROAS quietly collapses. The traffic doing the conversion isn’t the traffic you’d have chosen if Google had asked, which is precisely why Performance Max optimization is the single biggest lever in most paid marketing accounts this year.
Broad match is running wider than it ever has
Google has pushed broad matches hard over the last 18 months, and paired with Smart Bidding it’s framed as letting the machine find your best customers. In practice, your search term reports fill up with loosely related queries, your conversion data gets diluted, and your Target ROAS strategy slowly loses its calibration because it’s learning from noise.
We’ve audited accounts where a broad match was serving against queries so far from the original intent that the gap between click and conversion was roughly three times wider than it should have been. The algorithm wasn’t broken. It was being fed garbage and optimising faithfully toward it.
Your first-party data isn’t reaching the account
The move away from third-party cookies is fully operational now, not a slide in a 2022 conference deck. Enhanced Conversions and Customer Match have shifted from nice-to-have to load-bearing. If your conversion tracking still runs on basic tags with no enhanced data passing back, the bidding algorithm is making decisions with half its instruments dark. ROAS pays for the missing signal.
Smart Bidding is optimising toward the wrong events
This is the most common and most expensive issue we fix. Tracking gets set up once, years ago, and nobody revisits which actions the bid strategy is chasing. If your Target ROAS strategy is counting page views, form starts, or video plays alongside actual purchases, it’s spending to hit a blended signal that has little to do with revenue. The ROAS in your dashboard and the money hitting your bank stop matching, because they’re measuring different things.
The Google Ads Campaign Structure That Actually Fixes It
Fixing a falling ROAS in 2026 isn’t about spending more or pulling back. It’s about handing Google’s machine learning clean data, clear signals, and tight constraints so it optimises toward what you actually want. Here’s the Google Ads campaign structure for 2026 that does that.
Separate branded and non-branded cleanly
Branded search is your cheapest, highest-converting traffic. Non-branded is where you fight for new customers. Put them in the same campaign on a shared budget and Google leans toward branded every time, because it converts better and the algorithm rewards that signal. Your non-branded campaigns get starved, your blended ROAS looks great, and your new-customer acquisition quietly dies.
Split them. Separate budgets, separate tracking. Your blended number may dip slightly, but for the first time you’ll see your real growth picture instead of a flattering average. It’s also worth remembering that much of your branded volume really demands your SEO services and wider marketing built, so paying premium ad rates to recapture it should be a deliberate choice, not an accident of structure.
Build Performance Max with surgical asset group segmentation
Stop running one PMax campaign with one asset group holding everything. That hands the algorithm no direction at all.
Segment by product category, audience intent, or funnel stage. Each asset group needs a tightly defined theme, its own creative, and specific audience signals built from your Customer Match lists and remarketing pools. Given real signals, PMax behaves like a different product. Given nothing, it guesses, and it guesses cheap.
Feed Enhanced Conversions and Customer Match every week
This is the most underused lever in Google Ads right now. Enhanced Conversions return first-party data to Google in a privacy-safe way, recovering click-to-conversion matches the platform would otherwise lose. Customer Match shows PMax and Smart Bidding what your real buyers look like.
If you aren’t refreshing those lists with recent purchasers at least weekly, you’re training the system on stale signals. Fresh data shifts how the machine bids within days, not months.
Clean your conversion setup before you touch a single bid
Before adjusting any bidding strategy, audit your conversion actions. Only primary conversions that represent genuine business value should feed Target ROAS or Target CPA. Demote micro-conversions to secondary status so they still report but no longer steer the bidding.
In our experience this one change, done in isolation, is often where the first few points of ROAS come back, with no keyword or ad edits at all. It’s also the cheapest place to reduce Google Ads CPA, because you stop paying the algorithm to chase actions that never become revenue.
Anchor intent with Search before you scale PMax
PMax and Search coexist well, but they need a hierarchy. Run exact and phrase match Search campaigns against your highest-intent, highest-value queries first. Prove what converts and at what cost. Then use that conversion history to train PMax before you scale its budget.
Launching PMax on a cold account with thin conversion data is one of the fastest ways to burn spend while learning nothing. Give it a foundation and it scales on signal instead of guesswork.
What We're Seeing in the Numbers
Across the accounts we’ve restructured in the last six months, the pattern is consistent: the ones rebuilt around these principles are recovering ground, while the ones still bleeding are the ones running 2023 structures in a 2026 ecosystem.
To put a concrete figure on it, one DTC client we restructured this way moved from a declining 2.9x to a sustained 5.2x ROAS over a single quarter, with CPA down 41 percent, after a conversion-tracking rebuild and PMax asset group segmentation. That’s one account with its own margins and history, not a guaranteed outcome, and your mileage will vary with your category and starting point. But the direction of travel is the same everywhere the structure gets fixed, because the mechanism is the same.
The algorithm isn’t the problem. The signals it’s being fed are.
The Metrics That Tell You Where You’re Bleeding
If your account is in trouble, these five numbers will show you exactly where before you spend a penny fixing it.
Search Impression Share, split by lost-to-budget vs lost-to-rank. This tells you whether you’re losing ground to money or to relevance. The two have completely different fixes, and conflating them wastes budget.
Conversion rate by campaign type. PMax, Search, and Shopping each deserve their own benchmark. If PMax is dragging your blended average down, you won’t see it until you separate the view.
New vs returning customer conversion split. Plenty of accounts believe they’re running acquisition when they’re almost entirely retargeting existing buyers at a flattering ROAS.
Auction Insights overlap rate. If your campaign types are bidding against each other, your budget is fighting itself and you’re paying Google to referee.
Revenue per conversion vs reported ROAS. If these two don’t reconcile, your conversion tracking is broken and your ROAS figure is fiction. Fix this before you trust any other number.
How Social Schnell Approaches a Declining Account
Our paid media work isn’t about tidying dashboards. As a digital marketing agency, our job is to find where the machine is making expensive mistakes and correct them with structure, data, and intent.
Every engagement starts with a full conversion audit, because almost nothing downstream is trustworthy until that’s clean. From there we rebuild the campaign architecture around your actual business objectives, integrate your first-party data properly, and manage the only metric that matters: what lands in your accounts, not what the platform congratulates you for.
If your returns have been sliding and the usual levers have stopped responding, the structure is almost certainly the problem, and structure is fixable.
Conclusion
Rising CPCs are real. Sharper competition is real. But most ROAS drops in 2026 are structural, not market-driven, and that’s the good news, because structure is the part you control. The advertisers recovering fastest stopped trying to manage the algorithm from the outside and started feeding it the signals it needs from the inside.
Clean conversion tracking. Segmented Performance Max. Live first-party data. Branded and non-branded kept apart. None of these are advanced tactics. They’re the fundamentals Google’s current ecosystem demands, and most accounts still aren’t doing them properly. Get the structure right and the returns follow. That part hasn’t changed.
If your ROAS has been sliding and the usual fixes haven’t held, book a free Google Ads audit and we’ll show you exactly where your account is losing money. Contact us and we’ll take it from there.
FAQs
Why is my Google Ads ROAS dropping in 2026?
The most common causes this year are Performance Max sending budget to low-intent inventory, broad match diluting your conversion quality, incomplete first-party data starving the bidding algorithm, and conversion tracking that’s optimising toward the wrong actions. They’re structural problems, not bad luck, which is why spending more rarely fixes them.
Does Performance Max hurt ROAS?
Not inherently. PMax performs well when it has tight asset group segmentation, strong audience signals, and clean conversion data to learn from. Starve it of those and it defaults to the cheapest inventory it can find, which is where ROAS goes to die.
How do I fix Google Ads ROAS quickly?
Start with a conversion tracking audit so only real business-value actions feed your bid strategy. Then separate branded and non-branded campaigns and segment your Performance Max asset groups by intent. These changes typically show measurable movement inside 30 days because they fix the signal, not just the spend.
What is Enhanced Conversions and do I need it?
Enhanced conversions pass first-party customer data back to Google in a privacy-safe way, improving the algorithm’s ability to match clicks to conversions. With third-party cookies fully deprecated in 2026, it’s no longer optional if you want accurate optimisation.
What is a good Google Ads ROAS in 2026?
It depends on your margins and model, but most ecommerce brands should target 4x to 6x after accounting for all campaign types. The more useful question is whether your ROAS is climbing or sliding month over month, and whether it reflects real revenue rather than dashboard reporting.
How can Social Schnell help improve my Google Ads performance?
We rebuild campaign structure, fix conversion tracking, integrate first-party data, and manage ongoing optimisation against real business returns rather than platform-reported vanity metrics. Engagements start with a free audit that pinpoints where your account is leaking budget.
