Aug 4, 2026 · 6 min read · jdfelstead

Brand vs Non-Brand Audit: What Your Blended ROAS Is Hiding

Most Google Ads accounts report one number to the business. Total spend, total revenue, one ROAS figure. It looks tidy on a dashboard and it tells you almost nothing about whether your advertising is working.

The problem is that the number is an average of two completely different activities. Brand campaigns capture people who already know you and were coming anyway. Non-brand campaigns go and find people who did not. Averaging them together hides the performance of both.

A brand versus non-brand audit is the fastest way to find out what your account is really doing. It usually takes an hour, and it changes the conversation with the finance team more than any other single check.

Why the blended number misleads

Brand search is cheap. Your quality scores are high, your competitors mostly are not bidding hard on your name, and the people clicking already intended to buy. Return of 10x, 15x or higher is common.

Non-brand search is expensive. You are paying to interrupt someone who was looking at options. Return of 2x to 4x is a normal starting point for many ecommerce accounts, and lead generation accounts often sit lower still on a first-touch basis.

Put a small amount of brand spend into the same pot as a large amount of non-brand spend and the blended figure drifts upwards. Put a large amount of brand spend in and the blended figure can look excellent while the growth half of the account quietly loses money.

Here is the pattern that catches people out. An account spends £40,000 a month at a reported 6x. Split it, and brand is £12,000 at 18x while non-brand is £28,000 at 1.1x. The business thinks it is scaling profitably. It is not. It is harvesting existing demand and burning cash on acquisition.

Neither half is wrong on its own. What is wrong is making budget decisions from the average.

Step one: define the split properly

Before you can audit the split you need to agree what counts as brand. This sounds obvious and it is where most accounts go wrong.

Brand terms include your company name, common misspellings of it, your name plus a product category, and your name plus a modifier like reviews or discount code. It also includes your own product names if those names are unique to you.

Non-brand covers everything else, including competitor names. Competitor terms behave far more like non-brand than brand, so group them there or keep them in a third bucket of their own.

The grey area is your own product names when they have become generic, or when you sell a branded range that other retailers also stock. If you are a stockist rather than the manufacturer, searches for that manufacturer are not your brand. Treat them as non-brand or as a separate retail bucket.

Write the definition down. If you cannot state it in two sentences, your reporting will drift within a quarter.

Step two: find out where the split is actually happening

Campaign names are not evidence. Plenty of accounts have a campaign called Brand that is picking up generic traffic through broad match, and a Generic campaign that is quietly hoovering up brand searches because the negative lists were never finished.

Pull the search terms report for the last 90 days at campaign level and tag every term as brand or non-brand using your definition. In a spreadsheet, a simple formula that checks whether the term contains your brand name and its common misspellings will do most of the work in seconds.

Then compare two things:

The spend that landed in campaigns named brand, versus the spend that went to terms that are actually brand
The same for conversions and revenue

The gap between those two numbers is your leakage. Anything above five per cent needs fixing before the rest of the audit means anything.

Step three: the five checks that matter

Once the data is clean, run these five checks in order.

1. Brand share of spend. What percentage of total search spend goes to brand terms? There is no universal right answer, but if brand is taking more than a quarter of the budget in an account that is supposed to be driving growth, ask why. If it is under five per cent and a competitor is bidding on your name, you may be under-defended.

2. Non-brand return on its own. Look at the non-brand figure in isolation and compare it against your actual break-even, not against a target someone set three years ago. Break-even is calculated from contribution margin, not revenue. If your contribution margin is 30 per cent, a 3.3x return on ad spend is break-even, and everything below that is costing you money.

3. Brand cost per click over time. Chart brand CPC by month for the last two years. A steady climb almost always means competitors have moved in, and the money you are spending to defend your own name has gone up without anyone deciding it should.

4. Impression share on brand. If you are at 95 per cent impression share on your own name, the last few percentage points are the most expensive clicks in the account. Sometimes that is worth it and sometimes it is not, and the only way to know is to test it.

5. What happens without brand. This is the uncomfortable one. Run a controlled pause or a heavy reduction on pure brand exact terms for a fortnight in a limited region and watch organic traffic and total orders, not just paid ones. Many brands find some of that revenue simply moves to organic. Some find it does not. Either answer is more useful than the assumption you are working from now.

Step four: separate the reporting for good

The audit is only worth doing once if the split then survives in your reporting. That means three practical changes.

Split brand into its own campaign with exact match and a tight negative list, so it cannot absorb generic traffic. Add your brand terms as negatives across every non-brand campaign, including Shopping and Performance Max where possible through brand exclusions. Then rebuild your weekly report so brand and non-brand appear on separate lines, with a blended figure at the bottom rather than at the top.

The order matters. When the blended number sits at the top of the report, it becomes the number everyone quotes. When it sits at the bottom, people read the two halves first.

What people get wrong

The most common mistake is treating the audit as an argument for cutting brand spend. Sometimes it is. Often it is not, particularly in competitive categories where a rival will take the click if you do not. The point of the split is not to kill brand. It is to stop brand performance from making non-brand look healthier than it is.

The second mistake is doing the split once, presenting it, and never touching it again. Brand terms change as you launch products. Competitors come and go from your brand auctions. Broad match and Performance Max keep finding new ways to blur the line. Re-run the search term tagging every quarter at minimum.

The third is splitting the reporting but not the targets. If both halves still carry the same ROAS target, smart bidding will keep pushing budget towards the easy half, and you are back where you started with better looking spreadsheets.

The short version

Your blended ROAS is an average of a harvesting activity and an acquisition activity. Until you separate them, you cannot tell whether your advertising is growing the business or just taking credit for demand that already existed.

Split the terms, measure the leakage, check the five points above, then set different targets for the two halves. It is one of the few audit tasks that costs an hour and can change what you do with the entire budget.