Your agency tells you your cost per lead is $45, which sounds reasonable when you're spending $9,000 a month and getting 200 leads. The math checks out on paper. But when you look closer at where those 200 leads actually came from, you find that 120 of them are people who searched your business name. They already knew who you were and were going to contact you whether the ad was there or not. The other 80 came from people searching for your service generically, people who didn't know you existed until they saw your ad, and those are the leads your advertising actually generated.
The 120 brand leads cost about $12 each and convert at 30-40% click-to-lead because the person already decided they wanted you before they searched. The 80 non-brand leads cost about $94 each and convert at 5-8% because those people are comparing options and you’re one of several results on the page. Your real cost to acquire a new customer through advertising is $94, not $45. But $45 is the number that goes in the report, because it’s the blended average of cheap brand traffic and expensive new-customer traffic mixed together.
This post is part of our 12-point Google Ads audit. We look at campaign structure and CPA reporting as step 10 because by this point we’ve verified the conversion data, reviewed the search terms, and understand what’s actually being counted. Now we’re asking whether the numbers being reported mean what you think they mean.
Why Brand Traffic Makes Everything Look Better Than It Is
Brand traffic is when someone searches your business name, your owner’s name, or a specific branded term that shows they already know who you are. "Smith Accounting," "Smith Accounting Dallas," "John Smith CPA." These people aren’t shopping around. They’ve already been referred by a friend, or they’ve already visited your website once before, and now they’re coming back to take the next step. When they click your ad and convert, that conversion was going to happen with or without the ad. You’re paying for a click from someone who was already your lead.
Brand traffic converts at dramatically higher rates than non-brand. In most service businesses we manage, brand campaigns convert at 25-40% click-to-lead while non-brand campaigns convert at 4-10%. The CPC is lower too, often $1-3 for brand versus $8-25 for non-brand depending on the industry. So brand leads are cheap and plentiful, and non-brand leads are expensive and harder to get. When you blend those two numbers into a single "account CPA," the cheap brand traffic pulls the expensive non-brand traffic down to a number that looks comfortable. The problem is that the comfortable number isn’t real.
Think of it this way. If you stopped running all Google Ads tomorrow, most of those brand searchers would still find you. They’d click your organic listing instead, or they’d go directly to your website. You’d lose some of them to competitors who are bidding on your name (which is a separate problem we cover in our brand impression share post), but the majority would still convert. They weren’t generated by advertising, they were generated by your reputation, your referrals, and your other marketing.
You're just paying Google to be the middleman on a transaction that was already happening.
The Math That Changes the Conversation
Let me walk through a real example of how this works. Say your account has two campaigns: a Brand campaign and a Non-Brand campaign targeting your core services. Over 90 days, here’s what the data looks like:
| Campaign | Spend | Conversions | CPA | Conv. Rate |
|---|---|---|---|---|
| Brand | $1,800 | 150 | $12 | 35% |
| Non-Brand | $7,200 | 50 | $144 | 6% |
| Account total (blended) | $9,000 | 200 | $45 | — |
The $45 blended CPA is technically accurate. You did spend $9,000 and you did get 200 conversions. But that number communicates nothing useful about the cost of acquiring a new customer. The 150 brand conversions are people who already knew you, and the 50 non-brand conversions are the ones your advertising actually found. Your true customer acquisition cost is $144, not $45. If your agency is reporting the $45 number without breaking it out, you’re making business decisions based on a number that’s off by more than 3x.
This gets worse when the brand percentage is high. An account where 75% of conversions come from brand can show a $30 blended CPA while the non-brand CPA is actually $200. The higher the brand percentage, the more the blended number lies. And most agencies are incentivized to let it lie, because the blended number makes them look better.
What Happens When You Try to Scale
The scalability problem is where blended CPA reporting actually costs you money, not just confuses you. When the blended CPA looks good, the natural decision is to increase budget. More leads at $45 each sounds great, so you bump spend from $9,000 to $15,000.
But brand demand is relatively fixed. There are only so many people searching your business name each month, and that number is driven by awareness, referrals, and reputation, not by how much you spend on ads. If 150 people per month search your name, spending more on brand campaigns doesn’t create a 151st searcher. It just means you’re already capturing most of them. So the additional $6,000 in budget goes entirely to non-brand campaigns.
Suddenly your numbers look different. The brand campaign still generates 150 leads at $12 each. The non-brand campaign, now with $13,200 in budget instead of $7,200, generates maybe 85 leads at $155 each (CPA goes up as you push into more competitive auctions and broader intent). Your new blended CPA is $15,000 / 235 = $64. Your CFO sees CPA jump from $45 to $64 and panics. The agency gets a phone call asking what went wrong.
Nothing went wrong. The non-brand CPA was always $144+, you were just never looking at it separately, so you didn’t know what "more budget" would actually produce. The $45 CPA was never scalable because it was never real. The blended number created false expectations about what additional spend would return, and when reality caught up, it looked like performance declined when really you just finally saw what things actually cost.
How Smart Bidding Gets Confused
If brand and non-brand keywords live in the same campaign (which is more common than it should be), Smart Bidding has a structural problem. The algorithm optimizes toward a single CPA target for the entire campaign. It sees the brand keywords converting at 35% and $12 each, and it sees the non-brand keywords converting at 6% and $144 each. The blended campaign average looks acceptable, so the algorithm keeps bidding aggressively on the expensive non-brand terms because the cheap brand conversions pull the average down.
In practice, this means Smart Bidding will overspend on non-brand traffic as long as brand traffic keeps the blended number within target. It has no concept of "these brand conversions would have happened anyway, so don’t count them when deciding how much to bid on non-brand." It just sees conversions and cost, and if the blend looks good, it keeps spending. This is why separating brand and non-brand into different campaigns with different targets isn’t just a reporting preference, it’s a structural requirement for Smart Bidding to function correctly.
When we rebuild accounts, one of the first things we do is isolate brand traffic into its own campaign with its own bidding strategy (usually Target Impression Share at 90%+, not a CPA target). Non-brand campaigns then get their own realistic CPA targets based on what non-brand traffic actually costs to convert. This forces the algorithm to optimize non-brand performance honestly instead of hiding behind brand subsidies.
Performance Max Makes This Worse
Performance Max campaigns are particularly bad at this because they blend everything by design and give you almost no visibility into what’s happening inside them. PMax runs across Search, Display, YouTube, Gmail, Maps, and Discover simultaneously, and it reports a single blended CPA for all of it. When PMax shows a $40 CPA, you have no idea how much of that is brand search traffic it intercepted versus actual new-customer acquisition from Display or YouTube.
Without brand exclusions configured (which Google only made self-serve in late 2024, and which only apply to Search and Shopping inventory within PMax), Performance Max will happily bid on your brand terms and claim those conversions as its own. Your Search brand campaign loses impression share, PMax picks up those brand clicks, and PMax’s CPA looks amazing because it’s capturing conversions that were going to happen anyway. Meanwhile your non-brand Search campaigns look worse by comparison because they lost their brand subsidy.
We see this regularly on accounts running PMax alongside Search. The PMax campaign shows great numbers, the Search campaigns look mediocre, and the natural conclusion is "PMax is outperforming Search." The reality is usually that PMax is claiming brand conversions while Search is doing the actual new-customer work. Separating the data requires pausing PMax for 2-4 weeks and watching what happens to overall conversion volume. If it doesn’t drop significantly, PMax wasn’t generating incremental demand, it was just reassigning credit.
Mixed Intents at the Ad Group Level
The brand/non-brand split is the most dramatic version of this problem, but the same blending issue happens at smaller scales throughout the account structure. An ad group that contains both "emergency AC repair" and "AC maintenance plan" is mixing someone with an urgent need (high intent, converts fast, willing to pay premium) with someone planning ahead (lower intent, shops around, price-sensitive). The blended ad group CPA might look fine, but the emergency keywords are subsidizing the maintenance keywords, and you can’t see which is which.
This matters for ad copy too. The RSA serving both intent types can’t speak directly to either one. It can’t say "24/7 Emergency Repair" to the emergency searcher and "Annual Maintenance Plans from $149" to the maintenance searcher if both keywords live in the same ad group. One ad has to try to serve both intents, and it does neither well. Splitting them into separate ad groups with intent-matched ad copy lets you see the true CPA for each service type and write ads that actually speak to what each person is looking for.
The old SKAG model (Single Keyword Ad Groups) was an extreme version of this principle that’s no longer practical with how match types work in 2026. But the underlying logic is sound: group keywords by intent so your data isn’t blended and your ads can be specific. You don’t need one keyword per ad group, but you do need each ad group to represent a single coherent intent. "Emergency AC repair" and "24 hour HVAC" belong together. "AC maintenance plan" belongs somewhere else.
How to Check Yours
If your brand and non-brand traffic are already in separate campaigns (which they should be), this is straightforward. Pull up your campaign-level reporting, look at the brand campaign CPA separately from the non-brand campaign CPA, and compare both to the blended account number. The gap between your non-brand CPA and your blended CPA tells you how much brand traffic is subsidizing the reported number.
If brand and non-brand are mixed in the same campaigns (common on accounts that were never properly structured), you’ll need to look at keyword-level data. Sort by conversions descending and identify which converting keywords are branded terms. Add up the brand keyword conversions and cost separately from the non-brand keyword conversions and cost. Calculate each CPA independently. The difference will probably surprise you.
Also check what percentage of your total conversions come from brand. If it’s over 50%, your blended CPA is being significantly pulled down by traffic that wasn’t generated by the ads. If it’s over 70%, the blended number is almost meaningless as a measure of advertising effectiveness. You’re essentially reporting "how much it costs to be a middleman for people who already wanted you" and calling it your cost per lead.
The question to ask your agency: "What's my non-brand CPA?" If they can answer immediately with a specific number and show you how they calculated it, they're managing the account properly. If they hesitate, show you the blended number, or say "we don't break it out that way," you're not getting accurate reporting.
Why This Matters for Business Decisions
The blended CPA isn’t just a reporting nuisance, it drives real business decisions in the wrong direction. When the blended number looks affordable, businesses invest more in ads expecting the same return. When those investments don’t produce the expected results (because the incremental spend goes to non-brand at the real CPA), confidence in the channel drops. "Google Ads stopped working" is a common complaint that often just means "we finally pushed budget past the brand subsidy zone and saw what things actually cost."
If your agency has never shown you a brand vs. non-brand CPA breakdown, never explained what percentage of your conversions come from brand traffic, and never discussed whether your blended metrics are scalable, that’s a conversation worth starting. The answer might be that your non-brand performance is actually strong. But it might also explain why every time you increased budget, the results didn’t follow.
This is one of the 12 steps in our full Google Ads audit process. If you want us to pull your brand/non-brand split and show you what your CPA really looks like, request a free audit and we’ll run the numbers.
