Branded Search Is Lying to You: How to Actually Measure Brand for Ad-Driven Stores

Here's what trusting branded search as your brand scoreboard actually costs you. You see the line climbing in Google, you decide the brand is getting stronger, so you keep pouring money into the top of funnel that's secretly creating that line. And the whole time you've got no idea whether you've built anything that would survive you turning the ads off.

That's not a small misread. That's a budget decision you make every month based on a number that's mostly measuring your own ad spend back to you.

I heard a founder put it perfectly on a podcast recently. He said his team had run the test, and every extra ten dollars they spent on advertising bought them roughly two more branded searches. So the searches weren't a sign of love. They were a sign of spend. As he put it, he's one of the most-searched names in his category, and so is the giant luxury brand right behind him, because they both spend enormous amounts on ads. Nobody sits at the top of branded search without buying their way there.

Sit with that for a second, because it kills the favourite vanity metric in DTC. If branded search rises and falls with your ad budget, it cannot also be your independent proof of brand. It's the same dollar counted twice.

So what do you actually measure instead? I don't think there's one clean number, and anyone selling you one is selling you another branded search. What I run with clients is a stack of imperfect signals that, read together, tell you something branded search never will: how much of your business would still be here if you went quiet. Here's the stack.

Start by asking customers directly: the post-purchase survey

The single most useful brand signal in your whole business is a one-line survey on the thank-you page. "How did you hear about us?" Plain text box or a few options, asked at the moment of purchase.

People in the trenches call it HDYHAU, and the reason it matters is that it catches the thing attribution physically cannot. When someone types "heard about you from a friend" or "my sister has one", no pixel on earth was ever going to record that. Word of mouth is invisible to your ad platforms by design. The survey is the only place it shows up.

What you're watching for is the slice that says friends, family, or word of mouth. On the podcast I mentioned, the founders kept landing around a third of customers saying exactly that, month after month. That's a genuinely strong brand signal, because it's demand the ads didn't manufacture.

Here's my take on reading it honestly, though. A high word-of-mouth number is brand, yes. But it can also just mean your ad spend is low relative to your peers, so referrals are a bigger share of a smaller pie. So don't read the percentage in isolation. Read whether it holds, or even grows, while you're actively scaling paid. Referrals staying strong as new-customer volume climbs is about as close to proof of brand as this game gives you.

Read the whole business, not the platform: blended MER

Drop the per-platform ROAS for this and look at MER, your total revenue divided by your total ad spend across everything.

Why MER for brand? Because it's the one number that quietly captures all the demand the platforms can't claim. Your branded search conversions, your direct visits, your email, your word of mouth, the bloke who saw you on a mate's coffee table. None of that gets a tidy attribution row, but all of it lands in total revenue. So if your real brand strength is growing, MER tends to drift up over time even when your in-platform ROAS is flat, because you're getting more total revenue for each dollar of ads.

I treat a slowly rising MER trend as one of the better brand tells going. Not any single month, those bounce around. The trend over a couple of quarters. If you're spending roughly the same on ads but pulling more total revenue, something is doing the selling that you aren't paying for per click. That something is brand.

The hardest proof: repeat rates by cohort

If I had to defend brand with one set of numbers, it'd be these. There's a school of thought I agree with that brand really comes down to three things you can measure: pricing power, margin, and repeat purchase rates. Those are the maths of brand, not the vibes of it.

Repeat rate is the most accessible of the three, so start there. Group customers by the month you acquired them, then watch what share of each cohort comes back to buy again over the following months. You're not after a single blended repeat number. You want the curve, cohort by cohort, because that's what tells you whether the experience is actually creating loyalty or whether you're just renting one-time buyers from Meta.

A healthy DTC business often runs something like a third of monthly revenue from returning customers, the rest from new. That sort of split, paired with cohorts that keep coming back at a consistent or rising rate, is a real brand. People liked it enough to do the one thing ads can't make them do: come back without being chased.

One caveat I always flag. Repeat behaviour is partly your category, not just your brand. A consumable gets repeat purchases on built-in cycles. Something people buy once every few years won't, no matter how beloved. So compare your cohorts to your own past cohorts, not to a brand in a different buying rhythm. The question is whether your repeat curve is improving over time, which is the bit you actually control.

The leading indicator: direct and organic session trends

This is the one founders reach for first and trust the most, which is exactly why I'm putting it fourth with a warning attached.

Direct traffic and organic sessions are a fair brand signal, with one big asterisk: a chunk of "direct" and "organic" isn't really organic at all. When you run more ads, more people skip the ad and type your name in later, or come back to the site directly. So a slab of that traffic is just your paid spend wearing a disguise, the same trap as branded search.

You can still read it usefully if you read it the right way. Watch direct and organic sessions relative to your ad spend, not on their own. If your spend is flat and direct sessions keep climbing, that's a real signal. If direct only rises when spend rises, you've just found your ad budget again in a different report.

The version of this I rate most is what some operators call baseline: the revenue and sessions that aren't coming from any paid channel, tracked week over week. More people seeking you out without an ad in front of them is the cleanest version of "the brand is growing". Make it a weekly number your team actually looks at, the same way you look at spend.

The number you really want: your organic floor

Every founder eventually asks the real question. "How much of this business is actually mine, and how much am I renting from Meta?" Put differently: if I went dark on ads, what's the floor I'd land on?

The honest answer is that the only way to know for certain is to stop spending and watch what's left. And almost nobody should do that, because going quiet for the months it'd take to read cleanly can do damage you don't recover from. I've watched founders consider the six-month blackout as a test. It's not a test, it's a self-inflicted wound.

So you estimate the floor instead of measuring it, and the stack above is how. Your word-of-mouth share from the survey tells you what demand isn't ad-driven. Your repeat cohorts tell you what revenue is already baked in from people you've paid for once. Your baseline sessions tell you what traffic shows up without a campaign behind it. Stack those together and you get a rough, honest read on the part of the business that would still be breathing if you went quiet, without having to actually go quiet.

If you want to pressure-test it more gently than a full blackout, you can run small, contained experiments. Geo-based incrementality tests, where you pull spend in a few matched regions and compare them to the rest, give you a read on what's really incremental without betting the whole business on it. A handful of operators I respect run exactly these rather than trusting platform attribution, precisely because the platforms always claim more credit than they've earned.

The trap underneath all of this

There's a confession I hear from good operators more than you'd think. Some version of "I should have put more into brand years ago, but I couldn't measure it, so I didn't, and now the lack of it has caught up with me."

That's the real cost of leaning on branded search. Not that it's a slightly soft metric. That it flatters you into thinking the brand is fine, so the genuinely hard, slow, unglamorous brand-building never gets funded, until growth stalls and you can't buy your way out anymore.

None of the signals above is perfect on its own. The survey can be skewed by low spend, sessions can be ads in disguise, repeat rates are tangled up with your category. But read together, week after week, they triangulate something branded search never can: how much of your business is genuinely yours.

So here's the question I'd actually sit with. If your ad account went dark tomorrow, what's your honest guess at the floor you'd land on, and do you have a single number on a dashboard right now that would tell you whether that guess is anywhere near right?