How to measure brand awareness: 7 methods, and what each one actually tells you
A chief executive asks how many people have heard of the company, and whether that number is bigger than it was last year.
Most teams have three honest answers available, and none of them is good: "I don't know", or "I'll know in about six weeks when the next wave of research comes back", or a number from research that finished in March. It is now September, and the question was about September.
Brand awareness is measurable, by at least seven methods, and most teams are already running two or three of them without having decided which one they trust. The methods measure different things, at wildly different costs and speeds, and almost nobody writes down which question each one is capable of answering.
In this article we'll go through all seven, ordered by how much they tell you rather than by how easy they are to run. The first two we'd call instruments; the rest are indicators, useful and partial and frequently mistaken for the thing itself.
1. A brand tracking survey
The tracker goes first because it is the only method on this list that can ask a person a question.
A tracker recruits a sample of your target population and asks two things. Unprompted: "which brands can you think of in this category?" That is unaided awareness, and it is the closest thing marketing has to a measure of whether your brand exists in someone's head without being put there first. Prompted: "which of these brands have you heard of?" That is aided awareness, and the gap between the two is one of the most diagnostic numbers in brand measurement. No other method here produces it.
Trackers also carry attribute batteries (is this brand trustworthy, expensive, for people like me), which are the raw material of positioning work. If you want to know what people believe about you, this is the instrument, and there isn't really a second one.
What it costs you. Latency and money. A campaign that starts in April gets its verdict from a quarterly tracker in August, and between waves you have nothing. The sample sizes that fit most budgets also tend to produce quarter-on-quarter movements small enough that the honest reading is "no change detectable", which is true, unsatisfying, and gets over-interpreted in the room anyway.
The complaint about trackers is almost never accuracy, and we'd suggest you don't make it one. It's that they answer in seasons, and you fix that by putting something continuous underneath the tracker rather than by replacing it.
2. Share of Search
Share of Search is the proportion of all search demand in a category that goes to your brand. Not your search volume, but your share of the category's, and that distinction is the whole method.
The long version of the definition is in our guide to what Share of Search is, so here we'll stay on how the series gets built, which is where most attempts come apart.
Define the category as a consideration set, not an industry. It's the group of brands a buyer would plausibly weigh against each other, usually narrower than an industry classification and occasionally wider: a challenger bank competes with the incumbents and with two fintech apps that appear in no banking taxonomy. Get this list wrong and every number downstream is wrong with it, because each one is a fraction of this list.
Build a query set per brand. Brand name, the common misspellings, brand plus the main product line, brand plus "reviews". Names that are ordinary words, like Apple, Orange or Shell, need explicit filtering or the volume you get back is nonsense, and the set has to stay constant for the series to mean anything.
Pull monthly volume for every brand over the same window and the same geography.
Normalize. Each brand's monthly volume divided by the category total for that month. Everything before this step is keyword research; this step is what makes it a brand metric.
Smooth if the category is small, usually with a three-month rolling average, since single months in low-volume categories move for reasons that have nothing to do with the brand.
You can do all of this by hand, and plenty of people do, including in our own walkthrough of how to calculate Share of Search in a spreadsheet. Having built it both ways, we'd say the manual version is fine as a one-off and unreliable as a series: query sets drift, someone re-pulls with a different window, a competitor gets added in month nine, and the chart stops being comparable to itself without anyone noticing. Branquo exists to hold those decisions still.
Where the volumes come from matters as well. Google Trends is not where we get ours, partly because its numbers can be somewhat unreliable and partly because Google has never said publicly what the index is really counting. Branquo reads from the Google Keyword API instead, and the practical difference is that it returns absolute monthly volumes. Absolute volumes can be added together into a category total; an index scaled against its own peak cannot.
That denominator is the load-bearing part, because it separates two things absolute volume glues together. A brand's own branded search can rise 20% in a year while its share of the category falls, because the category grew 35%. On the raw number that brand is winning; on the share it is being outgrown. Those are opposite briefings.
Why we'd call it an instrument rather than a proxy. It's continuous, with a new reading every month instead of every quarter, and it's behavioral, counting what people did rather than what they later told a researcher. It also reads the whole category: a panel measures the people you paid to ask, while search measures everyone who went looking. There is also published work connecting Share of Search movements to later market share movements, which is the claim that makes it more than a curiosity.
The agency point. You can build this series for a brand you have no relationship with: no panel, no fieldwork, no procurement, no signed contract. So the first conversation with a prospect can open with their own category on a chart, their line and their three competitors' lines beside it, before anyone has agreed to anything.
If you want to see what your own category looks like, our free Share of Search tool will do it without a signup: type in the brands, and a report comes back in a few seconds.
What it costs you, and what to do about it. Four things, and they're real:
It only sees brands people can already name. This sits closer to aided awareness than unaided, and it won't surface the brand nobody has heard of. Pair it with an annual tracker wave if unaided recall is a board-level number for you; the series still tells you how you're moving against every competitor in between.
Branded search is contaminated by people who already bought. Logins, support queries, order tracking. Decide once whether to filter navigational queries out, then apply that rule to every brand every month, and the comparison stays readable even where the level is inflated.
It doesn't distinguish good attention from bad. A brand in a bad news cycle can look, in search, much like a brand in a good campaign, and Volkswagen's emissions scandal is the standard illustration. You usually know which one week three was, so annotate the series rather than reading it bare.
Small categories are noisy. Below a certain monthly volume the series is mostly variance that looks like a trend. Smooth it first, and if it's still noise, label the category as one where search is a weak instrument rather than reporting the line anyway.
3. Branded search volume on its own
This is the same signal as method 2 with the denominator removed, and it's what most teams already have sitting in Search Console.
It's better than nothing and it will catch a large campaign or a large crisis, but it can't tell you whether a rise is yours or the category's, so it flatters brands in growing categories and unfairly punishes brands in shrinking ones. Use it when you genuinely have no competitor set, and treat that as temporary, because naming four competitors is a shorter job than most people expect.
4. Branded organic and direct traffic
Sessions in your analytics platform from people who typed your name or arrived directly.
This measures arrivals rather than awareness: it counts the aware people who went looking and then completed the journey in a way your analytics could see, after cookie consent, ad blocking, app traffic and dark social have taken their cut. It also moves when your SEO changes and when someone redesigns a consent banner, and by construction it covers your own brand only.
5. Platform brand lift studies
Meta, Google and YouTube will run a controlled exposure test: an exposed group and a holdout, then a survey question put to both.
The methodology is sound and the causal claim is real, which is more than most of this list can say. The limits are that it measures one campaign on one platform among the people that platform reached, that it exists only while the campaign is running, and that it isn't comparable to anything outside that platform.
6. Social listening and mention share
Volume of brand mentions across social and news, often expressed as a share of category conversation.
The weakness is definitional: it measures who is talking, not who is thinking. Conversation is concentrated in a fairly unrepresentative slice of any population and spikes hardest for controversy, and categories differ enormously in how much anyone posts about them. It's genuinely good at detecting events and weak as a level.
7. Impressions, reach and opportunities-to-see
Reach and OTS describe how many people could have seen the advertising, which makes them a measure of media delivered.
Reporting that as awareness is reporting your spend back to yourself with a different unit on the axis, and everyone in the room knows it, which is why the follow-up is always some version of "yes, but did it work?" Keep these numbers for planning, and keep them off the brand health slide.
Where search stops and research starts
Go back to the question the chief executive asked in September. One method here can answer it in September: a normalized Share of Search series shows the brand up or down against its own category, this month, with the twenty-three months before it on the same axis and every competitor's line beside it.
It isn't built for the questions underneath that one, and some of those belong to research. The obvious one is why the line moved, since every method here is a level or a rate and none of them contains a cause. Unaided awareness only the tracker can reach, because every behavioral signal requires the person to already have the name. And what people believe about you belongs to survey and qualitative work completely; we wouldn't claim search behavior substitutes for it.
The useful arrangement is one instrument telling you when to reach for the other. Research commissioned on a calendar goes into field four times a year whether or not anything happened, whereas research commissioned off a continuous series goes into field because the March shift is real and sustained, with a specific month to ask about and a specific competitor to ask against. Share of Search answers when and by how much. The survey answers why, and in our experience it answers it rather better when something has already told it where to look.
What to do next
Start by writing down your category as a consideration set: every brand a buyer would weigh against you, by name. Not the industry, and not whoever happened to be in the last competitive deck.
If that list is hard to produce, or if three people on your team produce three different lists, stop there. The disagreement is your first finding, and it undermines any brand measurement you commission until it's settled, including the expensive kind.
If the list is straightforward, take the first four names and put a running series under them rather than a screenshot. A free Branquo account gives you one dashboard with four brands on it, one keyword per brand, and a monthly reading that keeps arriving without anyone having to remember to re-pull it. That's usually enough to see whether your category moves in search at all, and to have the chart ready the next time the question gets asked in September.
Then decide what the tracker is for. Most teams running quarterly research are using it to answer two questions at once: is the brand growing, and what do people think of it. The second is what a survey is uniquely good at, and we'd keep paying for it. The first is the one it answers three months late.