Why brand needs its own measures
Performance measurement counts actions that happened after an exposure and credits them to it. Brand activity is trying to change what people will do later, often weeks or months later, through a channel the company does not control, so the action it causes rarely carries its fingerprint. Judged by attribution reports, brand work looks like money spent on nothing, which is why it gets cut first and why the cost of cutting it appears slowly in rising acquisition costs. The IPA's published analyses of its effectiveness databank have argued that brand-building effects build over long periods and persist, which means a measurement system that only looks at the last few weeks will systematically undervalue them.
Brand measurement therefore has two jobs. The first is to show whether the activity was delivered: did it reach (ReachUnique people who saw an ad.) enough of the audience, at a frequency (FrequencyAverage times a person saw an ad.) sufficient to be noticed, with creative that was recognisably the brand's? The second is to show whether memory changed: awareness, recognition, consideration and the associations the strategy set out to build. The measurement plan discipline used elsewhere on this site applies here: name the question, the measure, the baseline, the method and the review date before the activity starts, so that the result cannot be argued into whatever the budget needs it to say.
Brand metrics
The measures below are the common set. Not every brand needs all of them; a small company can run the delivery, awareness and branded searchSearch queries that contain a company's or product's brand name. rows and learn most of what matters.
| Measure | What it tells you | How it is collected | Moves over |
|---|---|---|---|
| Reach and frequency | Whether the audience was exposed at all, and how often | Platform and media reporting; the reach and frequency calculator for planning | Days |
| Unaided and aided awareness | Whether the brand comes to mind, and whether it is recognised when named | Tracking survey of category buyers | Months |
| Consideration and preference | Whether people would consider or choose the brand | Tracking survey | Months to years |
| Asset recognition | Whether the distinctive assets identify the brand without its name | Tracking survey with unbranded stimuli | Years |
| Brand lift | Whether an exposed group differs from a control on recall, awareness or intent | Platform brand lift studies, as the help centres document them | Weeks, per study |
| Branded search volume | A behavioural proxy for interest in the brand | Search console and keyword tools | Weeks to months |
| Share of voice | The brand's share of category communication, paid, earned or in search | Media monitoring, ad intelligence, search data | Months |
Each row is a brand KPI (KPIA metric tied directly to an objective.) candidate. The one to report at leadership level is usually awareness or consideration among the strategy's audience, with reach and share of voice (Share of VoiceYour share of market visibility.) as the explanations for why it moved.
Awareness and lift, share of voice: the core guides
The core guides take memory and competition in turn.
Brand awareness and brand lift covers the memory measures: aided, unaided and top-of-mind awareness, how a brand tracking survey is designed and read, what the advertising platforms' brand liftThe difference in brand measures between people exposed to a campaign and a comparable unexposed group. studies measure and require as their help centres document them, branded search as a proxy, and the misreadings that turn a small, noisy lift into a strategy.
Share of voice as a brand metric covers the competitive measure: paid, earned and search share of voice, what the published evidence does and does not claim about the relationship between share of voice and market share, the idea of excess share of voice, and how a budget can be set from a target share rather than from last year's spend.
The core guides are complementary. Awareness and lift tell you whether the brand's own activity changed memory; share of voice tells you whether that activity was large enough, relative to competitors, to be expected to. A brand whose awareness is flat while its share of voice is falling has a budget problem; one whose share of voice is rising while awareness stays flat has a creative or consistency problem, which sends the question back to the brand foundations section.
Further guides
Further guides and tutorials turn the measures into a routine: a tracker set up so that its series holds, share of searchSearches for a brand as a share of searches for all the brands in its category, read as a trend beside the brand tracker and market share. pulled the same way every time, a report that ends in decisions and a funnel read stage by stage:
- Setting up a brand tracking survey — the brief, population and screener, a fixed question order, competitors, the baseline wave, cadence and change control.
- Tracking share of search for a brand — choosing the competitor set, building brand term lists, pulling search volumes consistently, handling seasonality and reading it against market share.
- Building a quarterly brand health report — the few measures to show, inputs from tracker, search and media, reading change honestly, and the decisions.
- Reading a brand funnel from tracker results — awareness, consideration and preference read as stages, conversion between them against competitors, noise and what each leak points to.
Linking brand to commercial outcomes
Brand measures are leading indicators; the business wants lagging ones. The link between them is real and never a receipt. Three methods make it visible. Incrementality tests, which hold back brand activity from a region or audience and compare outcomes, answer the causal question directly but need scale and patience. Marketing mix modelling, introduced in the marketing mix modelling basics guide, estimates the contribution of brand spend to sales over time from aggregated data, and is the method most suited to long-lived effects. The third method is cheaper and weaker: watching the performance accounts for the second-order effects brand work should produce, such as rising branded search, improving click-through on generic terms and falling acquisition costs, lagged against brand activity.
None of these produces a single number that settles the budget argument. Together they produce a pattern, and the pattern is what a sensible leadership team asks for: brand measures moving in the right direction, a model that attributes some of the commercial result to brand activity, and performance accounts that are getting cheaper rather than dearer. The branding hub sets out why the two kinds of marketing need each other; this section supplies the evidence that they do, for one company at a time.
A regional bank sets up brand measurement before a campaign
A regional bank planning its first sustained brand campaign commissions a quarterly tracking survey of 1,000 adults in its three-state footprint, with unaided awareness at 22%, aided awareness at 61% and consideration for a current account at 14% in the baseline wave. It records 9,400 branded searches a month and estimates its paid share of voice in the local banking category at 8% against a market share of 11%. The campaign runs for 9 months at $350,000 a month, reaching an estimated 78% of the footprint at an average frequency of 11. By the fourth wave unaided awareness reads 31%, consideration 19% and branded searches 14,800 a month, while share of voice has risen to 17%. A mix model built on 3 years of weekly data attributes roughly 6% of account openings in the campaign period to brand activity, and the bank's cost per funded account from paid search falls by about 15% over the same months.
Common measurement problems
| Symptom | Usual cause | Guide |
|---|---|---|
| Awareness moves but nobody trusts the figure | Small sample, changing questions or an audience that was never defined | Awareness and lift, Setting up a brand tracker |
| A lift study shows a result the team cannot interpret | Lift read as a business outcome rather than a memory change among the exposed | Awareness and lift |
| Branded search rises and falls with promotions, not brand work | Proxy read without separating promotional and seasonal effects | Awareness and lift, Tracking share of search |
| Share of voice reported without a market share to compare it with | Only half the metric collected | Share of voice |
| Budget set from last year's spend plus inflation | No share of voice target | Share of voice |
| Brand spend judged by last-click attribution | No incrementality or modelling in the plan | Marketing mix modelling basics |
| Measures collected but never baselined | Measurement started after the activity | Brand foundations, Setting up a brand tracker |
| Share of search flatters the brand | A competitor set that leaves out large brands; a thorough term list for the brand and thin ones for rivals | Tracking share of search |
| Brand reports show every measure and conclude nothing | No single headline measure; noise read as change; no decisions or owners at the end | Brand health report |
| Awareness is high but the brand is not chosen, and nobody can say where buyers drop away | Measures read one at a time; no stage-to-stage reading against the competitor set | Reading a brand funnel |
