Measuring brand without voodoo

Brand measurement has a deserved reputation problem. On one side, agencies waving six-figure "brand equity trackers" full of indices nobody can act on. On the other, performance teams declaring brand unmeasurable and therefore optional — a position that conveniently frees up budget for the channels their dashboards already flatter. Both are cop-outs. Brand is measurable the way fitness is measurable: not with a single number, but with a small panel of signals that move slowly, correlate with outcomes, and punish anyone who fudges them.

This is the panel we run for clients inside our growth practice — five instruments, one quarterly one-pager, zero voodoo.

Instrument 1: Share of search — the honest proxy

If you can only track one brand metric, track share of search: your brand's search volume as a percentage of the total search volume for all brands in your category. It's free-ish (Search Console, keyword tools, any trends data), it can't be gamed by your own media spend the way impressions can, and it leads revenue by quarters rather than weeks.

The boring specifics that make it trustworthy:

• Define the category set once. Five to twelve named competitors, written down, changed no more than annually. Redefining the set to make the line go up is the classic self-own.
• Use exact brand phrases plus common misspellings, and exclude navigational junk (your brand + "login", + "careers", + "glassdoor") — existing customers logging in tell you nothing about brand gravity.
• Report the share, not the absolute volume. Category volume swings with seasonality and news; share is the competed-for number.

When we rebuilt Copperline Mutual's voice and site, share of search was the quiet hero metric: flat for years, then a steady climb across the three quarters after launch while paid spend stayed constant. Illustrative numbers, of course — this is a concept studio — but the shape is what we see on real engagements: brand work moves search behaviour before it moves pipeline, and search behaviour is what convinces a CFO the brand line isn't a rounding error.

Instrument 2: A tiny recall study, run properly

Commercial brand trackers cost a fortune because they survey thousands of people monthly. You don't need that. You need one well-asked question, asked of the right 150–300 people, twice a year.

The question isn't "have you heard of us?" It's category entry point recall: "When you think about [buying accounting software / choosing a bank / ordering coffee online], which brands come to mind?" — unprompted first, then a prompted list. Unprompted recall is brand gravity; prompted recognition is brand presence. You want both, and you want them moving in the same direction.

Rules that keep it honest:

• Recruit from the category, not your list. Surveying your own customers about whether they've heard of you is a participation trophy.
• Fix the sample profile. Same screens, same quotas each wave, or you're measuring sample drift.
• Publish the methodology on the one-pager. One sentence — n, audience, date. It's what separates a metric from a vibe.

Twice a year is enough. Brand moves like a river, not a tap; monthly surveys just cost more to tell you the water is still flowing.

Instrument 3: The consistency audit

Brand health isn't only external. The cheapest measurement we run is internal: a quarterly audit of whether the brand is being executed. We score thirty recent artefacts — homepage, top five landing pages, ten ads, five emails, five sales decks, product UI screenshots, the 404 — against a short rubric:

• Is the voice on-position? (We use the voice chart sliders, not feelings.)
• Is the visual system being used — right type, right colour tokens, current illustration rules?
• Is the strategic promise legible within five seconds, or has everything drifted into feature soup?

Score each artefact 0–2 per criterion, average it, trend it. The number matters less than the gallery: put the worst offender and the best example side by side in the quarterly report. Nothing improves brand consistency faster than a leadership team seeing their own homepage lose to an email footer.

This instrument catches a failure mode the external metrics miss: brands die from internal entropy before customers notice. By the time recall drops, the decay has been compounding for a year.

Instrument 4: Pricing power and direct traffic quality

Two numbers hiding in data you already have:

Realised price vs. list. Brands with gravity discount less. Track the gap between list price and realised price (or promotional depth in e-commerce). A brand that's getting stronger sees that gap narrow even at stable volume. It's the single most CFO-legible brand metric that exists — it's denominated in margin.

Direct and branded-traffic behaviour. Raw direct traffic is a junk metric (it's full of misattribution), but the behaviour of brand-arriving sessions is gold: do people who arrive via branded search and direct convert faster, bounce less, buy more per order than cold paid traffic? Growing quality of brand traffic means the brand is pre-selling before your site gets a chance. Watch the conversion gap between branded and non-branded cohorts — a widening gap is the brand compounding; a narrowing one means you're paying to acquire people the brand should be delivering.

This is also where brand meets attribution honesty: brand metrics won't cleanly attribute to revenue, and pretending otherwise corrupts both your brand report and your media mix model. Keep them separate, show them side by side, and let the trends argue.

Instrument 5: The qualitative drift check

Once a year, run ten short interviews — customers, lost deals, one or two people who chose a competitor — and ask for associations, not opinions: "What three words? Who is this brand for? What would you expect them never to do?" You're not collecting testimonials; you're checking whether the market's picture of you still matches the strategy document.

The output we care about most is the "never" list — what people believe you'd never do. A brand with a clear "never" list has a position; a brand people can imagine doing anything is a brand that stands for nothing. If the market's "never" doesn't match your strategy one-pager's "we will not" column, that's the most actionable brand finding you can get.

The quarterly one-pager

All of this fails if it arrives as a dashboard. Leadership reads one page. Ours has:

1. Share of search — one line chart, eighteen months, competitor set footnoted.
2. Recall — prompted and unprompted bars, two waves overlaid.
3. Consistency score — one number, with the best/worst artefact pair as images.
4. Pricing power — discount depth trend, one line.
5. One paragraph of judgement. Written by a human. What changed, what we think caused it, what we're doing next. If the paragraph says "too early to tell" three quarters running, the programme isn't working and the page should say so.

Everything else — the crosstabs, the interview notes, the audit gallery — lives an appendix click away for whoever wants it. The one-pager's job is to make brand a standing agenda item with the same dignity as pipeline, not to win a dataviz award. Good governance for analytics applies: decide the metrics before you look at the numbers.

What we deliberately don't measure

• Social engagement as a brand metric. It measures the algorithm's mood and your posting cadence, not brand gravity.
• Awareness percentages without category context. "73% awareness" among people who'd never buy from you is expensive trivia.
• Sentiment scores from listening tools. Directionally interesting, precise-feeling, mostly noise. If you track it, track it as a smoke alarm, not a KPI.
• NPS as brand health. It measures a transaction's aftertaste. Useful, different instrument, different page.

Key takeaways

• Five instruments are enough: share of search, a small recall study, a consistency audit, pricing power, and a yearly qualitative drift check.
• Brand metrics move slowly — measure quarterly, judge annually, and never redefine the competitor set mid-flight.
• The consistency audit catches internal entropy before the market does.
• Denominate at least one brand metric in margin (discount depth) so finance reads the same page.
• Report on one page with a written paragraph of judgement. Dashboards get ignored; arguments get discussed.

FAQ

How soon after a rebrand should metrics move? Consistency scores should move within a quarter — that's execution. Share of search typically needs two to four quarters of sustained, distinctive output. Recall lags furthest. Any agency promising recall movement in eight weeks is planning to change the survey, not the market.

Can a small company with no research budget do this? Yes. Share of search is near-free, the consistency audit is internal labour, and a scrappy recall study among 150 category-relevant respondents costs less than a month of mid-tier paid social. The discipline is the expensive part, not the data.

How do you isolate the brand programme's effect from everything else? You don't, fully — and you should say so on the page. Hold spend roughly constant for a period if you can, log major confounders (PR spikes, competitor scandals, seasonality) as annotations on the chart, and trust the pattern across several instruments rather than any single line. Triangulation beats false precision.

What's the first sign a brand programme is failing? Internally: the consistency score slides while everyone insists they're "on brand." Externally: branded traffic converts no better than cold traffic — meaning the brand isn't pre-selling, it's just a name.

Should brand metrics be tied to bonuses? Cautiously. Tie teams to the inputs they control — consistency, output volume, discount discipline. Tying compensation to share of search invites gaming the metric set. Brand gravity should inform judgement, not replace it.