Audience audit · for businesses that buy traffic
You're paying for traffic. Do you know who shows up?
Your analytics counts sessions. It can't tell you who arrived, how much of it was human, or how your buyers react to what you're running. An Audience Audit measures your audience with independent instruments, then tests your marketing against modeled archetypes of the people you actually reach.
INDEPENDENT INSTRUMENTS · AGGREGATE STATISTICS ONLY · EVERY CLAIM GRADED
Schematic. The case study below shows the real numbers.
The problem
Sessions are not people.
Every dashboard you own reports how much: sessions, clicks, conversion rates. None of them can answer the question the spend depends on: who is actually arriving, and is it the audience you're paying to reach?
"48,231 sessions. 2.4% conversion. Top channel: paid social."
Counts and rates. Bots inside the averages. Demographics inferred by the same platforms selling you the traffic. Creative decisions made on hunch, because nothing here says who these people are.
Read from independent instruments on your property and your own first-party list. Every claim carries source, method, basis count, and grade.
One tells you how much you got. The other tells you who you got, and whether they're worth buying more of.
Case study · our own property, full disclosure
The Backcut, audited end to end.
The Backcut is our newsletter for tree service business owners, built with Meta ads. We audit it with the same instruments we put on client properties, which makes it the case study we can publish with nothing blurred. Every number below is live measurement.
01 · The site: who the ad spend actually lands
23.2% of Meta-ad devices resolve to identified people: paid-traffic quality, measured instead of assumed. 6% of pageviews are declared AI crawlers and 16% suspect automation, counted separately, never blended into the human number. Four issues in, the numbers now move issue over issue and the passport reports the trend instead of a snapshot. The live measurement document →
02 · The list: who actually subscribed
Among the 208 resolved subscribers carrying employment data: 98 are C-level, and the top titles read Owner (38), President (18), Business Owner, CEO, Arborist. Of those with company-size data, 109 of 185 run 1-to-25-person firms. Top industry: Environmental Services, which is where tree care lives. 74% male, modal ages 35 to 64; Texas, Michigan, California, and Georgia lead. Re-matched fresh for this re-issue: the rate held (25.1% → 25.9%) as the list grew. Composition is read from the resolved 26%; the unresolved remainder is reported, never projected.
The sentence this buys: the list the ads built is the list we set out to build: owner-heavy, small-company, propertied. We can say that with basis counts instead of vibes, and when a claim doesn't hold, the audit says that too.
03 · The archetypes: testing marketing before budget
Six modeled buyer archetypes are seeded from the measured composition, clearly graded as the modeled layer. Ads, landing pages, and even the brand's live Google results page get run against them, returning receptivity scores, objections, and blockers per archetype, before more money goes behind the creative. See a live panel test →
04 · The creative test: three real Meta ads, one panel
The loss story wins on every archetype except the one the data says can't be bought from cold social anyway; it also won a separate head-to-head comparison run, in which the two runners-up swapped order, so treat second and third as a tie. The panel's universal blocker: funnel suspicion. Five of six archetypes said some version of "the second this turns into a coaching pitch, I'm out," which is creative strategy and editorial policy in one finding.
The legacy frame converts identity instead of action (emotion 4.19, urgency 3.33): keep it for warm retargeting or add one line saying what the newsletter contains. The competitor-envy ad self-selects the growth segment and actively repels the referral-only veteran archetype (31%): fine if targeting can exclude him, waste if it can't.
05 · The actions: what the audit tells us to do next
High · Keep early issues upsell-free. Funnel suspicion was the universal blocker across the archetype panel, and the heaviest archetype is the most pitch-averse; on that segment the win is not losing it. (§03, §04)
High · Concentrate spend on the loss-story creative; pull the competitor-envy ad from cold audiences unless targeting can exclude the referral-only veteran it repels. (§04)
Medium · Watch the suspect-automation share issue over issue (18.4% at day 9, 17.7% at day 12, 16.0% at day 16). A five-point rise triggers a placement review; no action below that. (§01)
Medium · The measured age curve runs older than the creative assumes (45-54 and 65-and-older trade the top band issue over issue). Test one angle written for the semi-retired legacy operator before scaling spend. (§01, §03)
This section is the deliverable most clients actually buy: the same ranked-fixes discipline, run on your property and your list.
Deliverables
What you walk away with.
Who arrives on your property, by channel, with humanness measured at the pageview level. Passport format: graded claims, basis counts, yours to keep.
Your subscriber or customer list resolved against a 274M-person identity graph. Aggregate statistics only; your list never leaves your hands.
Your creative and your brand's live search results run against archetypes seeded from your measured audience, with per-archetype blockers and drivers.
Trust leaks, channel mismatches, creative blockers, competitors capturing your brand terms: ordered by expected impact, each one checkable.
How it works
Two script tags and your list. That's the setup.
Green means read from an instrument. Amber means modeled, and labeled as such. The two never blend.
Instrument
Two measurement tags on your property, about ten minutes in your tag manager. List matching and SERP testing start the same day; nothing waits on the tags.
Measure
A 30-day collection window on the site. FouAnalytics reads humanness; deterministic identity resolution reads who the audience is, in aggregate only.
Strategy meeting
Findings walked through live, ranked by expected impact, with the documents in your hands before the call.
What this is not
Three things we don't do.
We don't resell analytics you already own: this measures who, not how much. We don't angle for your media budget: we buy no media, so the findings have no thumb on the scale. And we never deliver personal data: aggregate statistics only, the same bright line as every document we issue.
How the instruments work →Request a strategy meeting.
Write with your domain and roughly what you spend in a month. We'll reply with whether the audit fits, what we'd measure first, and what it costs.
Write to [email protected]Instruments in minutes · first findings the same week