Has the paired, blind comparison been run and published?

Not yet. What is fixed: five brands, each read twice from the same brief, once by a human-led audit and once on nunNEO; an independent reviewer who does not know which read is which scores both on findings found, findings wrong and hours spent. What is still open: the five brands, the reviewer and the date. The result will be published on this page with its date, whichever way it comes out, and the claims above will be revised to match it. Until then, every figure in this piece carries the qualifier it was written with: on this record.

Does this hold outside dental, medtech and healthtech?

Two of the claims travel and one does not. Time to first finding and operator hours per read are properties of the method: the engine reads public surfaces against a competitor set at the same pace whatever the sector, and the record holds reads in professional services, biotech and pharma and motorsport that returned the same way. Those reads are the control in this study. The rubric, the competitor logic and the confidence rating held where the vocabulary changed. The findings quality claim is made for healthcare only. nunNEO is built around the healthcare buyer, the committee that reviews a vendor without taking a call and the claims discipline a regulated market demands, and the one engagement that produced the quality evidence was a dental company. Outside healthcare, treat the read as a structured first pass that a sector specialist verifies. Inside healthcare, the record is deepest in dental, with seven brands and all three GTM reads. The single biotech and pharma read is one data point, and pharma's claims environment is heavier than dental's, so the exceptions the checks surface deserve more of the strategist's time there.

What does a strategist need in order to run one?

Less than the hours suggest. The client's founder or marketing lead completes the intake, including the self-rating on the dimensions the engine will score. The read returns in about 20 minutes. The operator's two hours go three places: reviewing the items the engine's own checks flagged as a mismatch, turning the scored read into the findings deck and writing the call memo that sets up the decision call. Nothing is installed on the client's side and nothing is integrated; the read is made on public surfaces. The rubric, the 12 dimensions and how to read a confidence rating ship with the read, so the training is reading it once. Every read runs its own verifications before it is released, and a senior strategist reviews only the items that fail one, which means an operator who commissions a read receives a verified read with its exceptions marked rather than a raw output. An operator who wants reads inside their own practice starts at the link below.

Where does a human-led audit still win?

In four places, and the read is the wrong tool for each of them. A brand with no public surfaces yet: a pre-launch company, a stealth product line, a spin-out without a site. The engine reads what is public, and there is nothing to read. Questions that live inside the company: culture, employer brand, whether the leadership team agrees on what the company is. Interviews are the method for those, and the self-rating in the intake is the only inside view the read takes. Decisions that are political as much as analytical: brand architecture after an acquisition, a rename, a founder stepping back. The read informs the room; the workshop is where the room decides. And buyer reaction: whether a specific message lands with a specific committee. That is panel testing, and the read's job is to tell you which message is worth the test. In each of those cases the human-led method is right because of what it measures, and the comparison in this piece does not apply.

What happens to client information in a read?

The read is made on public surfaces: the company's own site and channels, the competitor set and what a buyer can find. The material a client supplies, the intake answers and the self-rating, is used for one purpose, the comparison that produces the perception delta, and is held in the engagement record under the same terms as any other client file. No client is named in anything public without written approval, and no customer, partner or pipeline fact leaves an engagement record at all. This study is that policy in practice: one engagement, no client name, no customer names and figures rounded wherever a precise one would identify someone.

The brand read landed in record time, delivered the very same day the intake arrived.

11 minutes ago

12 min read

Time to first finding: agency discovery one to three weeks, STR3 human-led audit 10 business days, nunNEO read about 20 minutes with findings the same day.

This is a direct comparison: nunNEO versus traditional human-led diagnostic work, evaluated on speed to first findings, operator hours required, cost per brand read, and the value of the results. The study covers nineteen brands from six different sectors, each counted once.

This point is worth emphasizing.

The diagnostic process is no longer the bottleneck in brand or go-to-market decisions—the decision process itself is now the slowest part.

A human-led brand audit typically takes 10 business days at STR3’s listed price, and 1 to 3 weeks according to published agency timelines. With nunNEO, the Brand Resonance Audit was completed about 20 minutes after the client submitted the intake form, and within two operator hours, it became a findings deck and a call memo—all on the same day. The initial client call focused on findings, not planning. The rest of this analysis is based on that direct comparison.

What was evaluated

On one side is the Brand Resonance Audit powered by nunNEO—a fast, data-driven assessment that measures a brand across 12 distinct dimensions and compares it to five to eight competitors. Each result comes with a confidence rating, source count, and consistency check, ensuring credibility and accuracy. Operator time is then used to translate these insights into practical recommendations for the client. On the other side is the traditional, human-led approach: STR3’s Mini Market Positioning Audit, priced at $5,500 and requiring 10 business days to complete, along with agency discovery phases that rely on workshops and five to ten stakeholder interviews.

nunNEO figures are drawn from a single engagement record. Human-led figures come from STR3’s public catalog and industry benchmarks, which provide context but are not definitive proof. Each brand in the study is counted once, using its latest

assessment.

The Brand Read Head-to-head

Time to first finding	10 business days (STR3 Mini Market Positioning Audit); 1 to 3 weeks (agency discovery)	About 20 minutes to a scored read; findings deck the same day
Operator hours per read	About 12 senior hours implied at STR3's list price; a workshop-and-interview model runs 8 to 15 hours of sessions before synthesis	About 2, to a findings deck and call memo
Cost per read, list	$5,500 (STR3); £1,800 to £7,500 mid-tier, £20,000 and up high-end (agency)	$2,500
What is read	What the interview sample recalls plus the surfaces the hours allow	Every public surface, 12 dimensions, 5 to 8 competitors, every time
Who it depends on	The consultant assigned and the stakeholders available	One rubric; system verification on every read; a senior strategist reviews only what fails a check
What arrives at the first call	A kickoff and a plan	The findings
Repeat read	A new engagement	The same read, re-run
Audit trail	Judgment in a deck	Confidence rating, source count and consistency check on every read; every figure dated with a method

The deltas, ranked

Ranked by gap size, largest first. The first three are measured; the rest are properties of the method that the record illustrates.

1	Time to first finding	10 to 15 business days	About 20 minutes to a scored read; same day to findings	At least 10x on the calendar; far more on the read itself	The diagnostic stops setting the project calendar. The first call becomes a decision call.
2	Operator hours per read	About 12 (STR3 list); 16 or more in a workshop model	About 2	6x to 8x or more	The operator's hours move from collecting to judging. One strategist carries what used to need a team.
3	Cost per read	$5,500 list; $950 at delivery effort (12 hours at the implied senior rate)	$2,500 list; about $950 in operator effort	55% lower at list; 5 to 6x lower in delivery effort	A read can be bought for a single decision. Four dated reads a year cost less than two human-led audits.
4	Repeatability	Each read is a new engagement	Re-run at will, on the same rubric	Quarterly scoring becomes practical	Brand becomes a measured variable in the operating review instead of a one-off project.
5	Consistency	Varies by consultant and interview sample	One rubric, 12 dimensions, a consistency check on every read	Comparable across brands and over time	Scores can be benchmarked and tracked; a human-led read of one brand cannot be set beside another.
6	Coverage	Bounded by hours and by who was interviewed	Every public surface and 5 to 8 competitors, every time	Breadth is no longer rationed	The finding that matters most is often on a surface nobody was paid to look at.
7	Audit trail	The deck	Confidence, sources and method on every figure	Procurement-readable	Pace stops being a reason for suspicion in a regulated market.

Findings quality: where human-led and machine reads are each weak by construction; machine breadth with system verification and human review on exceptions.

Cost per read, from three perspectives

At list. STR3's human-led Mini Market Positioning Audit is $5,500 and 10 business days. A Brand Resonance Audit on nunNEO is $2,500 and in hand the same day. Published agency audits run from £1,800 to £7,500 at the mid-tier over one to two weeks, and £20,000 and up at the high end. At list price, the read costs 55 percent less than STR3's own human-led equivalent and arrives 10 business days sooner.

In delivery effort. At STR3's implied senior rate, $5,500 buys about 12 hours. The nunNEO read used about two operator hours to reach a findings deck, or about $950 of senior time. That is a five- to six-times reduction in the cost of delivering the diagnostic, before the engine's own cost, which is small against either figure.

Per year. A brand-new read bought once a year is a project. Bought once a quarter, it is an instrument. Four nunNEO reads a year at list cost $10,000, less than two human-led audits at $11,000, and each of the four is dated, comparable with the last, and in hand the day it is asked for.

Findings quality: what the record shows and what it doesn’t

A read that arrives in a day is worth nothing if it is wrong. So this section states the evidence plainly and stops where the evidence stops.

What the record shows. In the engagement behind this study, the same-day read went to the one specification the company's buyers verify before a first call and found it published several different ways at once across the company's own surfaces. That inconsistency had stood through earlier, human-led work and through the company's own reviews; nobody owned the number. The read also scored two dimensions higher than the founder had rated himself, which is the opposite of the flattery bias a diagnostic built from the client's own interviews tends to carry. The claims discipline that followed retired four public claims and held eight pending evidence. None of those 12 had been questioned before.

Where human-led reads are weak by construction. They rest on five to 10 interviews, which means a sample of insiders and their recall. They vary with the consultant assigned. They arrive weeks after the surfaces they describe were last changed. And they tend to confirm the brief, because the people interviewed wrote it.

Where machine reads are weak by construction. The largest field experiment on general-purpose AI tools, run with 758 consultants at a global strategy firm, found more than 40 percent higher quality on tasks within the tool's competence and 19 percentage points less accuracy on tasks outside it, delivered with the same confidence. A general-purpose tool has no reason to go looking for the one figure a dental buyer checks first. nunNEO is built around that buyer, and every read runs its own verifications before releasing anything: a consistency check across sources, a confidence rating on each figure, and a source count. Where a check does not match, and only there, a senior strategist reviews the item. That is where the 19 points are caught: by a verification layer built for the task, with a human on the exceptions.

The operating model that falls out of this is machine breadth with system verification: the engine reads everything, on one rubric, checks its own findings for consistency and rates its confidence in each; a senior strategist reviews only what fails a check. On this record, the time advantage came with no accuracy penalty that anyone found. That is one record, not a benchmark. The test to run is a paired, blind comparison on five brands, a human-led audit and a nunNEO read each, scored by an independent reviewer on findings found, findings wrong, and hours spent. STR3 intends to run it and publish the result either way.

Self-rating against the machine read

Every intake asks the founder or marketing lead to rate their own brand on the dimensions the engine will score. The read is made without that self-rating, on the public surfaces and the competitor set, and only then compared with it. The result is a perception delta on each dimension: the distance between how the company sees itself and what a buyer can verify from the outside. Human-led audits rarely produce this number, because the consultant's read is built from the same interviews that carry the self-image; the two are not independent.

The delta is the finding. Where the self-rating sits above the read, the company is carrying a claim the market cannot see: a credibility and compliance exposure, and the first thing procurement or a competitor will test. Where the read sits above the self-rating, the company owns an asset it has not priced in: undersold, unpublished, or credited to someone else. Where the two agree, the dimension is calibrated and can wait.

In the engagement behind this study, two dimensions came back above the founder's self-rating, and the capability behind one of them became the positioning the sprint was built on. The other direction produced the claims register: four public claims retired and eight held pending evidence, each a place where the self-image had run ahead of what could be shown. Both halves of the map came from the same 20-minute read.

The operator's work is mechanical: rank the dimensions by gap size. Close the over-claims first, because they are a risk. Publish the undersold assets second, because they are growth sitting unused. Re-read in a quarter and watch the gaps close. For brand and GTM decisions, this replaces the question "what do we think our brand is" with "where does our self-image diverge from what a buyer can verify," which is the only version of the question a buying committee cares about.

One engagement's deltas are reported here. The processing record holds a delta on every dimension for every read across the 19 brands, which means the distribution- which dimensions healthcare founders most over-rate, and most under-rate can be published.

Where nunNEO sits against existing tools

No tool on the market does this combination, but every part of it exists somewhere, so the honest comparison is by category: what each is built to answer, at what pace and price, where published, and where its design stops.

Brand trackers on survey panels (Tracksuit, Latana, Qualtrics, Kantar, YouGov)	How aware and considered are we, and is it moving?	Classic waves run 8 to 16 weeks from field to insight; enterprise contracts	Measures movement, not the reason for it; built for consumer categories; reads no competitor copy	Healthcare B2B committees do not sit in consumer panels, and the number arrives after the quarter it describes
AI-moderated research trackers (Listen Labs)	Why did the number move?	Reports in under 24 hours by the vendor's account; pitched at enterprise and Fortune 500	Needs recruited respondents; no vertical rubric; no claims trail	Closest on pace and on explaining why; built for a different buyer and budget
B2B message-testing panels (Wynter)	How do target buyers react to this message?	Results in 12 to 48 hours from a verified B2B panel; from about $600 a test; annual plans from $20,000	Tests messages you bring; does not read your surfaces or your competitors; no healthcare lens	Complementary: the right place to validate a message once the read has told you which one to test. The human arm of the paired study
AI visibility monitors (Profound, Peec, Otterly)	Do AI answers mention us, and how?	Continuous, subscription	Presence in generated answers, not positioning	A useful signal downstream of positioning, not a diagnosis of it
General-purpose AI audits (prompted models, website-copy audit tools)	What is wrong with our copy?	Minutes; no cost to low cost	No competitor set, no vertical rubric, no confidence or source trail, no verification layer; outside its competence, 19 percentage points less accurate with the same confidence	This is the commodity the label "AI brand audit" is becoming. The rubric and the audit trail are the difference, and the paired test is how to prove it
Agency AI ecosystems in healthcare (Real Chemistry ANATOMI, EVERSANA's AI Agency)	How does the agency's own team work inside a retained engagement?	No published pace, price or method; pharma and life-sciences enterprise	Not sold as a standalone diagnostic; capability not publicly verifiable	A dental, medtech or healthtech company below enterprise scale cannot buy it, and could not audit it if it did

Against that map, the nunNEO read is different in seven ways, each of which is a line in the record rather than a claim: it is built around the healthcare buyer and the committee that reviews a vendor without taking a call; it carries a competitor set in every read; it returns in about 20 minutes and reaches a findings deck the same day; it attaches a confidence rating, a source count and a consistency check to every read; it verifies its own findings, with human review reserved for the items that fail a check; a claims register ships beside it; and it is priced for a company below enterprise scale.

Two things it does not claim:

It does not replace buyer-reaction testing; it tells you which message is worth testing.
It does not track awareness; instead, it diagnoses what awareness would be built on.

The record: 19 brands, six sectors

Each brand is counted once, by its latest read. STR3's own brands and set-up intakes are excluded. The engine is built for healthcare; most of the record is dental, medtech and healthtech. The reads outside healthcare are a useful control, because the method held where the vocabulary changed.

Dental and oral health: implants, imaging, practice software, enterprise data	7	3
Professional services, agencies and advisory	5	0
Healthtech and digital health	3	0
Healthcare services and consumer health	2	0
Biotech and pharma	1	0
Motorsport	1	0
Total	19	3

Method: a brand counts when the processing record holds a scored read for it; repeat reads of the same brand count once. Sectors are assigned from each brand's own description.

Brands read by sector: dental seven, professional services five, healthtech three, healthcare services two, biotech and pharma one, motorsport one; 19 in all.

What this means for the operator

The operator is whoever runs brand and GTM work for a healthcare company: a strategist, a fractional marketing lead, an in-house head of marketing. For that person, the comparison above changes the job in four ways.

The first conversation starts at the findings. The read is in hand about 20 minutes after the intake returns, and the findings deck the same day, so the kickoff is a decision call. In the engagement behind this study, the positioning decision was taken on pipeline evidence while the largest opportunity in the client's book was still in evaluation. A quarter later, the same work would have been a retrospective.
Hours move from collecting to judging. About two operator hours per read, against 12 or more, means a senior person spends the saved time on the decision, the architecture, and the claims, which is where senior judgment earns its rate.
Brand becomes a measured variable. A read that can be re-run quarterly on the same rubric turns brand from an annual project into a line in the operating review, with a dated score beside pipeline and churn.
Procurement gets what it asks for. Confidence, sources, and method on every figure, plus a claims register beside the brand, are the documents reviewers who never take a sales call want to see. That is deal risk removed before the deal exists.

The economics follow. Cost per read falls by more than half at list and by five to six times in delivery effort. Time to a defensible brand or GTM decision falls from weeks to the day the intake comes back, which is the quarter you get back when the decision window was open. And the standard of evidence rises at the same time, because every figure carries its date and its method. Decision velocity, measured, with the audit trail attached.

What this record doesn't show

The tool-side figures come from one engagement. The 19-brand record shows reach and method consistency, not 19 timed comparisons.
The human-led benchmarks are vendor-published and describe their own process.
Operator hours were logged after the fact from the session record; the 20-minute run time is the operator's figure for the engine's own processing.
The self-rating section reports one engagement's deltas; the cross-brand distribution is held in the record and not yet published.
No controlled accuracy benchmark exists yet; the paired test above is the next step.
Commercial results for any client are the client's to report.

Start with the Audit. str3.com/fix-brand-gtm

Sources

STR3 engagement record, one Brand Clarity Sprint; STR3 Service Catalog v2; nunNEO processing record, brands counted once by latest read.
STR3, Mini Market Positioning Audit, $5,500, 10 business days, str3.com/fix-brand-gtm.
Dell'Acqua, F., McFowland III, E., Mollick, E., Lifshitz-Assaf, H., Kellogg, K., Rajendran, S., Krayer, L., Candelon, F. and Lakhani, K. Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality. Harvard Business School Working Paper 24-013, 2023.
Major Tom. How much does branding cost? What to expect at each budget. Agency-published.
Clay. How Long Does the Branding Process Take? (2026 Timeline). Agency-published.
Inkbot Design. How Much Does a Brand Audit Cost in 2026? Vendor-published, updated Aug. 2026.
Webtonic. Marketing Audit Statistics: Cost and Pricing. Vendor-published, 2026.
Listen Labs. AI Tools for Brand Health Tracking: 2026 Decision Guide. Vendor-published; source for the 8- to 16-week tracker wave and the under-24-hour Pulse claim.
Wynter. Pricing, accessed Nov. 1, 2026: results in 12 to 48 hours; plans from $20,000 a year. Storyflow, positioning and message testing tools 2026: from about $600 a test.
Real Chemistry. Real Chemistry Launches ANATOMI, Business Wire, July 14, 2026. EVERSANA, EVERSANA and Google Cloud Bring AI Agency Concept to Life.