Do Health — NPS & PMF Report

Quantitative Overview

1,233 responses · NPS Post Blood Test · Data as of 12 May 2026

Total Responses

1,233

1,003 completed PMF

NPS Score

709P / 344Pa / 180D

Promoters

57.5%

709 respondents

Detractors

14.6%

180 respondents

PMF Score

34.1%

Very disappointed if gone

Superhuman PMF Framework: Do Health scores 34.1% on the core PMF metric (% who would be "very disappointed" if the product disappeared). The benchmark for strong product-market fit is 40%. Do Health is 5.9 percentage points below this threshold. The roadmap implication: accelerate what the "very disappointed" segment loves, and remove the blockers keeping the "somewhat disappointed" segment from converting.

Score Distribution

Detractors (0–6) Passives (7–8) Promoters (9–10)

106

238

159

550

PMF Disappointment Breakdown

Very disappointed

342 (34.1%)

Somewhat disappointed

545 (54.3%)

Not disappointed

114 (11.4%)

No longer using

2 (0.2%)

NPS × PMF Crosstab

How NPS groups map to PMF disappointment levels

PMF Segment	Promoters (9–10)	Passives (7–8)	Detractors (0–6)
Very disappointed	287	50	5
Somewhat disappointed	250	210	85
Not disappointed	17	25	72
No longer using	0	0	2

Has Do Health Helped You Make Meaningful Changes?

New metric in this wave — 1,233 respondents

46.7%

Yes — some changes
576 respondents

13.9%

Yes — significant changes
172 respondents

19.1%

Not yet
235 respondents

1.3%

No
16 respondents

60.6% of respondents report making at least some meaningful changes to their health habits — a strong early signal of behaviour impact. The "not yet" cohort (19%) is consistent with the early-journey profile: users who have had their first blood test but haven't yet reached their re-test window.

Strengths — What to Double Down On

Consistent themes across every NPS and PMF segment

Blood test results and biomarker insights are the undisputed core of Do Health's value. Joy ranks second across all question types. These two assets — blood insights and Joy — are what make Do Health meaningfully different from a standard health app or GP visit. Both deserve continued investment, not complacency.

Most Valuable So Far — Top Themes

Blood results

486

Motivation

Joy coach

In control

Bio age

Weekly plans

Based on 709 open-text responses

Joy AI Coach — Signal Summary

Positive mentions (open text) ~80%

Mixed / negative mentions ~20%

123 open-text responses mention Joy. Users who love Joy describe it as contextual, non-judgmental, and unlike other AI they've used. Users who are mixed find it too chatty, limited, or occasionally inconsistent with blood result guidance.

"Joy is one of the best AI tools I've used so far."

"Joy is aware of all my stats and tailors answers to my queries."

Strengths in Full

Ranked by mention volume across all question types

Blood test results and biomarker insights

The single most mentioned theme — 486 mentions in "most valuable so far" alone. Users are consistently surprised by 24–48hr turnaround. Plain-language explanations of what each marker means, and what to do about it, are called out repeatedly as something the NHS has never provided. Access to markers unavailable via NHS (ApoB, homocysteine, insulin) is a key acquisition reason.

486 mentions

Joy AI coach

Second most mentioned across all segments and question types. 54 direct mentions in "most valuable", 123 across open text. Users value Joy because it feels personalised, remembers their context, and is accessible without judgment. Users who normally dislike AI call this one different. Joy is the primary reason users feel the product is worth the price.

123+ mentions

Motivation to improve habits

81 mentions. Users describe the accountability loop — selecting tasks, logging them, knowing a re-test is coming — as a structure they haven't found elsewhere. The programme drives genuine behaviour change, not just information delivery. 60.6% of respondents in this wave report making at least some meaningful changes.

81 mentions

Feeling in control and empowered

45 mentions. The dominant emotional outcome users report is control and confidence, not just information. "I feel in control of my health for the first time." "I don't feel passive anymore." This is the core transformation the product delivers, and users are describing it unprompted. Strong positioning and retention signal.

45 mentions

Progress tracking and repeat testing anticipation

Over-indexed in the "very disappointed" PMF segment — users who came with strong motivation but haven't yet seen the feedback loop close. The re-test cadence is creating forward momentum and anticipation. This retention hook exists; the product isn't yet making it visible or rewarding enough early in the journey.

prominent in "very disappointed" segment

Biological age tracking

20 mentions. A smaller but meaningful signal. The longevity framing is motivating and distinctive. Worth protecting as both an acquisition hook and a retention driver as the product matures. Particularly resonant with the core demographic.

20 mentions

Opportunities — Blockers to Remove

888 responses to "biggest improvement" · 744 open-text responses

One signal dominates everything: users feel the app fails to deliver on its core promise of personalised guidance. This is not a feature request — it is a trust problem. Users can see their blood results but cannot connect them to the habits the app recommends. The product decision is not to add more content, but to rebuild the logic layer connecting test results to recommendations.

Biggest Improvement — Top Themes (888 responses)

Personalisation

252

Progress tracking

177

Clinic access

158

More biomarkers

Results clarity

Price/frequency

Open Text — Additional Blockers (744 responses)

Joy (mixed)

123

Progress tracking

Reminders/nudges

Result guidance

Not personalised

Clinic logistics

Backdate tasks

Blocker Deep Dive

Ordered by impact on PMF gap

#1 BLOCKER

Personalisation gap — 252 mentions

The top blocker by a wide margin. Users see their blood results but cannot connect them to the habits the app recommends. Weekly plans feel generic and repetitive. The product has no visible logic layer linking results to recommendations, making coaching feel arbitrary rather than evidence-based.

"The personalised plan is repetitive and doesn't feel personalised at all."

"The assessment of my bloods is meant to be personalised, except it isn't."

"Go for a walk, eat oily fish — it's just generic advice with my name on it."

#2 BLOCKER

Progress tracking — 177 mentions

Users have no reliable way to see whether what they're doing is working. Without visible progress, there is no feedback loop — users complete habits but feel nothing is accumulating. This is both a motivation problem and a data presentation problem. Especially acute in the "very disappointed" PMF cohort, making it a retention lever, not a nice-to-have.

"I need to be able to see if my actions are actually making a difference."

"I want to see my progress tracked more visibly over time."

#3 BLOCKER

Clinic access and logistics — 158 mentions

Clinic location and appointment availability affect overall satisfaction, particularly for the "somewhat disappointed" PMF segment. This is an operational priority but should not displace investment in personalisation and engagement — it drives friction, not core dissatisfaction or churn.

"Very disappointed I had to travel 25 miles for the nearest clinic."

Detractor Deep Dive

180 detractors (14.6%) — sharper version of the same problems

Detractors are more likely to describe blood testing as the only strong component, with everything else feeling like an underbuilt wrapper. This framing is a strategic risk: the blood test is the acquisition hook, but the product currently has no strong retention hook to follow it.

What Detractors Say

"The blood tests are helpful and convenient but everything else is poor."

"The plans and AI tool are pretty thin. It's only worth it for the bloods."

"Just 3 blood tests with not much else. It could be so much more."

"Joy is not reliable and even conflicts with blood work recommendations."

Detractor-Specific Signal

Detractors explicitly struggle to understand how blood results translate into recommended actions — more acutely than the broader user base. Unclear result interpretation is a churn driver, not just a friction point.

The 72 detractors who are "not disappointed" PMF (they won't miss the product) represent the cleanest churn risk. These users tried the product, found blood tests useful, but saw no reason to stay.

Three Product Decisions

Ordered by impact on the PMF gap. Data from 1,233 responses, May 2026.

These three decisions follow directly from the data and are consistent with the previous wave. The same problems showing up as the user base grows means they are structural, not cohort noise. The order reflects where the PMF gap is most likely to close fastest.

Rebuild the result-to-recommendation connection

Personalisation gap is the top improvement request by a 42% margin over the second-ranked theme. Users can see their markers but cannot see why the app is suggesting what it's suggesting. The fix is not more content — it is making the logic layer visible. Every recommended action should show which of the user's flagged biomarkers it targets. This turns "go for a walk" into "this targets your HbA1c and triglyceride results." One change. Highest-leverage area in the product.

252 mentions

Top blocker · All segments

Trust problem, not feature gap

Make progress visible and emotionally rewarding earlier

Progress tracking is the second-biggest blocker and over-indexes in the "very disappointed" PMF segment — the cohort closest to converting if the product delivers. Design a longitudinal view that connects task completion to biomarker trend. Add a before/after delta at BT2 that surfaces clearly, celebrates improvement, and explains flat or negative movement. This is the feedback loop the product currently lacks. Without it, even motivated users cannot tell if the programme is working.

177 mentions

Retention lever · "Very disappointed" segment

BT2 unlock

Ship reminders and accountability infrastructure

Reminders and notifications appear across 71 open-text responses. Users forget to open the app. There is no nudge layer. This is relatively low-complexity compared to personalisation but has a direct impact on habit formation and weekly engagement. Ship this while the deeper recommendation logic is being rebuilt. The B=MAP model (BJ Fogg) in Do Health's own principles document says the Prompt is as important as Ability — the product currently has no prompt layer at all.

71 mentions

Low complexity · High retention impact

Ship in parallel

Wave Comparison

May 2026 vs. previous wave

Metric	Previous Wave	May 2026	Change
Total responses	1,061	1,233	+172 ↑
NPS Score	43.0	43	Flat →
Promoters	57%	57.5%	+0.5pt ↑
Detractors	14%	14.6%	+0.6pt ↑
PMF — Very disappointed	34.6%	34.1%	-0.5pt ↓
Gap to 40% PMF threshold	-5.4pt	-5.9pt	Widening ↓
Behaviour change (new metric)	—	60.6%	New ✦

The signal: NPS is flat — consistent at 43 as the user base grows, which means quality is holding. The PMF gap is slightly wider, driven by the "somewhat disappointed" cohort not yet converting. The three product decisions above are the clearest path to closing it. The same themes appearing in both waves confirms these are structural problems, not noise.