Product Analytics · H1 2026 · Members only

Member engagement & conversion

Engagement depth, retention, and reservation conversion for January–June 2026.

Jan – Jun 2026Source · Mixpanel + FirebaseStaff activity excluded

Full analytics instrumentation was rolled out over the past year, so this report covers the fully-tracked Jan–Jun 2026 window.
All app metrics exclude staff/provider activity — 745,753 team_member events (~12%) removed. Definitions on the next slide.

Definitions

What each metric means

Active member

A patient account (user_type = member) with at least one in-app event in the month. Staff/provider excluded. One caregiver may hold more than one account (e.g., a parent and child).

Returning member

A member active in 2 or more distinct months of the window.

Month-over-month basis

Trends use chat only — the one event tracked across all six months. Other events were switched on through Q1 2026, so total-event counts can't be trended.

Chat session

One chat screen opened until the member navigates away. Staff/agent sessions are excluded.

Usage / time on screen

Active foreground seconds on a screen; the timer stops when the app is backgrounded. Sub-10s glances excluded.

Reservation stages

Started = step 1 (opens flow). Intake submitted = step 4 (name/email/phone). Completed = step 6 (booking confirmed).

Every figure in this deck uses these definitions. Started is not the same as completed; member is not the same as staff.

Membership · billing system of record

Growing — and an increasingly returning base

Active members at period-end — returning vs new that year

8,020

2023

18,972

2024

28,359

2025

30,643

2026*

Returning members New that year

25,071

returning members in 2026 — up from 161 in 2023

~98%

monthly membership retention, sustained since 2023

~12K

new members added per year — steady acquisition

77%

of active members are on annual plans

Server-side billing records. The returning base compounded from ~160 to 25,000 while new sign-ups held near 12,000 a year — the base is now 82% returning, on 35,757 total members (30,560 paying).

Staxbill billing records via Metabase (refreshed Jun 15 2026; retention export Jul 7 2026). New = joined that calendar year; returning = active at period-end who joined earlier. 2026* is year-to-date (June), so its new-member segment is a half-year.

Care delivery · clinical records

More care delivered — every year

Appointments completed · Jan–Jun of each year

9,912

2023

39,195

2024

59,400

2025

71,807

2026

Lab results delivered · Jan–Jun of each year

9,699

2023

54,239

2024

75,466

2025

125,636

2026

Same-window comparisons (January–June of each year) from server-side records, independent of app tracking. Appointments +21% and lab results +66% year over year — with appointments now running ~12,000 a month and lab results ~23,000.

Clinical database counts, monthly aggregates. Appointments = completed visits (cancellations/no-shows excluded), all channels. Lab records begin Feb 2022.

Care delivery · what members book & test for

The care mix — and a growing telehealth channel

Appointments completed by type · Jan–Jun of each year

Type	2023	2024	2025	2026
Sick visits	6,506	17,028	22,794	24,596
Telehealth	151	5,800	12,816	20,749
Well & newborn	2,659	6,943	9,891	12,867
Follow-ups	62	4,251	5,199	7,045
Wellness / IV	0	828	2,149	2,641
Immunizations	245	548	959	1,477
Weight loss	0	2,281	3,283	353
Other	289	1,516	2,309	2,079
All types	9,912	39,195	59,400	71,807

Most-ordered lab panels · H1 2026

Metabolic (CMP)

11,164

Blood count (CBC)

10,918

Lipid panel

9,688

HbA1c (diabetes)

9,446

Thyroid (TSH/T4)

8,964

Vitamin D

8,345

Telehealth went from 151 visits to 20,749 — now 29% of all care delivered. It barely existed in 2023; today it's the second-largest visit type. Lab ordering follows a preventive-care pattern led by metabolic, blood-count and lipid panels.

Clinical database — completed visits by type, Jan–Jun of each year (same window as the previous slide, so the totals match). Lab panels are the top 6 of ~1,289 types and count individual tests — one visit yields several — so they don't sum to the "results delivered" figure.

App adoption

Most members use the app — and keep coming back

1 in 3

of Baker's members are active in the app every month (~11,000)

2 in 3

used chat during H1 — 22,806 unique members

71%

of app members return in a later month

2,082

were active every single month of the half

App engagement is both broad and sticky: a third of the whole membership opens the app monthly, two-thirds message the care team, and most keep returning. Measured on chat — the one event tracked across all six months.

Mixpanel, members only, Jan–Jun 2026. Denominator is Baker's 35,757 total members (billing). 34% of chatters were active in 4+ of the 6 months.

In-app

Usage by screen

True usage by screen — quick glances excluded

Screen	Members	≥15s (true use)	≥30s	≥60s	≥5 min
Chat	22,806	21,194	20,173	17,310	5,488
Home / Newsletter	26,170	20,296	16,684	13,062	6,538
Reservation flow	15,320	10,389	8,892	6,951	2,549
Lab*	14,003	8,056	7,421	6,418	2,560
Studio	9,541	3,676	2,834	1,982	651
Your Health Record	4,472	2,383	1,516	825	206

Unique members (patient account); "true use" = a visit of 15s or longer. Home and Newsletter are the same event (the Home tab is the newsletter) and are merged; Yoga was the pre-launch tag for Studio and is merged into it. Screen timers only finalize when a member switches tabs — not when they leave the app — so these are minimums. *Lab is measured differently: because members check results and close the app, its timer often never fires — so Lab time is derived from the gap between a member's first and last tap inside Lab (tab, category card, PDF, PhenoAge). That's immune to the timer issue but excludes reading after the final tap, making it a conservative floor.

Core surface

Chat · members only

Chat usage — 1.72M member sessions

Member chat session length

0–30s

85%

30–60s

8%

60–90s

3%

90s+

4%

22,806

unique members chatting

41

median sessions / member

1.72M

member sessions (staff removed)

62,980

sessions over 90s

Members generate about 287,000 chat sessions per month; 85% are under 30 seconds. Staff/agent chat activity (745,000 events) is tracked separately and excluded here.

Channel Session Ended, member-side only, Jan–Jun. A session = one opened→closed chat channel. Top chat topics pending a message-category export.

Feature

Lab

Lab usage — top panels and depth

Most-clicked Lab categories (members)

Annual Physical

12,757

Cholesterol & Heart

6,003

Blood & Genetics

5,567

Vitamins & Minerals

5,536

Kidney Health

4,585

Thyroid & Endocrine

4,365

14,003

members opened Lab — ~2 in 5 of all members

~2 min

typical Lab visit; 6,418 members spent 60s+ (floor)

125,636

lab results delivered in H1 2026 (+66% YoY, clinical records)

1,748

used PhenoAge · 88% got a result

14,003 members opened Lab — roughly 2 in 5 of all members — with 77,110 opens in total. At least 6,418 spent a minute or more and 2,560 spent five minutes-plus reading results; the typical Lab visit runs about two minutes. Members go straight to named panels rather than browsing.

Button Click events on Lab screens, members only, Mar–Jun (Lab dwell tracking began March). Category = the lab panel a member opened; 77,110 Lab-tab opens preceded these.

Content

Newsletter

The Home tab is the newsletter — and members read it

Members by deepest read (single visit)

Opened (any)

26,170

Read ≥ 15s (true)

20,296

Read ≥ 30s

16,684

Read ≥ 60s

13,062

Read ≥ 5 min

6,538

The Home tab is essentially the newsletter, so time on Home is time with the newsletter open — the two are one surface, not two. 26,170 members opened it and 6,538 stayed five minutes or more. Sub-5-second U-turns excluded; counts are minimums.

Home / Newsletter dwell, members only, Mar–Jun (event added in Q1). Per engineering, "Newsletter Duration" and "Home Screen Dwell" are the same measurement because the Home tab contains the newsletter — they're merged here rather than double-counted. The timer finalizes on tab-change, so these are minimums. Per-article link tracking would need a release.

On the web

Reservations

Web reservations — the new-patient funnel

Started → intake → completed

Flow started2,486

Intake submitted22%546

Completed14%351

This is the new-patient channel — existing members book in the app. Bookings are pediatrics-heavy: new-patient well visits, sick visits, newborns and telehealth. Intake (step 4) = name/email/phone; completed (step 6) = confirmed. Studio classes book through a separate Webflow flow, and HRT is low-volume, so the funnel is shown in aggregate.

reservations.bakerhealth.com production, completed Apr 28 – Jun 30 2026. Consistent with the team's web-reservation funnel analysis; per-type completion rates aren't broken out because the completed-event type field isn't reliable.

Summary & next data

Summary

Growth & retention

Growing and loyal

Members grew every year — 30,643 active; returning base 161 → 25,071
~98% monthly retention; ~12K new members/year
App: 1 in 3 members active monthly; 71% return
22,806 chatted (~2 in 3); median 41 sessions

Conversion & features

Usage and bookings

71.8K appointments (+21%) & 125.6K lab results (+66%) in H1, YoY
2,486 new-patient web reservations started, 351 completed (14%)
14,003 Lab users in-app (~2 in 5); 1,748 used PhenoAge
6,538 members spent 5 min+ on the Home/newsletter surface

What's next

In development

AI chat assistant — Phases 1 & 2 target automating ~25–29% of member chats
Newsletter link-level tracking
Cohort retention reporting

Sources: Mixpanel raw events (app + web reservations), Jan–Jun 2026, staff excluded; Firebase for engagement totals. Definitions on slide 2.