A real Kula Intelligence output, anonymized. This is an Owners' Report produced by a paid Kula skill (Studio Accelerate — Phase A Integrity, then Phase B Yield), reproduced from a live two-location studio. Shown here as 'Saltwater Yoga'; owner, teacher and competitor names and all figures are changed. The analysis, the structure and the recommendations are the real deliverable.
Studio Accelerate · Phase A — Integrity

Saltwater Yoga — Owners' Report

What your two years of data actually says — and the decisions it puts in front of you.

Prepared for Dana & Marco · June 2026

Window Jan 2024 – May 2026 (29 months) Studios Santa Monica · Venice Beach · Online Basis verified to your raw exports

01The headline: steady, not growing

Revenue has held flat for two years.

The business turns over about $83,000 a month in organic revenue (excluding the ClassPass visits that reimburse separately). That number has barely moved in 29 months — no upward trend, no decline, just a stable line with two one-off spikes. 🟢

The "+5% versus last year" you might see in a quick look is misleading: it's almost entirely one month — June 2025 — which was a teacher-training intake, not broad growth. Strip that out and the year is flat. 🟡

Monthly organic revenueAverage ~$83k

02The real story is underneath: your two studios are diverging

Flat at the top hides one studio cooling and another heating up.

Most of your membership money is billed online, so the raw location report is misleading — it makes both rooms look small. When we push every online dollar back to the studio where that member actually trains, the true picture appears: 🟢

$1.64M 69%
Santa Monica · down 20%
$665k 28%
Venice Beach · up 93%

Santa Monica is still your engine — roughly 2.3 times the size of Venice Beach — but it's slowly cooling, down about 20% across the window. Venice Beach is genuinely growing, nearly doubling. The two move almost as mirror images, and that's exactly why the top line looks flat: Venice Beach's growth has been quietly covering Santa Monica's decline. The gap is narrowing, but Venice Beach has not overtaken Santa Monica. 🟢

Santa Monica (true)Venice Beach (true)
The most important sentence in this report: the business isn't standing still — one studio is handing growth to the other. That's a very different problem to manage than "nothing is moving," and it's invisible unless you split the two rooms properly.

03Membership is slowly leaking — and it's a Santa Monica leak

You're working hard just to stand still on memberships, and the losses are concentrated.

Your active autopay membership base peaked at 282 in late 2024 and has eased to about 205 now — down roughly 10%, with the drop accelerating through 2026. Across the whole period you signed up almost as many members as you lost: a treadmill, not growth. 🟢

By year: 2024 added 56 net members; 2025 lost 16; 2026 so far has lost 63. And when we split it by studio, the leak is clearly Santa Monica — cancellations there have run roughly double new sign-ups through 2026. Venice Beach's membership flow is small and close to balanced. 🟢

One event you should know about. When the membership plans were re-platformed to the new "Direct Debit" products in May 2025, about 68 of 312 members (22%) did not carry across. That single changeover explains a large part of the mid-2025 drop. Some genuinely left; some may simply have slipped through the cutover. It's worth knowing how many of those you could win back. 🟡

04Teacher training and cross-studio members

Two smaller facts worth holding.

Teacher training is small but lumpy. It brought in about $53,400 over the two years — only ~2% of revenue — earned in a handful of intake months. The big June 2025 month that flatters the annual figure was mostly a 200-hour intake (~$19k), not underlying growth. 🟢

About one member in six trains at both studios (15–17%, easing slightly). Most of your members are loyal to one room. That matters for what's coming next: if Santa Monica members leave, Venice Beach can't simply absorb them — they're closer to two separate communities than one shared pool. 🟢

05The quarterly scorecard

Everything above, quarter by quarter, by studio.

QuarterStudioRevenueNewCancelNetSessionsBoth-studio %
2024-Q1Santa Monica$164k180*15+165
Venice Beach$38k46*5+41
2024-Q2Santa Monica$193k6328+35
Venice Beach$52k206+14
2024-Q3Santa Monica$181k5638+1855216.9%
Venice Beach$59k2118+3296
2024-Q4Santa Monica$209k4447−351916.8%
Venice Beach$48k1412+2269
2025-Q1Santa Monica$198k4645+152815.9%
Venice Beach$61k14140284
2025-Q2Santa Monica$195k4159−1853916.4%
Venice Beach$96k2316+7287
2025-Q3Santa Monica$151k3063−3356815.9%
Venice Beach$86k1119−8292
2025-Q4Santa Monica$141k7758+1954715.9%
Venice Beach$96k2017+3276
2026-Q1Santa Monica$134k2979−5053215.6%
Venice Beach$78k1428−14288
2026-Q2†Santa Monica$110k2532−735614.3%
Venice Beach$67k1213−1189

Revenue is the true-studio view (online re-attributed by where members train, training income excluded). "Both-studio %" is members training at each studio that quarter. * 2024-Q1 sign-ups include everyone already a member when the data starts, not a quarter of new joins. † 2026-Q2 is April–May only, so its totals are partial. Sessions and both-studio data begin Q3 2024, when class attendance tracking starts.

Read across the Santa Monica row and the pattern is stark: roughly the same number of classes every quarter (~520–570), but revenue sliding from ~$209k to ~$134k. Same room, same cost to run it, fewer paying bodies. Venice Beach is the reverse — flat class count, revenue roughly doubling. 🟢

06New on your doorstep: Method Lab

A well-funded competitor opened in Santa Monica last month.

Method Lab opened at 1431 Montana Avenue, Santa Monica on 18 May 2026 — reformer, heated mat, yoga, their own teacher training, intro offers from $50. 🟢

Two honest points. First, it cannot be blamed for your Santa Monica decline — that's been running for over a year, well before they opened. The Santa Monica problem is yours, and it's internal. Second, it makes that problem more urgent: a fresh, free-trial-led competitor is now pulling the same Westside catchment, arriving exactly when your strongest studio is already soft. The upside is you have a clean "before" baseline — we can measure any Method Lab impact cleanly against the trend from here. 🟡

07Why members actually leave

The raw cancellation count overstates real churn by about a quarter — and the leak is Santa Monica.

We've now built the cancellation report. We don't yet have Mindbody's recorded reason field, so the reasons here are inferred from each member's membership type and the timing of their last bill — directionally sound, not final. With that caveat, the picture is clearer and less alarming than the headline number. 🟡

Of 700 membership exits across the window, roughly three in four (494) are genuine stops, but nearly one in four (162, 23%) are simply fixed-term passes reaching their end date — a renewal job, not a member walking out. Those expiries jump from low single digits to 26–27 a quarter from mid-2025, tracking your mid-year sale push exactly. Much of the scary-looking late-2025 and early-2026 churn is sale passes ending, not people quitting. Strip them out and true voluntary churn is materially lower than the raw count. 🟢

Where it does bite, it's Santa Monica. Of the ongoing stops we can place to a studio, 76% are Santa Monica (328 vs Venice Beach's 103). The same room that's losing revenue and now faces Method Lab across the road is also where members are actively leaving. The worst single quarter is 2026-Q1 — 81 ongoing stops, the highest in the data and not explained by expiries. That's the quarter to watch as fresh data lands. 🟢

Two honest limits. The 494 "ongoing stops" is still a ceiling, not a count of quits — it blends voluntary leavers, failed payments and freezes, which only Mindbody's reason field can separate. And the May-2025 migration loss is understated here (18 tagged versus the ~68 the carryover analysis found), meaning true voluntary churn is likely lower still. Pulling the MBO cancellation export remains the one move that turns this section fully green and tells you, for Santa Monica's 328 stops, how many actively quit versus dropped for billing — the difference between a retention fix and a billing fix. 🟡

08The decisions in front of you

Seven moves, ranked by speed and certainty — not by headline size.

The strategic frame underneath all of them is the same: defend Santa Monica. It's 69% of revenue, it's been softening for over a year, and Method Lab is now across the road. The open question worth settling alongside it — is Venice Beach's growth net-new members or Santa Monica members drifting across? If it's net-new, feed it; if it's internal migration, your flat top line is a warning, not a comfort. 🟡 Every move below pulls toward that frame. The first three cost almost nothing and the demand is already measured.

Grow what's already winning

1. Add a sibling to your two ceiling slots.

Santa Monica Sat 8am turns away 22 people; Sun 4pm turns away 36 — measured waitlist, not estimate. Those classes clear $506–552 a sitting; a new early-Saturday or Sunday-afternoon class costs $97–112 to staff. This is the cheapest growth on the board: the demand is proven, you're simply out of seats. Put a proven draw on it before you touch anything else. 🟢

2. Reassign three Santa Monica slots — don't re-hire.

Mon 7:30am, Thu 4pm and the Saturday marquee each pair a softening teacher with a stronger mover who lifts +4 to +8 heads in the same room. Swapping is near-zero cost and loses almost no regulars. Your busiest classes are already won; the lever is putting your best movers on the viable-but-underfilled slots, not on the ones already full. 🟡

Stop the leaks

3. Close the day-28 cliff.

Two in three new starters fall at or before month one, and the intro-pass fall lands the exact week the 28 days expire. You can see who's convertible before then — future regulars train roughly twice as often in their first fortnight. This is the highest-value fix in the business and the one Whispers is built for. The Phase B appendix models what moving it 15% is worth. 🟢

4. Re-engage the unused class passes.

858 passes still carry unused sessions; 251 were never opened once. These people already paid and never got the value — your warmest, cheapest re-engagement list, and the natural bridge from a casual pass into a habit. 🟢

Tighten the cost base

5. Consolidate the Venice Beach dawn doubles.

Mon/Wed/Thu run two back-to-back 6:00 + 7:15am classes into a half-empty 20-cap room — 7 to 8 bodies each — while the same times fill on Tue/Fri. Demand supports one class on those days, not two. Collapsing each to a single morning takes about $15k a year of teacher pay out of classes that barely clear their own cost, with only a handful of regulars to re-home. 🟢

6. Realign pay to contribution.

You pay a flat rate across a four-fold spread in what teachers actually draw — and your two best value-per-dollar teachers sit on the bottom rate. Not a cut: a realignment that rewards the movers filling your rooms. Design this deliberately before it becomes a retention risk. 🟡

Unlock the rest

7. Pull the two exports that turn guesses into numbers.

The MBO cancellation export turns section 07 green — splitting genuine quits from failed payments, freezes and sale-pass expiries, so a retention fix isn't mistaken for a billing one. The ClassPass remittance prices the 27% of your seats currently showing $0 — one in four bodies in the room. Both are admin pulls; both convert a guess into a decision. 🔴

How to read the confidence marks:  🟢 we'd stake the number on it  ·  🟡 direction is sound, a detail needs your eye  ·  🔴 we need data from you before standing behind it.

Prepared by Kula for Saltwater Yoga · Studio Accelerate, Phase A (Integrity). All figures reconcile to your Mindbody exports; ClassPass reimbursed visits and teacher-training income are separated from studio revenue. Studio revenue is allocated to the room where members train, not where payments are billed. Data runs to 31 May 2026. This report shows what is happening; it deliberately contains no projected savings or revenue forecasts — those belong to the next phase, only once you've sniff-tested everything here. More human, not less.

Studio Accelerate · Phase B — Yield (Scenario)

The intro-pass engine, and what moving it is worth

Phase A above is what is. Everything below this line is what could be — a scenario, not a promise.

Two things sit in this appendix. First, an interactive view of your intro-pass cohorts, month by month — the actual survival curves from your data, so Dana can see which intakes held and which fell apart. That part is real: every dot traces to Mindbody. 🟢

Second, a model of what a 15% lift does — 15% more intro passes sold, 15% more people surviving the month-one cliff, and how that ripples through the whole curve into revenue. The ladder of value it runs on is real (your own realised dollars per starter). The 15% itself is an assumption — an input you can drag, not a result we're banking. Treat the numbers as "if we move this, here's the arithmetic," not "we'll save you this." 🟡

B1Your intro cohorts, month by month

Pick a starting month. This is the share of that intake still attending, period by period.

Nine in ten new people arrive on a casual class or the 28-day intro, and the intro pass is the one route that holds: intro starters stick at roughly two to four times the rate of drop-ins at every point. But the cohorts are not equal — some months convert two in five past the cliff, others barely one in eight. The pattern is the asset. 🟢

Intro cohort explorer

Tap a month. Solid line = fully observable. Greyed months are still maturing (not enough time has passed to see the full year yet). "Pooled" = Jan–May 2025 combined, the five intakes with a complete 12-month run.

Selected cohort — % of starters still attending Pooled Jan–May 2025 reference

"Start" is the first 30 days (on the intro pass). "M1" is the first month after the 28-day pass expires — the cliff. Past M1, anyone still attending is on a paid plan, i.e. they converted. Source: stage3 cohort tables, reconciled to Mindbody. 🟢

B2What a 15% lift is worth

Same engine, two dials: sell more passes, and lose fewer at the cliff.

This runs on a real ladder. We took your 209 intro starters from Jan–May 2025 and measured what each actually paid you over their first year, grouped by how long they stayed. Someone lost at the cliff was worth about $110. Someone who locked in past month eight was worth about $1,236 — roughly eleven times more. That gap is the whole game: retention isn't a softer metric than acquisition, it's where the money is. 🟢

Drag the dials. The left number is the model's headline. 🟡

The 15% scenario

Defaults are set to 15%. The ladder and curve are real; the percentage lifts are your assumption.

+15%
+15%
Extra realised revenue / year
Extra per 100 intro starters
Extra members reaching lock-in / yr
Today — members still attending, per 100 starters With your dials — pass lifts the whole line, cliff lifts the tail

Baseline = 428 intro starters/yr (your 2025 intake) × the realised value ladder above. The cliff dial moves people out of the $110 tier into the retained tiers in today's proportions; the pass dial scales the whole intake. Independent levers, multiplied. 🟡

The honest version in one line: your intro pass already works — it's the only door that holds. The leak is the cliff at day 28, and that's the cheapest member you'll ever keep. Phase A says the cliff is real. This says what closing even part of it is worth. The next page says how.

B3Intro pass: should it be shorter or cheaper?

Tested: could a 3-week ($80) or 2-week ($51) pass hold the same conversion? The data says don't shorten it — but it hands you something better.

The instinct is reasonable: if the convertible behaviour happens early, trim the dead back end and drop the price. The data contradicts it. The members who convert are precisely the ones still attending in weeks 3 and 4. Non-converters fade from week one; converters barely decline. 🟢

Who's still in the room, week by week

Share of intro starters attending in each week of the pass, split by whether they went on to convert. Jan 2025–Feb 2026, n=521, 36% converted.

Converted — still attending Didn't convert
88%
of converters still attend in weeks 3–4
78%
of converters still attend in week 4
2%
convert after a silent first fortnight

Cut the pass to two weeks and you remove the room from ~88% of your future members while they're still deciding. Almost nobody starts late, so you catch no hidden segment by trimming. The back half is cheap to provide — non-converters have already faded — and expensive to remove. 🟢

The same data hands you the real lever: the early-intensity signal. By day 14, 91% of eventual converters have already taken 2+ classes, and conversion climbs cleanly with first-fortnight visits: 🟢

Conversion by visits in the first two weeks

You can see who's converting before the pass is half done. So the move isn't a shorter pass — it's a day-14 conversion ask to the high-intensity group: keep the full 28-day runway, act on the signal early. Same outcome you wanted (faster decision, less dead time), none of the downside.

The recommendation: decouple price from length. Length protects conversion — keep the 28 days. Price is your volume lever — if you want to test cheaper, test $80 or $51 at the full 28 days, not a shorter pass. And whatever you test, run it as a measured A/B on one studio for a quarter and watch conversion, not sales — because a shorter pass might compress visits earlier through urgency, and history can't prove that either way. 🟡

B4How you close the cliff: Whispers

The right nudge, to the right person, at the moment they're about to drift — sent like a human would, because behind it there is one.

A Whisper is a small, timed, personal contact triggered by something real in the data — a milestone hit, a gap in attendance, a pattern only the system can see across thousands of bookings. It is not a blast. Each one is aimed at one person and reads like their teacher noticed. More human, not less. 🟡

1 · AN OCCURRENCE Something happens in the data — a milestone, or a silence. 2 · THE STUDIO NOTICES The system flags it and drafts a Whisper. A human approves the tone. 4 · THE MEMBER STAYS They feel seen, not sold to. They come back — past the cliff. 3 · THE WHISPER — THREE KINDS MILESTONE — celebrate Trigger: a number worth marking. "That's 25 classes, Elena — quietly one of our most consistent. Proud of you." SURGICAL — re-engage Trigger: 10 days unseen, SMS / app. "Haven't seen you in a little while — your Thursday 7am spot's still yours." AFFINITY — the teacher From the instructor they see most. "Hey, it's Theo — hoping to see you Thursday 7am. Saved you a mat." ONE MORE VISIT, AT THE RIGHT MOMENT A member who returns the week after the cliff is the one who climbs from $110 toward $1,236.

Milestone whispers build the relationship.

They cost nothing and they bank goodwill before you ever need it. "Well done on 25 classes" tells a member the studio is counting — quietly, on their side. The affinity instructor is the natural sender.

Surgical whispers stop the drift.

The system sees the gap a human front desk never could — ten days unseen, a missed regular slot. One precise SMS or app nudge, naming their actual class, lands before the habit breaks. This is the one aimed straight at the cliff.

The affinity follow-up closes it.

A second touch from the teacher they train with most turns a reminder into a relationship. Most-visits-with is in your data already; the whisper just routes it to the right voice.

The cliff playbook — four Whispers, build-ready

Each one below is specified the way it would be deployed: a precise trigger your data already supports, a window, a sender, a channel, the message, and the guardrail. The split happens at day 14 — the signal tells you whether a starter is warming or cooling, and they get opposite Whispers.

2aThe warming starter — open the conversion

Trigger intro day 12–14 AND ≥2 signed-in visits in first 14 days  ·  Window day 12–14  ·  Sender studio (warm, branded)  ·  Channel member app, fall back to SMS

"You're three classes in and clearly finding your rhythm — love to see it. When you're ready, here's what carrying on looks like after your intro. No rush, your pass is still running."

Guardrail. An opener, not the hard ask — that's the teacher's job at step 3. Names their actual count. Never discount-led. Suppress if a membership is already booked. 🟢 trigger is fully in your data (visit timing predicts conversion 16%→47%).

2bThe cooling starter — surgical re-engage

Trigger intro day 10–14 AND 0 visits in last 7 days AND total visits ≤1  ·  Window day 10–14, before the habit's cold  ·  Sender studio  ·  Channel SMS (highest open), app mirror

"We haven't seen you in a little while — and your intro still has plenty left on it. There's a spot in [their first class / a beginner-friendly Thursday 7am] this week with your name on it. Want us to hold it?"

Guardrail. Help, not guilt. Offers a specific, easy next class — booking friction is the enemy. One touch only; if no response in 5 days, route to the day-25 continuity Whisper, not a second nag. 🟢 trigger measurable; effect size unproven here — pilot it.

3The affinity-teacher ask — close it

Trigger intro day 18–24 AND ≥3 visits AND still attending (visit in last 7 days)  ·  Window day 18–24, while still in the room  ·  Sender the instructor they've attended most (most-visits-with)  ·  Channel personal — app/SMS in the teacher's name

"Hey, it's [Theo] — it's been so good having you in class these few weeks. Your intro wraps up soon; I'd love for you to stay with us. Happy to talk through which membership fits how you like to train — just say the word."

Guardrail. This is the highest-leverage touch you have — keep it human and one-to-one, never a template blast. Sender is computed from attendance, not guessed. Relationship-led, not price-led. 🟡 mechanism strong (Stage 5/6 affinity), effect to be measured.

4Kill the gap — the continuity bridge

Trigger intro day 25–28, pass about to expire, no membership booked  ·  Window day 25–28  ·  Sender studio  ·  Channel app + SMS

"Your 28 days are nearly up — but your momentum doesn't have to be. Roll straight onto membership and keep your classes without missing a week. One tap to continue, or reply and we'll set it up with you."

Guardrail. Make continuity the default path, but the member must take the action — no silent auto-charge. You carry a trust dent from the May-2025 billing cutover; an unconsented charge would reopen it. Clear opt-in only. 🟡

Sequencing rule: a starter receives at most one Whisper per stage. 2a and 2b are mutually exclusive (the day-14 signal decides which). A booked membership suppresses every later Whisper. Total touches across the 28 days: two to three, never more — the line is "more human, not less," and over-contact breaks it.

B5The drop-in tripwire — turning your leakiest door into a ladder

A one-time, hidden upgrade offered at the moment a drop-in is bought — $38 becomes a $15 yes for two weeks unlimited, then a staircase to membership.

Drop-in is your widest and leakiest front door: roughly 1,000 new drop-in starters a year (since Jan 2025), most paying once and never returning. The reason the funnel breaks here is the size of the jump — $38 to a $104 intro is a 3× decision at the counter. The tripwire reframes it: at the point of sale, a one-time, hidden offer to roll the $38 into a 2-week unlimited pass for just $15 more. Not a discount — a smaller decision, using the drop-in as the down payment. The take-rate also measures, for the first time, what share of your drop-ins were ever convertible. 🟡

The ladder

Rung 0 · Drop-in — the class they came for
$38
Rung 1 · The $15 tripwire — hidden, one-time, at checkout → 2-week unlimited
+$15
$53
Rung 2 · The fence-sitter extension — +3 weeks, credited to membership if they join
+$90
$143 / 5wk
Rung 3 · Membership — 2× / week or unlimited
$117–120 / fn

Rung 2 is the catch for "I like it but I'm not ready," not a toll everyone pays — the main path runs Rung 1 → the day-10 upgrade Whisper / affinity ask → membership. Crediting the $90 to membership removes the "I paid more than the intro" objection. 🟡

The one trap to engineer around: a 2-week pass that just ends re-creates a cliff, and B3 proved 2 weeks is too short to convert on its own. The fix is the day-10 warming Whisper, repurposed as the ascension trigger — the tripwire and the Whisper playbook are one machine, not two. 🟢

Tripwire take-rate model

Drag the rates. Baseline = today's drop-in starters at one-and-done value. The rates are your assumptions — the model is the arithmetic on them.

1000
25%
25%
Extra realised revenue / year
New members created / year
Drop-ins pulled into the funnel / yr

Values used: one-and-done drop-in $46 (Stage 4 realised), tripwire-but-didn't-convert $53 (the pass), converted member $646/yr (conservative vs the B2 retained-intro average of ~$777). Excludes the $90 rung and any second-year value, so it understates. 🟡

The recommendation: build it — it's the sharpest attack on the drop-in leak you have, and it productises acquisition: measurable take-rates at every rung, an ascension that doesn't depend on a good front-desk day, and clean CAC/LTV metrics that lift enterprise value. Two conditions: keep it hidden and one-time so it never cannibalises the public $104 intro, and fence the $53 unlimited to off-peak so tripwire-goers fill your empty weekday-dawn rooms (Stage 8) instead of the weekend marquees that already turn people away. 🟡

B6Where your warmest members leak to ClassPass

Some starters sample you, then keep attending through the aggregator that pays $0 in your ledger. Here's how many — and the honest worth of chasing them.

The pattern: a member arrives on a drop-in or intro pass, then continues on ClassPass instead of joining you directly. Intro holders defect at nearly double the drop-in rate — they're exactly the warm, convertible audience the aggregator most wants. 🟢

First routeStartersWent to ClassPass after…and never convertedStill active on CP*
Drop-in1,84966438
Intro pass1,16580547
Total3,0141469715

* Attended on ClassPass since 1 Mar 2026 — the reachable "live" list today. The 97 spans the full two-year window; most have already drifted. Source: stage3 entry × attendance plan_name. 🟢

The honest read: stock is small, the prize is prevention. Across two years, 97 warm starters chose ClassPass over membership — but only 15 are still active on it today. Re-engaging the live 15 is a modest, clean win; the real value is stopping the next ~49 a year from defecting at the cliff in the first place — which is the cliff Whispers (B4), not a win-back campaign. 🟡

What it's worth to re-engage

Two numbers. Left: winning back today's live list to membership. Right: the value this group is currently routing to the aggregator. Drag the win-back rate.

35%
Re-engagement worth / year
Members won back
~$5.9k
Gross/yr this group sends to CP*

Worth = live 15 × win-back rate × ~$716 realised first-year membership value (conservative, per B2). *The 97 non-converters logged 719 ClassPass visits over the window ≈ $13.7k of attributed value (~$5.9k/yr) — but that's gross: ClassPass remits some of it off-system, so the net leak is smaller and stays 🔴 Held until the remittance export lands (decision 7). Re-engagement worth is the cleaner number. 🟡

5The ClassPass win-back — bring them home

Trigger attended on ClassPass ≥2× in last 30d AND no active SY membership AND prior drop-in/intro purchase  ·  Window rolling  ·  Sender the teacher they've attended most  ·  Channel in-studio / app

"So glad you keep coming back to class — you're clearly part of the furniture now. When you're ready, joining us directly gets you [priority booking / unlimited / member events] ClassPass can't. Want me to show you the membership that fits how you train?"

Guardrail. Never knock ClassPass or make them feel caught — they chose you, just through another door. Lead with the relationship and the member-only value, not a price war you'll lose. Reachable because they're physically in your room. 🟡 effect unproven — pilot on the live 15.

Workings — how the model is built (so you can pull it apart).
W-B1 · Cohort curves 🟢. Source: stage3_active / stage3_entry, intro bucket, true-new starters, anchor Jan 2025. Survival = members attending in each 30-day period ÷ cohort base. Only fully-observable periods are plotted; censored tails are hidden, not estimated. Pooled = Jan–May 2025 (n=209), the five cohorts with a complete 12-month window.
W-B2 · Value ladder 🟢. The 209 Jan–May 2025 intro starters, each member's actual non-ClassPass sales in their first 365 days, grouped by retention depth: lost at cliff 118 ppl × $110 · held 2–4mo 32 × $260 · held 5–7mo 17 × $613 · locked 8mo+ 42 × $1,236. Total realised $83,633. Locked:cliff ≈ 11×.
W-B3 · The 15% lifts 🟡 assumption, not result. Two independent dials. "Passes sold" scales the intake. "Cliff survival" moves that % of cliff-fallers into the retained tiers in current proportions. Annualised on 428 intro starters/yr (your 2025 intake). The 15% figures are unproven for this studio — the matching Stage 9 model recommends a measured 3-month pilot before scaling.
W-B4 · What could make this wrong 🔴. Realised value is historical; future ladders may differ (Method Lab opened across the road 18 May 2026). A lift in passes sold could draw lower-intent starters who retain worse. Whisper effect sizes are modelled, not yet measured here. Verify with the pilot before banking any annual figure.
W-B7 · ClassPass leak & win-back 🟢 counts / 🔴 dollars. Switchers = true-new drop-in/intro starters whose first ClassPass signed-in visit post-dates their entry (stage3_entry × booking_attendance, plan_name ILIKE classpass). 66 drop-in + 80 intro went to CP after; 97 of those never converted (no active period ≥1); 15 still active on CP since 1 Mar 2026. The 97 logged 719 CP visits ≈ $13.7k attributed value (×$19.05/visit, Stage 8 basis) over the window, ~$5.9k/yr — gross; ClassPass remits part of it off-system (🔴 Held, decision 7), so net leak is smaller. Re-engagement worth = 15 × win-back rate × $716 first-year membership value (conservative). Direction ≠ intent: attending on CP after a paid pass may be substitution or an existing CP user who sampled once.
W-B6 · Drop-in tripwire 🟡 scenario. Population: ~1,000 true-new drop-in starters/yr (stage3_entry, bucket=dropin, since Jan 2025) 🟢. Baseline per starter $46 (Stage 4 drop-in one-and-done realised). Scenario applies operator-set take and convert rates; tripwire-non-convert valued at $53 (pass only), converted member at $646/yr (conservative vs B2 retained-intro ~$777). The $15/$53/$90 prices and the 2×/week $117-fortnight tier are confirmed SY products. Take and convert rates are unproven — pilot before banking. Excludes the $90 rung and second-year value (understates); assumes the 2-week cliff is managed by the day-10 ascension Whisper (if not, convert rate falls). Operator dependency: confirm Mindbody/Wix can present a conditional one-time upsell at point of drop-in sale.
W-B5 · Length-vs-price test 🟢 measured / 🟡 on the counterfactual. Intro true-new starters Jan 2025–Feb 2026 (n=521, conversion observable), signed-in visits bucketed into 7-day windows from pass start; converted = attended in any 30-day period after the first (the same retention proxy as the curve). Measured: converters attend in weeks 3–4 at 79%/75% vs non-converters 56%/40%; 88% of converters still attend weeks 3–4, 78% in week 4; conversion rises 16%→47% with first-fortnight visits; 91% of converters reach 2+ visits by day 14. 🟢 traceable. The 🟡: this is correlational under the current 28-day design — a shorter pass could compress visits earlier via urgency, which no historical cut can measure. Hence the A/B recommendation, not a switch.

B7What you actually have now

Not a report. A capability — and a team you could never have put on payroll.

Everything in this appendix would normally take a CFO, a data team and a strategy firm to produce. A studio your size has none of those, and now doesn't need them. With KulaOS they run in the background, all the time.

The shift is this: every change you make from here is tracked through your own data, automatically, to the cent. Did the new Saturday class actually pay? Is the cheaper intro selling or just discounting? Did the cliff Whisper move conversion, or just feel good? You make the move; the system watches it land and reports back. That's the end of guessing whether it worked. 🟢

A 24/7 CFO
Every dollar reconciled, every change measured against the line it was meant to move
An analyst that never sleeps
Cohorts, retention curves, yield maps, slot economics — on demand, not once a year
A strategy desk
Consultant-grade analysis of every lever, in plain English, costed and ranked
A retention manager
Watching every member's pattern, surfacing the ones about to drift before they do

For a three-to-five person studio, that's four senior hires you'll never make — running quietly, every day, on the data you already own.

But it doesn't make the calls — you do. KulaOS can't tell you whether to protect a teacher whose numbers dipped, keep a community class that loses money, or trust what you feel on the floor over what a curve says. That judgement is yours and your team's, and it's the part that matters most. The system surfaces what's worth considering and proves whether it worked; the people supply the wisdom. The data is the instrument — your team is still the one playing it. Supporting, not replacing. More human, not less.

Phase B is a scenario layer, kept deliberately separate from the Phase A integrity findings above. The cohort data is real and traced to Mindbody; the uplift percentages are operator-set assumptions, not commitments. Nothing here should be banked until a measured pilot confirms the effect sizes for Saltwater Yoga. More human, not less.