01 | Introduction

ONE8T Wellness | Growth System Proposal

Prepared for Skye Kaiss by Mario Paguio, Kaizen Collective. 14 July 2026.

Six months after opening, Marda Loop carries 690 active paying clients and $396,624 of net revenue, and its fastest daily pace is happening right now, this month. West Springs is in presale behind it. The plan is four Calgary locations, then franchises across Canada.

This document is the operating system for that plan. The problems you raised on our call are real, and every one is answered in here, including the one you tested me with. But a business built to franchise does not need four repairs. It needs every stage of the member journey, first ad to year two, running on machinery a new location inherits on day one. That is what we operate across STRONG Pilates' 150 studios, and this document fits it to ONE8T.

Before writing it, I read your systems with the access you granted: both Grow accounts (46 automations, 69 triggers, 46 email templates), both Meta ad accounts (337 ads), Marda Loop's complete member and payment history from Core, and the full training log of your current AI (85 corrections since January). Every number in here comes from those reads, taken 13 and 14 July 2026. Where a number needs a definition to be trusted, the definition is printed beside it.

How to read it: section 02 is what the reads found, including the conversion number Hapana overstates by three times. Section 03 is the system and the timeline. Section 04 is the build, down to the field names. Section 05 is our Canada portfolio, counting definitions attached. Section 06 is the price, written so it can be handed to a franchisee. Section 07 is what you own.

The plan at a glance

PhaseWhat it deliversLive
1. The Core to Grow bridgeSession times flow into Grow. Timed confirmations and reminders. The book and pay flow without an account wall.Bridge week 2. Booking flow week 3 to 4.
2. The AI hostOne agent for the brand, connected to your real schedule and inbox, corrected by your team same day.Side by side week 3. Cutover on your word.
3. The lifecycle libraryThe journeys between first visit and year two: the intro decision arc, member onboarding, milestones, absence recovery, pack expiry, payment recovery, saves and winback, campaigns.First tranche week 2. Full library by week 6.
4. The growth engineAds, creatives and landing pages run as one journey with true exclusions, reported on one dashboard.Restructure from week 1.
5. The templateEverything above packaged so location 3 to 100 switches it on instead of rebuilding it.Hardens as 1 to 4 ship.

Investment: $5,000 AUD once for the template build, then per studio either $400 AUD a week (full service) or $200 AUD a month (Grow and AI only). CAD context and full scope in section 06.

02 | The Problem

What you want, and why the current stack can't deliver it.

A. The asks the platform can't meet

You wantThe stack todayConsequence
Reminders that say the session timeCore never pushes booking time to Grow. 0 of your 46 automations can reference it. Your own confirmation SMS reads "arrive 30 mins before your scheduled session time" with no time in it.Every message is generic. The 98% show rate depends on people checking the app.
Book and pay in one pass, no accountCore's simple signup covers purchases only. Booking still requires an account and password.The midnight lead hits a wall at the exact moment they were ready.
Correct your AI the same dayAll 85 corrections filed since January were applied manually by the vendor's team. Their interface says so on every ticket.Errors live for days. One logged incident: "Agent fabricated membership package details."
One brain for the brandTwo agents trained separately. One knowledge base is empty; everything it knows arrived through correction tickets.You teach everything twice. Location 3 would mean a third student.
AI that books off the real scheduleThe AI has no connection to Core or Grow. Its integrations page shows every CRM as NOT CONNECTED.It can only send links. 5.3% of its conversations end in a tracked goal.
One accountable partnerThree vendors. None owns the journey end to end.The problems live in the gaps between them, and the gaps belong to nobody.
The system today
Four tools, three vendors. The two links that matter most do not exist.
Meta Ads agency managed, 337 ads Landing pages book = create an account Hapana Core bookings, members, payments Grow CRM 46 automations, blind to time HireBob AI no CRM connection sync drops the session time summaries only, one way no schedule access Dashed = disconnected or one way. The AI cannot see a booking. Grow cannot see a time.

B. What the accounts show

Six findings from the read, in order of what they cost you.

Finding 1 | Conversion

True intro to membership conversion is roughly 11%, not the 33.5% your reporting implies.

1,357 people have bought an intro. Hapana's built in conversion field marks 33.5% converted, but it counts any later purchase (session packs, PLUS ONE add ons, even comps) as a conversion. Rebuilt from actual product names, two independent methods land at 10.9% and 11.7%. The matured cohort (intros older than 90 days) runs 12 to 14%.

Addressed by Phases 1 to 4 together: this number is the whole lifecycle, not one fix. Benchmarks in section 05 run 31.8% to 41.7%.

Finding 2 | Cost per lead

Your cost per lead has doubled against the account's own baseline.

Lifetime: $15.08 across 2,429 leads. Last 30 days: $32.77 (84 leads on $2,753). Structural causes are visible in the account: 183 ads spread across 35 ad sets, audience frequency at 2.15, and a lead exclusion list holding roughly 20 people, so paid impressions keep reaching people who already converted.

Addressed by Phase 4. Same window benchmarks in section 05.

Finding 3 | The AI

The AI is a sealed unit you rent, not a system you run.

85 corrections since January, every one applied manually by the vendor. No connection to your CRM or booking data. In its own insights, 5.3% of conversations end in a tracked goal. In the sampled threads, three separate leads told it to stop messaging them, a side effect of its email plus SMS double touch default.

Addressed by Phase 2. Its genuine strengths (3.4 second responses, polite decline handling) carry over; its accumulated rules import as the day one rulebook.

Finding 4 | Post purchase

There is no engine after the sale.

No closeout automation (you send mass texts by hand), no structured retention arc after a purchase, no winback. 195 genuine cancellations are logged; a reason was captured for 82% of them, which is a usable base nobody is using.

Addressed by Phases 1 + 3: the lifecycle library exists precisely for this, and it becomes possible once the bridge feeds Grow attendance and time.

Finding 5 | Revenue

The underlying business is compounding. The layers above it leak.

$396,624 net revenue since opening. May was the biggest full month at $74,515. July is running $2,603 a day over its first 13 days, the fastest pace yet. Demand is not the problem; capture is.

Finding 6 | Repeatability

Nothing is templated for location 3.

Every system above was hand built for one location at a time. On the current stack, your third opening starts from zero, with the same three vendors and the same gaps.

Addressed by Phase 5.

Marda Loop net revenue by month, CAD
From your Core payment records, net of refunds. July shows its first 13 days; its daily pace ($2,603/day) is the best of any month so far.
$44.9k $52.9k $60.9k $64.7k $74.5k $63.2k $33.8k Jan Feb Mar Apr May Jun Jul* *13 days

C. The whole map

The asks from our call live in the first half of the member lifecycle. The second half is where compounding revenue lives, and today it is empty. This is the difference between fixing complaints and installing the operating system.

The member lifecycle, mapped against your accounts
Top: the stage. Middle: what runs at Marda Loop today. Bottom: what the system runs at every location.
Attract Book Show up Decide Onboard Attend Renew Recover 183 ads $32.77 a lead booking exists behind a wall reminders exist no times 11% convert no day staging nothing runs after joining no milestones no visit data 1,134 packs no expiry journey 195 churned no winback focused ads pay first flow timed comms day by day arc 60 day journey milestones expiry journeys saves + winback Measured end to end: one weekly dashboard reads every stage Solid border = partly covered today, where your asks live. Dashed = does not exist today. Sage = run by the system, at every location, from the same template.
03 | The Process

Five phases. One operating system.

The same architecture we run across the STRONG Pilates network: infrastructure first, then the AI, then the journey library that works every stage of the lifecycle map above.

The system once the build lands
One bridge, everything connected. The AI, the journeys and the ads finally see the same truth Core sees.
Meta Ads Kaizen managed Booking flow pay first, no account wall Hapana Core source of truth The Bridge Public API, sync every 15 min Grow CRM fields, journeys, one inbox ONE8T AI host one brain, all locations Weekly dashboard members, revenue, CPL bookings + times + packages 4 live fields per contact live schedule writes to the same inbox exclusions, nightly Solid sage = the connections this plan builds. Everything lands in accounts ONE8T owns.
PhaseOverviewReplaces
1. FoundationThe bridge reads Core through the Public API and gives Grow the session data it has never had. Timed communications go live. The book and pay flow removes the account wall.The wait for Hapana's V2
2. AI hostOne agent for the brand, running inside your inbox, booking off the live schedule, corrected by your team in plain English the same day.HireBob / SalesApe
3. Lifecycle libraryNine journeys covering every stage from intro decision to winback, cloned from patterns running at fleet scale and fitted to ONE8T's products. The full breakdown is in section 04.The gaps nobody owns today
4. Growth engineAd account rebuilt to a focused structure with true exclusions, creatives produced from your raw footage, landing pages aligned to the booking flow, one weekly dashboard.The marketing agency
5. TemplateEverything above captured as the ONE8T template with central controls, so each new location inherits the full system on day one.Starting from zero at location 3
Timeline and your critical path
What we deliver each week, and the only things we need from your side.
Week 1 Week 2 Weeks 3 to 4 Ongoing Bridge build, then live at Marda Loop Booking flow mockup build and ship on your domain AI host brain built side by side, cutover on your word Journeys intro arc and onboarding first, full library by week 6 Ads restructure, exclusions, creatives, then always on From you approve reminder copy review journeys + AI go / no go on cutover 15 min weekly

Access is already largely in place from the audit: Grow, both ad accounts and Core were connected on 13 to 14 July. Nothing on your critical path takes more than 30 minutes in any week.

04 | The Breakdown

How each phase is actually built.

Phase 1 | The Core to Grow bridge

Give Grow the four fields Core never sends.

"When somebody books an intro, say tomorrow at 2pm, Core does not push what time that is. So I can't send out a text reminder of what time the freaking session is. Did you fix that?"

Skye, on our call, 6 July
  1. Read. The Hapana Public API (the same interface Core's own apps use) exposes each client's future bookings. The bridge polls it every 15 minutes during studio hours, with a full reconciliation sync nightly. No dependence on Hapana's roadmap, no V2 wait.
  2. Write. Four plain custom fields appear on every Grow contact: Next Session Date, Next Session Time, Last Visit, Active Package. Your team sees them on the contact card like any other field.
  3. Trigger. Journeys key off those fields: confirmation with the real time, a reminder the evening before, a nudge the morning of the session, and a follow up if they miss. If someone reschedules, the next poll updates the fields and every pending message moves with them automatically. You approve every message before anything sends.
Phase 1 data flow
Core bookings + times Bridge every 15 min Grow contact fields Next Session Date / Time Last Visit / Active Package SMS, tomorrow 6pm: "Your ONE8T suite is booked for 2:00pm tomorrow. Arrive 1:30." The same fields feed the AI (Phase 2), the closeout journeys (Phase 3), and the ads exclusion sync. Today's version of that message: "Please arrive 30 mins before your scheduled session time." Your live confirmation SMS. It cannot name the time, because nothing in Grow knows it.

The book and pay flow

1. Pick a session live availability, via the API 2. Details + card name, email, phone, pay 3. API does the rest account + booking created 4. Confirmed with the real time No account wall. The password step disappears from the lead's path; the account still exists in Core, created by the API.

This is the one custom build in the plan. Design mockup in week 1, live in weeks 3 to 4, on your domain. The same pattern runs on the STRONG Pilates network, which is where the bridge design comes from.

Phase 2 | The AI host

One brain, connected, corrected by you.

The agent's structure
ONE8T brand brain framework, offers, policies, voice Marda Loop folder hours, open studio flows West Springs folder presale stage, founder offers Location 3 folder inherits everything, day one Live schedule + inbox books real sessions in Grow Same day corrections plain English, logged in the repo 85 imported rules six months of training kept
  • Connected, not sealed. It reads live availability through the bridge, writes every conversation into your Grow inbox, and sees purchase state, so it stops selling the moment someone buys.
  • Correction mechanic, with a clock. It says something wrong at 9pm; your team messages the correction in plain English; it is applied the same day and logged, timestamped, in a repository you can read. No ticket queue. Compare: 85 corrections to date, each one waiting on the vendor's staff.
  • Your knowledge survives the switch. The Nervous System Training document and every rule those 85 tickets taught (pricing, PLUS ONE handling, promo eligibility, escalation triggers) import as the day one rulebook.
GuardrailRule
ChannelsOne channel per touch. No email plus SMS double sends.
Hard stopsStops on purchase. Stops on a clear no. Both are system rules, not habits.
Account dataNever guesses balances or credits; directs to the app or the studio.
Pre openingNever confirms sessions at an unopened location; offers the reservation path or Marda Loop.
EscalationCall requests, repeated booking failures and complaints notify a human immediately.

Target behaviour, one exchange:

Lead: "Can I move my Thursday session?"
Host: "You're booked Thursday 2:00pm at Marda Loop. Nearest open suites are Thursday 4:30pm or Friday 9:00am. Want me to move you to either?"

Grounded in live schedule data. Today's system cannot see any of those times.

Cutover: the new host runs side by side in week 3 while the current system keeps working. It takes over when you say so, and cancelling the old vendor is your lever to pull, on your timing.

Phase 3 | The lifecycle library

The journeys between first visit and year two.

This is the architecture we operate across the STRONG Pilates network: an intro lifecycle staged by day, a 60 day member journey, attendance milestones, absence recovery, birthday touches, monthly campaign pushes, presale machinery. Below is that same architecture fitted to ONE8T's product ladder: the $40 and $60 intros, 3 for $99, session packs, Base, Starter, Elevate and Peak memberships, PLUS ONE guests.

JourneyFires whenWhat it doesAt Marda Loop today
Intro decision arcAn intro is bookedTime aware touches from booking to decision day: real time confirmations, day counters, the right nudge on the right day, a human handoff if it stalls.Partial. Journeys exist but are blind to time and day.
Member onboardingFirst membership paymentThe first 60 days: welcome, habit building cadence, check ins, an early win moment.Missing.
Attendance milestones10th, 25th, 50th, 100th sessionCelebrate the milestone, ask for the review at the high point, nudge the upgrade at the right moment.Missing. A review ask exists but is not tied to visits.
Absence recovery7, 14, 21 days without a visitEscalating win back, with a human task at day 21.Partial. Tag based, with no visit data behind it.
Pack expiry and repurchaseLast session used, or pack expiresA repurchase journey before the habit breaks.Missing. 1,134 packs have already run out.
Payment recoveryA payment failsRetry, card update link, a human task if it persists.Missing. 14 open failures right now.
Save and winbackCancellation intent, then 30, 60, 90 days after churnA save sequence first, then structured winback offers.Missing. 195 churned, with reasons on file for 82% of them.
Birthday and momentsThe dateA personal touch from the studio, not a coupon blast.Missing.
Campaign cadenceMonthly calendarThe mass texts you send by hand become planned pushes: offer, audience, exclusions, measurement.Manual today.

Status comes from the read of your 46 automations. "Partial" means a version exists but cannot see session times, visits or day counts until the bridge (Phase 1) feeds it. Every journey ships with copy for your approval, and every journey is part of the template that location 3 inherits.

Phase 4 | The growth engine

Fewer ads, true exclusions, the whole journey measured.

Ad account restructure
Today 183 ads, 35 ad sets, budget spread too thin to exit learning Exclusion list: about 20 people Under this plan A handful of focused ads per objective, each with enough budget to exit learning Exclusions synced nightly from Core's real list Creative production included: you supply raw photos and video, we produce statics, carousels and cutdowns.
  • Landing pages align to the booking flow, so the ad's promise and the checkout are one motion instead of a handoff between vendors.
  • West Springs presale runs as a campaign: founder tiers, waitlist nurture, opening countdown, the machinery we run for new studio openings through the year.
  • One weekly dashboard covers members, net movement, intros, revenue and cost per lead. Section 05 shows a live example.
Phase 5 | The template

Build once. Switch on per location.

The ONE8T Template bridge, fields, journeys, emails, AI brain, booking flow, dashboard Marda Loop live now, first to inherit West Springs presale playbook included Locations 3 and 4 Franchisees "here's the vendor, here's the cost"
  • Instantiating a location takes days, not months: journeys, emails, fields and the AI's location folder are cloned from the template; the dashboard gets a new row.
  • HQ keeps control structurally. Offers, journeys and pricing live in the central template. Locations get local hours and tone, not their own rules.
  • The franchisee pitch becomes one line: the investment table in section 06 is written so you can hand it over as is.
05 | Benchmarks

Our Canada portfolio, same window, same definitions.

Five STRONG Pilates studios we manage in Canada. Names hidden because these are clients' books; every figure pulled live on 13 and 14 July 2026.

Definitions, so the comparison is apples to apples: a lead is Meta's lead event (form or pixel), counted identically in your column and ours, window 13 June to 12 July 2026. Note your internal lead counts can read higher than Meta's because a two person booking is two people to you and one lead event to Meta. Intro to member = intro purchasers who later hold a recurring membership, from each studio's member system.
StudioActive membersCPL, 30dRevenue, 30dIntro to memberNote
Canada Studio A260$6.11$83,98331.8%Six straight weeks of net member growth, churn 4.2%
Canada Studio B150$5.52$38,47341.7%Newest of the five, open about 4 months
Canada Studio C243$9.07$99,52035.4%Biggest revenue month in the portfolio
Portfolio average CPL across all five: $10.23 ($7,875 spend, 770 leads). The average is printed, not just the best studios.
Cost per lead, last 30 days, CAD
ONE8T Marda Loop today $32.77 Canada portfolio average, all five studios $10.23 Most efficient studio $5.52
Intro purchasers who become members
Your true rate (Finding 1) against two of the portfolio's.
ONE8T Marda Loop, all time ~11% Canada Studio A 31.8% Canada Studio B 41.7%

The weekly view, from a live client dashboard

Canada Studio A, last 30 days, name hidden. This is the report that replaces Dash View and the Excel exports.

260
Active members
+10
Net growth, 30d
4.2%
Churn, 30d
$83,983
Revenue, 30d
$6.11
CPL, 30d

Weekly net member movement, six positive weeks in a row.

06 | Investment

Priced to be handed to a franchisee.

ItemAUDApprox CADCovers
The ONE8T Template
one time, for the business
$5,000about $4,900Core to Grow bridge, timed communications, book and pay flow, the full nine journey lifecycle library with its email and SMS templates, the AI brain, the dashboard. Built once, owned by the business, inherited by every location.
Full Service
per studio, per week
$400about $390
roughly $1,700/month
Meta ads management with creative production included, Grow and the lifecycle journeys, the AI host, monthly campaign cadence and presale execution, the weekly dashboard.
Grow + AI
per studio, per month
$200about $195Automations, the AI host, dashboard and support, for any location where ads are handled elsewhere.

CAD figures converted at 0.98 (ECB reference rate, 13 July 2026), indicative only; billing is in AUD. Ad spend is separate, stays in your Meta account, and is paid to Meta directly.

Against today's stack

LineTodayUnder this plan
AIAbout $1,600 CAD a month for two locations, sealed, vendor correctedIncluded, connected, corrected by your team
Automations and CRM workAd hoc, quoted per changeIncluded
Creative productionSeparate ($5,000 in photoshoots so far, edits extra)Included; you supply raw footage only
ReportingDash View plus manual ExcelWeekly dashboard included
Agency management feeYour current fee, on top of ad spendReplaced by the $400/week, all of the above inside it

No charge until the template work begins.

07 | Ownership and Continuity

Built in your accounts. Documented as it's built.

  • What is yours outright. The fields, journeys, emails, funnels and pipelines are built inside ONE8T's own Grow accounts. If we ever part ways, they stay, and they keep working.
  • What is a service. The bridge and the AI run on our infrastructure, and their full configuration, instructions and change history live in a shared repository your team can read at any time. Nothing about your own system is hidden from you. Today, the equivalent knowledge sits inside your vendor's back office.
  • How we work. You deal with me directly; my team executes; I review before anything reaches you or a lead. This is the same operating model I run across the STRONG Pilates network of 150 studios.
  • Continuity. A business partner, a trained team, and documentation deep enough that the work survives any one person. This document was produced from a one day, fully documented read of your estate; every build gets the same treatment.
08 | Next Step

One decision starts it.

Reply with the tier for each location (full service, or Grow and AI only). Access is already in place from the audit, so week one starts the same day: the ad account restructure begins, and the booking flow mockup and reminder copy land with you for approval.

Mario Paguio
Kaizen Collective | mario@kaizencollective.com.au