Regulars — Repeat-Purchase Agent on SOP + Brain

Picks out repeat purchases and special offers that matter to the shopper, based on order history, product logic, and customer behavior. Builds and sends its own campaigns, and measures its own result in dollars.

Section 1 · Idea Selection

1.1 The Bet, in One Sentence

We're building an autonomous repeat-purchase and special-offer agent for DTC brands on Shopify and Klaviyo: it replaces the retention agency. Offers are built from three sources:

• an auto-generated repeat-purchase SOP;
• product descriptions and product logic;
• shopper behavior.

Together this is a knowledge layer unique to each brand, and it self-improves: every offer is tested against a control group, the result feeds back into the model, and experience carries over between stores — so performance goes up over time instead of decaying.

Core ICP: brands at $10-30M GMV with a repeat-revenue share above 35%. The $3-10M tail is fixed-fee only, no percentage — their lift can't be measured with statistical confidence. Pricing: a fixed-fee ladder at $3.5-7K/month by brand size, plus 10-12% of measured lift.

Target: $100K MRR is honestly reachable by month 5-6; month 4 only under a best-case run (details under "Path to $100K").

The Engine

1. Auto-SOP. Scenarios aren't hand-written: order history reveals what the best-outcome customers did (LTV, full basket), in what sequence, at what intervals.
2. Anchored to product logic, not basket statistics. The offer rests on the derived path and on product attributes (ingredients, usage protocol, what complements what) — not on bare "customers who bought this also bought." That's what separates us from Rebuy, Nosto, and Klaviyo AI: they recommend by basket co-occurrence, with no understanding of the product.
3. A measurement layer. Every offer is measured against a control. Pricing rests on this (percentage only on proven lift), so does protection against discount cannibalization, and so does filtering false patterns out of the auto-SOP. It's also the self-learning mechanism: the system sees what worked and reweights offers accordingly.
4. Cross-store experience transfer. Purchase cycles and product bundles look similar within a niche, so a new brand doesn't start from zero — it starts from patterns accumulated on other stores, and gets value from day one while its own data builds up. The edge compounds with every client; Klaviyo can't do this, its models are locked inside a single account.

How It Works

1. Connect the merchant's data in read-only mode: customer base, order history, catalog, existing repeat-purchase and bundling policies.
2. Connect the email channel (Klaviyo).
3. The system finds patterns in the history: individual purchase cycles, product bundles, best-customer paths. From these it auto-builds a repeat-purchase SOP: which triggers, for whom, when. The brand's existing policies aren't discarded — they're extended.
4. It scores the upsell opportunity: how much money is being left on the table, and where. This same estimate doubles as the sales audit.
5. The agent builds and sends campaigns by time and by event. Each one is measured against a control group, and the result feeds back into the model.

What We Sell

Land, a narrow entry point: a timely repeat purchase — the same product, on the date it runs out for that specific customer. Expand within the account: filling gaps in the bundle by product logic; special offers only where they genuinely add a sale; light post-purchase touches (review requests, "how did you like it"). We don't take on complex service workflows.

1.2 Why This Bet, Not Another

1. The only idea on the list (see 1.5) that directly drives sales: the result is measured in the client's revenue, not in reports or advice.
2. Easy to scale: software with no manual work per client, one integration stack, auto-configuration by default.
3. Minimal client involvement: read-only connection, no need to change the brand's processes, the agent runs on top of the existing stack.

1.3 Why You Specifically Should Own It

I've done by hand what this agent automates. I built bundles, special offers, and time-based upsells in e-commerce, so I know what a well-made offer looks like. I wrote SOPs for a beauty salon and know how to turn a fuzzy "what comes after what" into explicit rules — exactly the layer that feeds the auto-SOP, and beauty happens to be the top-scoring niche in our filter. I've built customer journeys and understand where in that journey which email belongs.

Weak spots, stated plainly: the ML and measurement statistics fall to a technical co-founder; scaling sales is my growth area.

1.4 Why This Becomes Profitable Fast in This Market

1. We measure the client's pain in dollars and lift LTV: the conversation is about revenue from the start, not features.
2. The client pays for the result: percentage only on measured lift, and the fixed fee is lower than what they already pay an agency. Easy to say yes.
3. The first pilots run on the holding's own business units: acquisition is close to free, no cold sales cycle to start.
4. We offer a narrow, legible service with per-client tailoring: one job, done end to end, on that brand's own data — not a platform with a hundred settings.

1.5 Competing Bets Considered and Rejected

Competing ideas from my own list, each with the reason for rejection:

1. Investment-attractiveness scoring agent: researches the market and trends, produces a report, flags changes. Rejected: sells a report, not a result; a one-off purchase, value can't be tied to the client's money, long trust cycle.
2. Brand persona agent: defines brand value and points of differentiation against product, competitors, and audiences. Rejected: the output is subjective, can't be paid on results; essentially a one-off agency project with no recurring revenue.
3. Business strategy and OKR-setting agent. Rejected: impact on revenue is indirect and lags by quarters, the buyer champion is fuzzy, competes with free templates and consultants.
4. Mystery-shopper agent: calls or messages in, checks the service script is followed. Rejected: measurable, but the deal size and market are small; it finds problems rather than generating revenue; scale is capped by the number of checks run.
5. Workflow and user-journey diagnostic agent: maps the journey cross-platform, flags where to improve touchpoints. Rejected: heavy integration across many systems, months to first value; it hands out advice instead of taking action, and advice doesn't sell on a percentage.

What all five share: they sell knowledge or a signal, and someone else's hands turn that into money. The chosen agent is the only one that takes the action itself and measures its own result in revenue. That's where point 1 in 1.2 comes from.

Adjacent versions of the idea itself were also rejected: salons and services as the first vertical (small deal size, complex procedures, a v2 candidate); "just a smarter Klaviyo" (a feature of the channel owner, gets absorbed); a company-memory platform (sells knowledge, not money); competitor price monitoring (a different product, a different buyer).

ICP: Data-Driven Filter

Everything reads off Shopify, Klaviyo, and CRM at the audit stage, before any promises.

Notes. A repeat share above 70% is also a poor fit: everything's already been squeezed out, nothing left to add. A catalog under 15 products leaves only repeat orders with nothing to bundle; above 500-1,000 it's already a browse-driven marketplace where our approach falls apart. Good-fit niches: beauty, supplements, pet products, coffee and food. Not a fit: furniture, electronics, jewelry.

How the Market Solves This Today, and the Risks

The capabilities largely already exist:

• Klaviyo predicts next-order date and churn risk
• Rebuy and Nosto upsell
• Recharge runs subscriptions
• Shopify Sidekick generates campaigns

The gap isn't in the capabilities, it's in using them: turning this into money needs a dedicated person, and mid-market doesn't have one. The result: either static flows, or an agency doing it by hand for $3-10K/month.

The evidence is indirect (we test the assumption in the first audits — read-only access shows what's turned on and how it's configured): triggered emails generate 41% of email revenue from 5% of sends, and the market for agency retainers exists precisely because brands can't pull this off themselves.

The real competitor, then, isn't Klaviyo (we sit on top of it) — it's the agency. Against an agency we're cheaper, more consistent, measurement is built in, and we learn across many stores at once.

Risks, in descending order:

1. Absorption: the platforms automate fast, and "we do it for you" isn't a moat by itself. Differentiation rests on three things only: honest measurement, cross-store experience transfer, pay-for-result.
2. Sliding into an agency: closing the gap by hand turns this into a service, not a product. Auto-first only.
3. Honest measurement shows a smaller number than inflated attribution. The truth is a harder sell.
4. No market precedent for a "percentage of measured lift" contract. Can only be validated by actually selling it (hypothesis H3b).
5. Dependency on the Shopify and Klaviyo APIs.
6. Small brands don't have enough data, hence the bottom of the segment is fixed-fee only.

Measurement

The foundation for pricing and for H4. Principle: we don't compare periods, we compare parallel groups. The control group in the same period is the "baseline" — seasonality, promotions, and the brand's own campaigns hit both groups equally and cancel out.

1. Sticky holdout, 5-10% of the base: a customer sits in control for the whole quarter, which captures compounding effects. Rotated once a quarter.
2. For each flow, the contract specifies whether it's an add-on (test = the brand's program plus us, control = the brand's program alone) or a replacement (our flow vs. their old one). The brand's own program runs uniformly across both groups.
3. Ghost sends: we measure only among those who were actually triggered; the control logs the moment the email would have gone out. Baseline conversion rises from 0.31% to 5-15%, and the required sample size drops by an order of magnitude.
4. Cadence: a monthly cut is directional (for internal steering and the client report); the quarterly reconciliation is the one that counts (the final percentage is billed off it). On large brands, with a big enough lift, the monthly cut converges on its own.
5. If the brand had no campaigns before: control equals silence, the measurement is cleaner, the lift is bigger. Triggers are computed off purchases, not email history, so the engine has something to learn from regardless.
6. Measurability threshold: roughly 250K triggered profiles per quarter. Below that, the percentage model isn't offered.

Section 2 · Core Hypotheses

Each hypothesis: a statement of belief, the test we'd run, the pass signal, the fail signal.

H0 · Measurability

H0: A sufficient share of ICP brands is physically measurable within a quarter. Test we'd run: on data from 5-10 brands from the audits, measure base size and the actual monthly share of triggered profiles (the single biggest unknown on this whole map). Pass signal: 30%+ of brands have 250K+ triggered profiles per quarter. Fail signal: below 15% — the percentage model doesn't apply to the tier; pivot to fixed-fee only, or move the ICP upmarket.

H1a · Pain, Missed Repeats

H1a: Through the window of overdue or missed repeat purchases, the median brand is leaking 6%+ of annual revenue, of which 1.5%+ is recoverable. Test we'd run: audit 8-10 brands; the cycle-length rule is locked in before we look at the data. Pass signal: median leak 6%+ and recoverable 1.5%+. Gray zone (pivot): 4-6% — decided by H1b. Fail signal: under 4%, or recoverable under 0.8%.

H1b · Pain, Bundle Gaps

H1b: A missed bundle top-up leaves 1%+ of revenue recoverable. Test we'd run: the same audit, a separate metric. Pass signal: 1%+. Fail signal: under 0.5%.

H2a · The List Exists

H2a: Brands passing a dual filter (size and base volume) form a nameable, countable list. Test we'd run: Store Leads by revenue tier, plus base-size measurement in the audits. Pass signal: 100+ brands pass both conditions. Fail signal: under 30.

H2b · Channel

H2b: At least one channel converts to a booked audit above the floor, with the denominator being everyone we reached. Test we'd run: cold outreach, 300-500 emails, plus warm intros. Pass signal: warm 10%+, or cold 1%+. Fail signal: warm under 5% and cold under 0.3%.

H3a · They Pay the Fixed Fee

H3a: After a paid audit, the brand signs an LOI or a pilot with the fixed fee within 30 days (the threshold sits at the upper bound of the 14-30 day market sales cycle for this deal size). Pass signal: 3+ out of 10. Gray zone (pivot): exactly 2 of 10 — expand the sample. Fail signal: 0-1 out of 10.

H3b · They Take the Percentage

H3b: A majority of those who sign choose the percentage option over a flat fixed fee. Why this is risky: pay-for-result is thriving in the market (Chargeflow: 25% of recovered chargebacks; Intercom Fin: $0.99 per resolved ticket; Rebuy: money-back on attributed revenue) — but always against a directly observable event. We looked twice and couldn't find a single contract paying on lift measured against a control group: it needs a measurement both sides trust. We lock in the choice — percentage or fixed fee — explicitly in the LOI. Pass signal: at least 2 of 3 signees take the percentage. Fail signal: everyone insists on flat fee only — pricing pivots to a larger fixed fee.

H4 · Effectiveness (Core)

H4: The agent drives lift on the affected slice against a fixed holdout, and at sufficient volume this is visible within the window. Test we'd run: a pilot on 2-3 brands with 250K+ triggered profiles per quarter, fixed 5-10% holdout, ghost sends, CUPED, add-on-or-replacement locked into the contract. Pass signal: directionally within the window, a point estimate of 12%+ with an 80% interval entirely above zero, and the partner commits to riding out the quarter; final pass is 15%+ at the quarterly reconciliation (around week 21, past the window). Gray zone (pivot): estimate 5-12%, or the interval touches zero at sufficient volume — ride out the quarter, revisit the percentage at reconciliation. Fail signal: estimate 5% or below, or the interval crosses zero at sufficient volume. A volume shortfall is not a failed test — it's the absence of a test — go get more base.

H5a · Channel Throughput

H5a: The agency channel, separate from the holding's pilots, delivers 8+ qualified intros a month, 3+ demos, 50%+ base-fit rate. Fail signal: under 4 intros a month, or fit under 30%, or intro-to-demo under 10%.

H5b · CAC

H5b: Fully loaded channel CAC (founder time, tools, partner share) per signed client. Pass signal: under $8K, paid back by the fixed fee in 2.5-4 months. Gray zone (pivot): $8-15K — not scalable, fix the funnel. Fail signal: $15K+, which is market-normal for a $40-80K deal, meaning there's no warm-channel edge; pivot to white-label or rebuild the channel. Holding pilots aren't counted in CAC.

H5c · White-Label

H5c: A deal structure that preserves the agency's own margin actually works. Pass signal: 2+ agencies accept the terms within 8 weeks. Fail signal: zero — the agency track is shut down, we go direct only.

Testable only in Stage 2; early proxies within the window:

H6 · Retention (Stage 2)

H6: Annual NRR of 120%+ can't be measured in 4 months. Proxy within the window: among customers with a repeat purchase, an add-on offer produces a 90-day lift of 10%+ vs. holdout, unsubscribes rise by under 0.5pp, margin stays at or above control.

H7 · Cross-Store Experience Transfer (Stage 2)

H7: Experience from other stores benefits a new client from day one. Full validation needs 10 live brands; within the window we have 2-3, so we rehearse it on historical audit data instead. Test we'd run: hide one brand's data, train the model on the rest, replay that brand's real history, and compare the "borrowed-experience" model against a model trained on its own data. Repeat for each brand in turn. Pass signal: borrowed experience hits 60%+ of the fully-warmed model's accuracy in month one. Gray zone (pivot): 30-60%. Fail signal: under 30% — transfer doesn't work, no compounding edge.

Pricing

Logic: the fixed fee covers costs and doesn't scare anyone off (it's lower than what the brand already pays an agency for this work); the earnings sit in the percentage of measured lift.

Pilot quarter for everyone: fixed fee $2,500-3,500. After the first confirmed reconciliation, the fee steps up the ladder. Cheap to start, price rises after proof, entry conversion doesn't break.

Why we don't charge on sends, catalog size, or order count: those are cost/activity metrics (and Klaviyo's own axes), not result metrics, and they create a perverse incentive to send less. Cost to serve a client is roughly $1-1.5K/month, so the fixed fee is profitable on its own — but we justify it to the client on the budget it replaces, not on a promised ROI: ROI before the pilot is a model, the promise only lives in the percentage.

Billing rhythm: a directional monthly report, money on the quarterly reconciliation. The first 4-8 weeks bill on interim operating metrics, flagged as such in the contract.

Configuration: automatic by default; larger clients get human confirmation; manual configuration is the exception only. A manual-everywhere mode would turn this right back into an agency.

Who pays: the brand's retention lead or founder, out of an existing budget (agency plus apps). If it doesn't work: the pilot is risk-reversed — no lift, no fixed fee charged; if there's still no result, we part ways.

GTM: Two Non-Overlapping Channels

1. Direct, primary. Entry via the audit: brands with no agency, or unhappy with theirs. Full-scale marketing starts month 5, once we have pilot case studies.
2. Agencies as a white-label distributor, not a source of leads to replace. The offer: the agency keeps its retainer, we're the engine sold wholesale, they serve more clients with the same team. ARPU is lower than direct, CAC near zero.

Rule: direct marketing never targets a partner agency's own clients. A "friends first, competitors later" play is off the table — the market's too small, we'd burn both channels. Gate: H5c; if zero agencies agree, the track shuts down and we go direct only.

Section 3 · Validation Sequence

Logic: first, cheaply validate pain and measurability on other people's data (E1-E3); then the channel and paid intent (E4-E6); only then the expensive pilot (E7). Each next step is unlocked by the previous result, not by the calendar.

Hypothesis map: H0 measurability, H1a/H1b pain, H2a list, H2b channel, H3a fixed fee, H3b percentage, H4 effectiveness, H5a throughput, H5b CAC, H5c white-label, H6 retention (in-window proxy), H7 experience transfer (in-window proxy).

3.1 Experiment Timeline

Sequencing: E0 in week one (nothing to lead with otherwise); E1-E3 run in parallel in the early weeks; E4-E6 overlap with them; E7 only after the E3 measurability map. E7's final quarterly reconciliation lands around week 21, outside the window.

Product Form & Technical Validation

The product doesn't need to exist whole from day one — two milestones, with a mandatory checkpoint between them:

1. Audit tool (read-only plus lost-repeat-purchase math). Needed by week 1, for E1, E3, E6. An analytics script, no sending, no ML, cheap to build. It's also what we sell with.
2. Offline engine validation (E-eng, weeks 3-8). Before selling pilots on a promised result, we build the SOP on historical audit data, replay it on held-out data, and check whether the engine would have produced measurable lift — and whether it can actually be built in a reasonable time. Gate before E7: no offline result, or it doesn't come together, no live pilots launch.
3. Full engine (builds its own SOP, sends through Klaviyo, measures via ghost sends, holdout, CUPED). Needed by week 9, for the E7 pilot.

3.2 Decision Tree

Enabling steps (E0 landing page, E2 connector) aren't gates in this tree: missing them delays the start, it doesn't kill the bet. Each row is one decision point.

A volume shortfall in E7 is not a failed test — it's the absence of one — we go get more base. Holding pilots aren't counted in CAC (H5b), they're the floor of the funnel.

3.3 Resource Requirements per Experiment

*if we need to bring in business analysts

Total cash for the validation phase ~$4-5K excluding engineering salaries; E6 audits recoup part of it.

3.4 Success Metrics by Phase

Discovery (weeks 1-4)

12+ interviews, 50%+ name the pain as a top-3 issue, 4+ audits with real data, a working read-only export, measurability map computed.

Paid Intent (weeks 5-9)

Cold-to-audit conversion 1%+, warm 10%+, 10+ demos total, 2+ paid audits or LOIs with a fixed fee (money or a signature, not "interested").

Stage 2 Promotion (week 16)

2-3 pilots with a signed measurement methodology, at least one has paid the fixed fee for a second consecutive month, sales cycle under 30 days in at least the warm channel, the $100K model recalculated on actual conversions.

3.5 Kill Criteria

3.6 Pivot vs. Kill Decision Framework

The Rule

Pivot a component, kill the bet. Bet = customer profile × pain × offer × channel × pricing model. Kill only when two or more irreplaceable components fail at once.

Pivots

Cold doesn't work, shift all weight to warm; reconciliation doesn't fit the window, bill on an operating metric with a quarterly true-up; measurability holds only at the top, move the ICP upmarket; the audit doesn't sell but the LOI does, drop the audit as a gate.

Kills

K1 and K3 together on day 30; K2 plus rejection of the interim metric; a 60-day sales cycle against an 84-day market median. Discipline: a pivot costs 2 weeks; at most 2 per window; a third required pivot equals a kill.

Section 4 · Path to $100K Revenue in 4 Months

4.1 First Sale Path

Goal: first paying customer within 14 days. Three approaches, run in parallel.

Option A. Holding Business Units as First Pilots

• Who: the e-commerce lead at a business unit with a Shopify store and repeat purchases. Introduced through the CEO — an internal conversation.
• Price & contract: pilot-quarter fixed fee $2,500-3,500/mo, no percentage during the pilot. Three months, read-only access, measurement methodology as a contract appendix, exit after reconciliation, right to use it as a case study.
• Role in the funnel: delivers the first sale within the 14-day window. It's an internal sale, so it's excluded from channel CAC and conversion — counted only as the funnel floor.

Script:

1. "We'll connect to your Shopify and Klaviyo read-only and show you, in dollars, within a week, how much revenue is leaking through overdue repeat purchases."
2. "Not impressed? We walk away, no obligations. Impressed? We launch a pilot, every campaign measured against a control group."
3. "The pilot rate is $2,500-3,500 a month; the holding doesn't pay a percentage during the pilot — we need the data and the first reconciliation, you get the revenue."

Option B. Warm Intro from a Klaviyo Agency

• Who: the founder of a mid-sized agency focused on retention for DTC. A shortlist of 15 is built (Flowium, Chronos Agency, Hustler Marketing, Sticky Digital, Underground Ecom, and others), priority on those with clients at $10M+ GMV.
• Price & contract: the same ladder for the brand; a wholesale rate below direct ARPU for the agency (set in negotiation, condition H5c). Three-way contract: agency owns the client, we own measurement and results.
• Role in the funnel: delivers the first warm intro, feeds the pipeline for months 2-4.

Script:

1. "We're not taking your clients or touching your retainer — we're the engine you resell under your own brand, same team, more clients served."
2. "To start, one intro to a client you don't have the bandwidth to do retention work for by hand — we'll run a paid audit for them, you keep the relationship."
3. "If it works for you, let's talk wholesale pricing — you take the margin on top."

Option C. Direct Shortlist with a Paid Audit

• Who: a retention lead or founder from the shortlist. Selection method: Shopify Plus, Klaviyo for 6+ months, niche in beauty/supplements/pet/coffee, size signals of $10-30M. Priority on brands with no agency, or unhappy with theirs.
• Price & contract: the audit fee is credited toward the first month of the pilot, then a pilot quarter at $2,500-3,500, stepping up the ladder after reconciliation. The audit is its own small contract (this is the H3a test), followed by the pilot quarter.
• Role in the funnel: rarely produces payment within 14 days (the cycle runs 14-30 days after the audit); it fills the pipeline with audits in progress.

Script:

1. "Brands with your profile are leaking 6%+ of annual revenue through overdue repeat purchases — we'll compute your number from your own data within a week."
2. "The audit is paid, but its cost is credited to your first month of the pilot if you continue."
3. "The pilot costs less than what you pay an agency, and the percentage is only on lift measured against a control. No lift, no percentage."

4.2 Founder-Led Sales Motion

I close the first 10 deals personally. Three role configurations:

1. Product owner solo on audits and agency negotiations — works because we're selling a number, not charisma; caps out at 2-5 signings a month.
2. Product owner plus CEO on holding pilots and anchor brands — the CEO lends the group's weight and answers the "will you still be around" question that matters for a contract with a quarterly reconciliation.
3. Product owner plus a technical co-founder on demos where the deal hinges on trust in the measurement (the core of H3b) — the person who built the methodology defends it. Default is configuration 1; holding and anchor deals get configuration 2; configuration 3 joins at the methodology stage of any deal that includes the percentage.

When to hire a salesperson, by signal: hire once 10 deals are closed, the script repeats without improvisation, H5a has passed, and the pipeline is consistently wider than the founder's own 2-5 signings a month. Don't hire yet, fix it instead: if the pipeline exists but audit-to-pilot conversion is below 3 of 10, a salesperson would just scale a broken motion.

4.3 $100K MRR Math · Honest Decomposition

Key fact: the percentage is zero in months 1-3, the quarterly reconciliation hasn't happened yet — all of the early MRR is fixed fees.

A realistic mid-point by month 4 is closer to $60-75K; the top of the range is a stack of best cases at the ceiling of our signing rate.

There is no self-serve tier: it requires a data connection, an agreed methodology, and a holdout contract — this isn't a card-and-checkout button. The $3-10M tail on the $2K fixed fee is a separate stream, not a priority within the 4-month window. Hitting $100K by month 4 needs everything at once: mid-tier only, holding pilots in months 1-2, ghost sends bringing the percentage in ahead of reconciliation on large bases. Otherwise, honestly: month 5-6 with the ladder, month 7-8 on the old flat fixed fee (which delivered $36-48K by month 4).

4.4 Why This Math Is Realistic

Comparables

There's no direct comparable with a disclosed ramp, said plainly. Indirectly: Sierra and Decagon (pay-for-result is growing fast in B2B, Sierra hit $20M in 8 months of selling) — evidence for the mechanism; Chargeflow (percentage of result, $10M+) — evidence the model survives; 11x as a cautionary tale (their revenue was found to be overstated), which is why we model our own MRR off fixed fees only.

Market Conditions

Klaviyo has 3,912 customers at $50K+ spend; after our filters, 1,200-2,000 brands remain; we need 15-20, which is 1-1.5% of the pool.

Workeron's specific advantages: warm holding pilots in months 1-2, the group's weight in contract negotiations, an assembled agency list for white-label with near-zero CAC. Risks: signing pace (ceiling of 2-5 vs. the 8-12 needed — the binding constraint), H3b has no precedent, the quarterly reconciliation pushes the percentage past the window.

4.5 Sensitivity Analysis

Rows come straight from the MRR model (they account for the reconciliation lag and the signing-rate ceiling), not a linear recompute off the plan.

In every branch the company doesn't die: fixed fees cover costs, the timeline shifts, not the outcome. The model is more sensitive to signing pace than to price, so the channel (H2b, H5a) is what we protect first. Note the differing baselines: $75-90K is the pure mid-tier plan, while the $55-90.5K in 4.3 is the full decomposition including anchors and the services tier.

4.6 Honest Failure Mode

The likely shortfall is pace, not market rejection: holding pilots start in months 2-3 due to internal approvals, the external channel delivers 2-3 signings a month instead of 8-12.

By How Much

$40-70K by month 4 instead of $100K, 30-60% below target.

What We'd Still Have

2-3 live pilots with a signed methodology; the first quarterly reconciliation just ahead (soon a proven dollar lift — the key asset for months 5-6 sales); H0-H3a validated on real data; profitable unit economics on the fixed fees.

Conclusion

$40-70K by month 4 is a working trajectory to $100K by month 5-6, not a failure. A real failure shows up earlier and is caught sooner: H3a (0-1 of 10 pay) or H4 (5% lift or below) — both kill signals trigger before month 4.

First Experiment & Sales Entry Point

The lost-repeats audit (read-only Shopify + Klaviyo + CRM). It simultaneously qualifies the brand against every filter and opens the sale: we walk in not with a pitch, but with the dollar figure of money left on the table on a slide. It kicks off the H1 → H2b → H3a chain.

Out of Scope for Regulars v1

1. Complex service workflows and dependencies.
2. Salons and services as a vertical (a v2 candidate).
3. Medical and regulated scenarios.