The new business, in plain English

This mirrors the full working plan, written so you can read it without decoding anything. Everything here has a source and a date behind it in the long version. Nothing in it is decided.

4 August 2026 · read this, then answer the questions

→ The 22 decisions waiting on you

Most have a recommendation attached. Three of them block the next piece of building — and four are new as of 5 August, from the billing report.

→ The verification verdict — read this one first

Six researchers, six skeptics, 449 source checks. It went against us. What survived, what was fabricated, why this might not work as a business, and the legal finding that changes the pricing model.

→ Work done, never billed — the original report

The biggest lever, taken apart: what it looks like, why it happens, how we'd fix it, whether we can take a percentage, how long the money takes to arrive. Partly superseded — it carries a banner listing what died.

Where we landed

We sell an operating layer that goes inside a mid-sized service business and finds three things: money they've already earned but aren't collecting, revenue that never became a job, and work they're paying people to do that a system should do. We run it, we prove every dollar in a shared ledger, and we take a share of what we produce.

They pay nothing up front except a refundable commitment that proves they're serious. No retainer. No lock-in clause. They can leave whenever they want.

The correction that rebuilt this whole plan Until today the plan was built outward from our costs — payment fees, duplicate software, Rizza's hours. You caught it twice. Reasoning from our own inefficiency quietly capped the business at the size of our own inefficiency. It's now built from the target company inward, and the numbers moved by more than a hundredfold.

Where the money actually is

This is the part that changed most. On a restoration company, the two biggest costs are field crews and the estimating software the insurance carriers effectively require — and we can't touch either one. Software and payment fees, which the old plan led with, come to about 3% of what's available.

Here's the real map, on a $15 million independent restoration company.

WhereWhat it isYear one
Money already earned, not collectedSupplements and overhead-and-profit they're entitled to but don't document · files that survive a carrier audit · equipment days they can't prove · unpaid deductibles · the handoff from drying the house to rebuilding it$250–500K
Revenue that never became a jobMissed and after-hours calls · the referral engine nobody runs, when 93% of their leads are referrals · marketing money pouring into the most expensive channel they have$200–600K
Cash sitting stillInsurers pay in 60–120 days. The best operators get it to about 32.$120–150K
Cost and capacityBack-office labour · crews losing half a day a week to admin and driving · lost equipment · and, small: software and card fees$180–500K
Knowing anything62% of them cannot calculate profit on a single job$150–450K
Total we'd find$900K–$2.2M
What we could defensibly claim$400–900K
Our share at 20–30%$80–270K per client, per year
The problem I told you was fatal isn't fatal I said cost savings run dry after year one. That's true of the software-and-fees group and false of everything else. Claims recovery happens on every job, forever. The referral engine dies the day nobody runs it. AR discipline is a standing practice. Which means you may not need the messy profit-share switch at all — a straight share of savings stays clean to measure and stops running out.
And the thing you won't like I wrote a "who else already sells this" line under every single lever. The answer is never "nobody." Supplements have five vendors and the maker of Xactimate itself. Missed-call answering is free inside software they may already have. There's a shop advertising restoration virtual assistants at $8/hour doing supplement recovery — which is your model, aimed at your target, already running.

So we don't win on any single piece, and we shouldn't pretend to. What nobody sells is all of it, joined up, run by someone accountable for the number, on one contract. The owner is expected to buy nine subscriptions and wire them together himself, and he doesn't, because he's running a restoration company. That's the business.

Who we go after

About 9 in 10 restoration companies are independent — roughly 54,000 of them against about 6,000 franchise locations. That closes a worry the plan was carrying: avoiding franchises costs us about a tenth of the market, and it's worth it, because franchisees are contractually forced onto their franchisor's software and run 8–14% lower profit on royalties they can't cut.

On size: $5 million and up for the full offer. A $2M company yields maybe $12–30K a year to us while the audit, the build and the operator cost almost exactly what they cost at $15M. That's break-even at best. But $1–3M isn't wasted — it's the productised, low-touch tier, which is the community idea you raised on Sunday. Thousands of potential members instead of hundreds of clients.

On vertical: restoration has your story, but insurance sets their prices and cash takes four months. Foundation repair is customer-paid, so nobody controls their price, margins are better, cash is fast, and their main constraint is cost per lead — your decade, not something we'd learn. My suggestion is to lead the marketing with restoration and run the first paying pilot on foundation repair.

What we charge, and for how long

A refundable commitment first — around $7,500 for smaller companies, more as they get bigger. It's not a fee and not a retainer. If we don't hit the number in 90 days they can take it back. It exists so they commit.

The clause that matters most in the whole contract "Saved" has to mean found, documented and agreed — not in the bank. Insurance takes 60–120 days to pay. If we define it as cash received, we'd end up refunding people on engagements that genuinely worked.

Then a share of what we produce — and this is where your rolling-clock idea lands, with one change.

Things we hand over (a switched payment processor, a killed subscription) run whether we're there or not, so only a clock keeps us paid. Things we operate (claims recovery, the referral engine, AR chasing) stop the month we stop — so they're already permanent and already cancel-anytime, with no clock needed. On a restoration company the handed-over stuff is that 3% again.

Everything on 24 monthsEverything on 5 yearsSplit them
Five-year total~$394K~$894K~$803K
ShapePeaks year 2, then collapsesGrowsGrows
Can they cancel?Only by waiting it outOnly by waiting it outAny time
How hard to sellMediumHardestEasiest

The first column is the trap: a pure rolling-clock model pays us for finding and nothing for running. The moment we stop finding new things, revenue decays to nothing no matter how well the systems are still working — and findings taper on every account.

Why I'd skip perpetuity A permanent share of someone's profits is an encumbrance a buyer finds in due diligence and drops the price for. And we're planning to tell owners "every dollar of profit we add is worth three to five dollars when you sell," because restoration is being actively bought up right now. Those two sentences can't both be in the deck.

What sales resistance actually is — from your own clients

I went into your business files for this. It's never about how long the contract runs. It's about paying when they can't see anything happening. Bob Brown, more than once: "Kate has produced nothing recently" and "you are evading my question about costs." Mike, in July: spending is "very, very high in actuality compared to what we discuss in the meetings." And your own words about that account — "they're good until the end of the year currently." Annual horizons. Not five-year ones.

So the rule is: never let the fee outlive the visible work. A fee on savings from eighteen months ago is a fee with nothing attached to it — which is exactly the conversation Bob was already having with us on a live engagement.

How we sell it, and how we report

The thing we sell with and the thing we report with are the same picture — the web you already have running here. They pitch on a web of what they're losing, then watch it fill in with what we actually recovered.

Then there is no report. There's a live link that's always current — a report can be late, a live page can't. Monthly is one line and one link. Quarterly, new strands appear, which is your rolling line-item idea made visible and the proof we're still finding rather than coasting.

The single best thing we could build When one of the handed-over items reaches the end of its clock, it greys out on its own and the fee drops — without them asking. A fee the client watches go down on schedule is the strongest possible proof the countdown is real. Bob's question can't even start against a page where every charge shows its own end date.

What could kill this

→ Now the decisions

Twenty-two questions, most with a recommendation you can just tick. Answers save as you type — hit "copy all" when you're done and paste them back to me.