The 2× Playbook: Eleven Industries

You are not buying a company. You are buying a filesystem that does not yet know it is a filesystem — invoices, schedules, renewals, call logs, work orders, half of them retyped by hand between systems that don’t speak — and the gap between those two facts is the return. That sentence, which began as a note in our acquisition research, turned out to be the whole thesis. This paper is the long version: which industries the math favors, which levers move first, and the sequencing rule that separates the doubles from the obituaries.

The arithmetic of the double

Start with the anatomy of a typical lower-middle-market service firm: call it five million in revenue at a twenty percent margin. Labor runs half to seventy percent of its cost base. Inside that labor, a large fraction — thirty to sixty percent depending on the vertical — is structure-shaped work: collecting, reconciling, scheduling, drafting, routing, reporting. Structure-shaped work is what systems now carry well. Remove a conservative slice of it and the cost line moves more than most owners’ last three initiatives combined.

Now the revenue side, which is the half the automation debate forgets. These same firms leak demand chronically — calls unanswered, leads unworked for days, renewals unmanaged, accounts never rounded. The machinery that fixes cost also fixes capture: instant response, systematic follow-through, renewal and attach run by rule. Put both moves on the same bridge and the generic case lands near six and a quarter million at a thirty-plus percent margin. The earnings have doubled, and nothing exotic happened — no new market, no new product, no heroics. Arithmetic, sequenced.

The inefficiency you delete is the value you create.

The filter: flow-through beats disruptable

The intuitive way to rank industries for this play is by automation surface — how much of the work AI can carry. The intuition is wrong, and the study’s clearest proof is residential solar: the largest automation surface of the eleven, and a rank near the bottom anyway. Thin, policy-whipsawed margins and a cancellation-prone pipeline mean the gains leak out before they reach the earnings line.

What actually predicts the double is flow-through: how much of each recovered dollar survives to EBITDA, and how much of each EBITDA dollar survives to the exit multiple. Recurring revenue, pricing power, licensing moats, and low churn are flow-through; project work, subsidy dependence, and bid-market pricing are leaks. The best 2× targets are not the most disruptable — they are the most flow-through.

Three shapes of the opportunity

Ranked this way, the eleven industries sort into three archetypes. The Recurring Compounders — pest control, HOA management, commercial insurance — already own the annuity; the work is raising retention, attach, and recurring mix until earnings and multiple move together. The Throughput-Constrained Producers — SOC 2 audit, recruiting, franchise placement, data services — have more demand than licensed or skilled capacity; the work is releasing capacity with systems and selling every hour it frees. The Thin-Margin Turnarounds — bookkeeping, sales training, solar, real-estate brokerage — cannot cut their way to the double; the work is changing what the business sells: attach, productization, or model shift.

The ranking

Tier one, where the math runs cleanest: residential pest control — the textbook case, roughly three-quarters recurring, with booking rates at the front desk so far below the ceiling that the fastest EBITDA move touches no marketing budget; SOC 2 and compliance audit — the purest automation alpha in the study, because the compliance platforms already deliver evidence by API and the audit firm is the last manual node before a legally required signature; HOA management — a per-door annuity whose binding constraint, manager capacity, is exactly what back-office automation manufactures.

Tier two, real doubles with one hard problem each: commercial P&C insurance, where retention and account-rounding move earnings and book value at once; franchise brokering, a speed-to-lead business answering an eighth of its inquiries promptly in a market where conversion flows almost entirely to profit; client accounting, where the automation gains are peer-reviewed rather than promised and the advisory attach is worth multiples of the compliance retainer; sales training, whose product famously decays within a month — until AI roleplay makes reinforcement, the actual product, deliverable at software cost.

Tier three, for operators who know exactly what they are doing: BI and data services, where the escape from the rate card is productizing the project exhaust; recruiting, where systems double desk throughput and the prize is the contingent-to-retained migration; residential solar, a distressed-entry-only consolidation; and real-estate brokerage, where a 1.7% median margin leaves nothing to cut and the entire play lives in the eighty-percent-margin attach economics next door. Each of the eleven has its own page in our industries section, with the levers named and priced.

The organization the math builds

Run this playbook to its conclusion and a specific org shape emerges — three layers. A thin human trust layer faces the market: the people who answer for the work, whose names carry the relationships. An agentic production core carries the structure: research, drafting, scheduling, reconciliation, reporting, running at machine scale. And a human signature layer gates everything that leaves the building: the decisions, the approvals, the accountable yes. AI does the crystallization; trust stays human. Companies that delete the third layer to buy throughput spend their reputation to save a salary — the one trade this model forbids.

The Bench rule

The sequencing discipline deserves its own name, and the sector already paid for the lesson. A venture-backed bookkeeping firm raised nine figures selling software prices on largely human delivery, betting automation would catch up to the pricing. It did not. The company burned through the capital and collapsed abruptly, stranding thousands of small-business clients. The rule its corpse teaches: automate first, then price and sell. Efficiency must be real before it is priced, and priced before it is scaled. Every failure mode in this playbook is some way of running that sentence backwards.

Evidence and honest caveats

The pattern has public proof points and they deserve their asterisks. An AI-native accounting startup reports gross margins around sixty percent against a mid-teens industry norm — self-reported, but directionally consistent with the controlled studies. A Texas holding company reports nine-figure earnings within two years of applying the model to community management at scale — unaudited, and we label it so. The per-industry benchmarks in this study mix peer-reviewed research, regulatory data, and operator interviews; where a number is self-reported, the industry page says so. We would rather publish a smaller number with a source than a larger one with a shrug.

For the owner; for the acquirer

If you own one of these businesses, the uncomfortable version of this paper is that the repricing arrives whether you participate or not — the only open question is whether the double accrues to you or to whoever buys you at the old multiple. The dignified move is to run the playbook yourself, or at minimum to instrument the business so the value it already creates becomes visible before a buyer prices its absence.

If you are acquiring, the discipline is the filter: underwrite flow-through, not automation potential; price the trough, not the story; and sequence like the rule says. The gap between a filesystem and a company that knows it is one is the most honest arbitrage available in the lower middle market this decade. It will not stay unpriced for ten years. It is unpriced now.

Sources & method

This publication condenses an eleven-industry internal research program: composite feasibility scoring across automation surface, flow-through, demand leakage, regulatory moats, and exit-multiple mechanics, built from published studies, regulatory filings, industry benchmarks, and operator interviews. The full scoring appendix travels with the working file, and the source behind any figure quoted here — including the ones we flagged as self-reported — is available on request.