About This Model
The delivery discount is easy to assert and hard to picture. The mechanism sits across three places a reader rarely holds at once: the intake, the cost base, and the exit multiple.
So this document builds a company from published benchmarks — cost structures, operating ratios, valuation adjusters — applied to a profile common enough in vertical B2B software to be recognisable. It carries the discount. It goes through resolution. The arithmetic runs line by line, and the sensitivities that break it come before the engagement economics.
The Company: Entrenched, Growing, Worth Less Every Year
A vertical B2B software group at roughly $20M revenue and 150 people, four real management layers: the 100–249 stage, where the binding constraint is that nobody can hold the whole picture any more. Four products, two of them acquired. The oldest platform is 25 years old, single-tenant, serving customers whose operational processes were built around it a decade ago. Nobody leaves voluntarily. The moat is real.
It passed through the 50–99 stage without graduating from it: no function ever took ownership of the intake, and no one owns the boundary between the core product and its services layer. That boundary is the mechanism behind every line in the table.
| Line | Entry | Basis |
|---|---|---|
| Revenue | $20.0M | Profile assumption |
| Revenue quality | ~75–80% support and maintenance, not subscription | Common in pre-SaaS vertical platforms on legacy pricing |
| Billed services | ~20% of revenue, at or below break-even | Sector norm break-even to 5% margin; custom-heavy vendors run negative |
| Product development | ~30% of revenue | Equity-backed norm 30–34%; efficient quartile nearer 10–15% |
| Support and success | ~12% of revenue | Against a ~9% benchmark |
| Hosting | ~7% of revenue | Legacy single-tenant estate; modern estates ~5% |
| ARR per head | ~$115,000 | Median ~$141k; top tier $200k+ |
| Blended gross margin | ~60% | Below the 70% threshold at which software multiples apply |
| Gross revenue churn | ~10–13% | Consistent with delivery-driven incidents |
| EBITDA | ~−$1–2M, breakeven to slightly negative | Follows from the cost lines above |
Run the thesis report’s six indicators against this company and all five leading indicators fail: ARR per head well below the ~$141k median while revenue grows; product development at the top of the band with no roadmap surface to show for it; nobody can say what share of last year’s commitments shipped without touching the core; support three points above norm while churn holds; and commitment-to-release time lengthens on quarterly releases. The sixth, services mix at 20%, is the lagging one the company already knows about. It has been running the cleanup for two quarters.
Why the Cleanup Misses
Assume the cleanup succeeds on its own terms. Services repriced to break-even, delivery labour reclassified, eight quarters restated, billed services down from 20% of revenue to 14%. The income statement is materially cleaner. Blended gross margin moves from roughly 60% to roughly 63%. It does not reach 70, and the reason is the point of this document.
Of the 30% of revenue in product development, a meaningful share is not product development. It is customer-specific work absorbed into the core: conditional behaviour, client integrations, fields that behave differently for one account. It was never billed as services, so the cleanup cannot find it. On this profile, a third to a half of the product-development line is absorbed customer work: $2–3M of a $6M spend. The company now reports 14% services against a clean separated P&L, and a buyer still underwrites it at services multiples, because the cost of delivering the next customer has not moved.
The Operating Bridge: 60% to 80% Gross Margin
Resolution follows the order in the companion documents — mandate, casting, team, rebuild and routing — over 24–36 months. Each line names its mechanism so it can be disputed on its own.
| Line | Entry | Resolved | Mechanism |
|---|---|---|---|
| Revenue | $20.0M | $22–24M | Tail pruned; retention and renewal repricing offset it. Modest, deliberately |
| Product development | ~30% (~$6.0M) | ~12–15% (~$3.0M) | Absorbed work leaves the core for the ring; cost base repositioned; genuine roadmap spend remains |
| Configuration delivery | inside the lines above | ~3% of revenue | Configuration authored declaratively under a human gate: where AI changes the unit economics |
| Billed services | ~20% at break-even | ~8–10% at 35–40% margin | Residual genuinely bespoke work, priced, outside the boundary |
| Recurring mix | ~20–25% true subscription | ~85–90% | Support repriced as subscription at renewal, over an 18–24 month cycle |
| Support and success | ~12% | ~9% | A standard core generates fewer incidents; ring changes removable without residue |
| Hosting | ~7% | ~2–3% | Single-tenant estate to multi-tenant, event-driven infrastructure |
| Release cadence | Quarterly | 6x against the sector norm: 2 production releases/month | Measures how often the product ships, not output per engineer |
| Blended gross margin | ~60% | ~78–82% | Composite of the lines above |
| EBITDA margin | ~−5% to 0% | ~25–35% (~$5.5–7M) | Composite of the lines above |
The Cost Bridge, Lever by Lever
The composite decomposes by lever. Components are ranges and are deliberately not summed: they overlap at the edges, and false precision is how modelled work becomes dishonest.
| Benefit | Owner | Baseline at entry | Measure |
|---|---|---|---|
| Single-customer work out of the R&D line | Product boundary owner | Share of development spend on single-customer work | Same, quarterly |
| Services repriced and separated | Commercial lead | Services mix and services margin | Same, quarterly |
| Hosting on a modern estate | Engineering lead | Monthly hosting cost at the signed baseline | Same, monthly |
| Support at sector norm | Support lead | Support and success as % of revenue | Same, quarterly |
| Churn from delivery incidents | Customer lead | Gross revenue churn | Same, annual |
| Cost: pruned tail accounts | Commercial lead | Revenue and headcount carried on the tail | Revenue lost against margin recovered |
Every row is modelled. Cash lines reach EBITDA; the rest move the multiple. They are not summed. Where the platform is replaced mid-hold, the report states which measures stay comparable across the change and which do not.
| Lever | Contribution | Mechanism |
|---|---|---|
| Experience arbitrage | ~$1.5–2.0M | Senior engineering and product capability at 40–55% of tier-1 cost, on the company’s books |
| AI-assisted modernisation | ~$1.0–1.5M | ~25% effort reduction on the remaining base once the ring absorbs configuration under the gate |
| Hosting and infrastructure | ~$0.8–1.0M | Event-driven multi-tenant estate replacing single-tenant instances |
| Services repricing | ~$0.8–1.2M | Residual bespoke work moved from negative margin to 35–40%, priced and contracted |
| Tail rationalisation | ~$0.2–0.4M net | Bottom-tail accounts wound down; support, CSM and custom-work capacity released. Net of gross profit lost |
The Cash Shape: Phase by Phase
The bridge does not arrive in a straight line, and a board that has not seen the shape in advance treats month 6 as failure. Costs rise before they fall: diagnostic fees and retention packages land first, transition waves double-run for a quarter each, and the unprofitable tail is wound down before its capacity is redeployed.
| Period | Revenue | EBITDA margin | What is happening |
|---|---|---|---|
| M0–6 | $20M → ~$19M | −5% → −7% | Phase 0 fees and retention packages committed; wave 1 hires double-run with incumbents; tail culling begins and takes revenue out first |
| M6–12 | ~$19M | −4% → ~0% | Waves 2–3 land; legacy freeze releases capacity; arbitrage savings begin to outrun transition cost. EBITDA crosses zero ~M12 |
| M12–18 | ~$19–20M | ~5–10% | Rebuild at full cadence; hosting migration lands; modernised platform releases ~M18 |
| M18–24 | ~$20–22M | ~15–20% | Client migrations run; renewal repricing converts support to subscription; services reprice at renewal |
| M24–36 | $22–24M | ~25–35% | Run-rate. Recurring mix reaches 85–90% as the repricing cycle completes |
EBITDA margin through the transition · modelled
Costs rise before they fall: Phase 0 fees and retention land first, waves double-run, the tail is wound down. Savings then outrun the transition, and repricing builds the run-rate.
Modelled on the profile in this document. Half-year detail in the table above.
Enterprise Value: Entry to Exit
Entry and exit are priced on different instruments, and stating both is what makes the arithmetic legible.
At entry the asset is breakeven, so no EBITDA multiple applies. Entrenched vertical software of this quality trades on revenue, between the ~1.1x IT services and ~2.3–2.6x non-SaaS software medians: at 1.5x revenue, entry enterprise value is approximately $30M. At exit the asset has $5.5–7M of EBITDA at 25–35% margin, a standardised core shipping twice a month, and 85–90% recurring revenue. Lower-mid-market software of that shape clears around 9–12x EBITDA, and larger private deals 13–19x; this model prices the exit at 10–11x, deliberately below current private medians. Central case: ~$5.5M × 10.5x — approximately $58M, with the bridge’s modest revenue growth sitting above it as headroom, not summed into it.
Modelled · central case · $M enterprise value
~1.9x on enterprise value, before fees and structure.
Revenue growth in the bridge sits above the central case as headroom; it is not summed.
Roughly 1.9x on enterprise value over 24–36 months, before fees and structure. The dominant channel is EBITDA creation, not re-rating, which is what insulates the case against exit-multiple compression.
| $4.0M EBITDA | $5.5M EBITDA | $7.0M EBITDA | |
|---|---|---|---|
| 9x exit | $36M | $50M | $63M |
| 10.5x exit | $42M | $58M | $74M |
| 12x exit | $48M | $66M | $84M |
Entry EV $30M throughout; read any cell against it for the multiple on enterprise value. The grid spans 1.2x to 2.8x. The bottom-left cell, where EBITDA disappoints and multiples compress, still returns capital. The downside is flat, not fatal, and that is the property that makes the case underwriteable.
Nine Sensitivities to Run First
The sensitivities matter more than the base case. Run them before anything else, and read each as an outcome.
- Grant no mandate, and every line in the bridge becomes a recommendation, dead at the first renewal-pressure exception before anything technical is attempted.
- Let a material account take a boundary exception, and absorption resumes. The architecture survives in name; the discount returns.
- If gross churn holds above 10%, stop. Every other line can succeed and the asset still declines. This bridge fixes margin, mix and predictability; it improves retention only where churn was delivery-driven.
- Start the transition without written retention, and the two people who know why the platform was built that way now decide your timeline. Retained in week 4, or do not start.
- If configuration replaces custom code at half the assumed rate, gross margin lands at 77–79% and product development at 18–20%. The thesis holds; the above-80 claim does not.
- If a third of the support base declines conversion at first renewal, recurring mix lands nearer 70% and the classification uplift arrives a cycle later. Margin improves regardless.
- If the engineering baseline is understated by 20%, the largest line in the bridge is mis-sized and the EBITDA arrival point moves by roughly a year. Cost the baseline bottom-up before committing.
- Point new capital at sales before the intake, and commitments arrive faster into an unchanged intake. The condition compounds and the bridge never starts.
- Let the mandate quietly lapse after a board reshuffle, and everything below step 1 reverts. The least discussed failure in this class of work, and the reason decision rights are papered by resolution.
Illustrative Engagement Economics
The shape of the engagement that executes this bridge. Illustrative, provisional, and subject to counsel and CFO confirmation in any real structure.
- The delivery team is hired onto the company’s own books: permanent payroll in R&D and cost of goods. No agency margin, no annuity. The cost base in the bridge is the company’s, not a vendor’s.
- The operating fee is a separate, clearly named professional-fees line, structured as non-recurring transformation cost. A quality-of-earnings review adds it back, so the fee is invisible to the exit multiple. A fee blended into run-rate OPEX is not.
- The carry is tied to incremental EBITDA against the day-zero baseline, vesting over the engagement, paid at exit or refinancing. The operator’s wealth event is the same event as yours.
- The founder variant is a fractional engagement under professional fees with a success fee or equity participation at the raise, where the diagnostic underlying this bridge is what prices the raise as software.
One Starting Move
Run your own numbers into Exhibit 1. If the five leading indicators fail, the shape applies whatever the sixth — the services line — says. If product development is high and the roadmap surface is thin, the absorbed quantity is there, and it is findable in a fortnight.
Then run the test that settles it faster than any financial line: take last year’s customer commitments and sort them into three piles — shipped by the standard product, shipped by configuration, shipped by changing the core. Most companies cannot produce the sort. The ones that can usually do not like the third pile.
EstimatorRun your own revenue, margin and services mix against the published ranges. Run the estimator
Take-away packThen put your six indicators against the benchmarks. Download the scorecard
Take-away packReady to test it on your own backlog? Download the three-pile audit worksheet
A Reinessence Group company.
Method. Every figure in this document is modelled; the profile in Chapter 2 is constructed, not observed. Operating benchmarks: SaaS Capital, 2026 Spending Benchmarks for Private B2B SaaS; Benchmarkit, B2B SaaS Performance Metrics; SaaSRise, 2026 SaaS Benchmarks Report and The 2026 SaaS Headcount Reset; Aleph, ARR per Employee 2026. Valuation: Software Equity Group, 2Q26 SaaS M&A and Public Market Report and gross-margin and NRR cohort analyses (2024–2025); Aventis Advisors, SaaS, Software and IT Services Valuation Multiples 2015–2026; Hampleton Partners, Enterprise Software M&A Report 2H2026; Multiples.vc, Public Software Valuation Multiples (October 2026). Where surveys disagree, the driver is funding model, and the ranges reflect that rather than selecting the convenient figure. The ~25% modernisation-effort assumption and the hosting range reference the delivered engagement described in the thesis report, where the hosting reduction was measured on a comparable load, on a greenfield product, and covers monthly hosting only, not total cost of ownership. Re-In-Essence classifies claims as realised, operational or modelled and never sums the classes. Nothing in this document is realised or operational.
Important information
This report is for general information only. It is not investment, financial, legal or tax advice, and it is not a recommendation about any particular company. Market figures are published third-party data, cited as their publishers report them. Modelled figures illustrate stated assumptions; they are not forecasts. Results from past engagements do not predict results elsewhere. Nothing in this report is a promise or guarantee of any outcome. Take advice from your own advisers before acting on it. Full terms.