Portfolio OS.
The operating model.

Four Functions Installed, Two Aligned

The operating system has six functions. Four are installed in strict sequence, each holding because the one before it is in place. Two are aligned rather than installed: Strategy arrives as the board mandate, and Sales and Customer Success are pulled into the product cadence and reviewed as an operation. Taken together, the four installations are the target operating model and the order of operations is the plan that gets there — stated here in plain terms rather than a named framework.

Six functions · four installed, two aligned

Each function holds because the one before it is in place. Sales is where the restored asset converts.

1StrategyAligned — sits with the board and the mandate

2ProductInstalled — evidence-led cadence, one sprint ahead

3PeopleInstalled — leaders, not implementers

4OperationsInstalled — published calendar, bug budget

5EngineeringInstalled — core, ring, edge under the gate

6Sales / CSMAligned — pulled into the cadence; reviewed as an operation

Strategy is granted, not installed: it arrives as the board mandate. Sales and CSM are aligned through the product cadence — the function that pulls the right people into one room with the customer — and reviewed as an operation, end to end.

3.1 Product and Org Design: From Pro-Serve to True SaaS

The first installation reshapes authority, not process. Companies at this stage are staffed with implementers — people whose job, as lived, is to say yes to whatever the customer asks. The shift is to an opinionated product function with the standing to say no — and the mechanism that makes the no commercially survivable: the request routes to the configuration ring or the edge, and the maintenance obligation is negotiated on the customer’s side of the boundary, priced and capped.

The mechanism, concretely. A 3-phase product flow — product flow, user stories, wireframes — with a minimum of two iterations per phase, a documented ready-for-dev gate, and the cadence running one full sprint ahead of engineering. A governance overlay on the discovery tooling, where the leadership review fires only when a customer-evidence pack is attached to the ticket: the structural fix for opinion-driven roadmaps, and the gate through which enterprise sales can no longer bypass technical reality. A client committee segmented by ICP, because the right research method differs by tier: deep-dive interviews for enterprise accounts, A/B prototype tests for mid-market, usability testing at the volume end. And a standardised user-story format as the contract between Product and Engineering, which is where spec-mismatch rework goes to die.

This is the empty chair from the thesis document, filled — by a pairing at the intake: one person accountable for what and when, one for how and how much, neither able to commit on the other’s behalf.

Move Capability Up, Not Just Cost Out

Org design applies to teams, not only to people. Cost programmes move work one way: out of the business, into a shared centre, priced per transaction. That is right for commodity work and wrong for the team holding the product knowledge.

In a fatigued asset the delivery function treated as a cost centre is often the deepest reservoir of platform knowledge in the company. The stronger move is upward. Its standards become the group’s standards, its senior engineers review the other teams’ architecture, and a cost line becomes the capability the rebuild runs on.

The move holds only if it is more than a rename: standards someone is obliged to follow, reviews someone is obliged to act on. Adoption by respect is where it starts. Written authority is what makes it survive a leadership change.

3.2 People: From Implementers to Opinionated Product Authority

Six components, none novel; the discipline is the installation. Build leaders, not implementers: senior leadership’s job is growing other leaders, not being the last decision-maker on every question, carried by a structured delegation framework, named escalation routes and an explicit decision-rights map. Time management and delegation, taught as installable practice rather than motivational coaching. Communication lines: senior leadership accessible at structured moments (demos, kick-offs, mid-reviews) rather than as the default escalation route. Documentation culture, so institutional knowledge becomes portable. A leadership-development model that produces the next generation of senior operators inside the company. And a weekly cadence — feature planning, programme delivery, release management, process governance — as the operating heartbeat without which nothing else holds.

Underneath the components sits the culture work: intellectual honesty. In organisations that drifted into reactive mode, the people who make the most sense are routinely judged as difficult. Changing what evidence is required, who gets heard and how disagreement is handled is the real installation, and it is installed by the senior team modelling it. People are then assessed against the role as it exists now, not as it existed at the last stage.

3.3 Operations: From Firefighting to a Published Calendar

Process installs only where an owner exists to hold it; otherwise it decays, and the decay is misread as proof that process does not work here. The roadmap becomes a published release calendar. Capacity is planned monthly: squads, tech leads, vacation factoring, QA and release planning in one document the team owns. A published bug budget makes remediation a planned line item. Bugs are a planned cost, not an exceptional one.

The structural move underneath is the team model: from one centralised queue, where every roadmap competes for the same engineering bandwidth, to product-led autonomous teams, each owning a product surface with its own engineers, tech lead and roadmap. It is a textbook scale-up move that is rarely executed inside a 10-year-old organisation, because the political cost of breaking up a unified resource pool is real. Once through, dependencies get named instead of implied, and senior engineers stop spending their time on cross-team mediation. That is what makes the cadence survive the next set of competing demands.

3.4 Engineering: From Custom Branches to Core, Ring, Edge

The three layers

The core — lockedProduct code only, standardised, identical for every customer. Strict typing, full test coverage. No single-customer edge case touches it.

The ring — configurationCustomer-specific need absorbed as rules and data, never code branches. Cleanly separated — exactly where agentic tooling builds custom configuration cheaply and safely.

The edge — priced and ownedBespoke integrations, adapters, legacy bridges. Outside the main application, priced, under its own statement of work, with a named owner.

Customer requests are never refused; they are routed. Nothing lands in the core uninvited.

With the first three disciplines secure, automation is applied to a model that works. The product restructures into a hard common centre, genuinely identical for every customer; a ring absorbing customer-specific need as configuration, rules and data, never code branches; and an edge for genuinely bespoke work, priced, under its own statement of work, with a named owner. Three tests distinguish a real ring from a claimed one: the share of last year’s commitments satisfied inside it with no core change, whether a ring change can be made by someone who is not a core engineer, and whether it can be removed without residue.

The Fleet and the Gate: AI Under Governance

Agents make the ring cheap: ring work expressed as declarative configuration is bounded, specified, testable against a known contract, exactly what language models do well. Applied without the boundary, the same tooling writes technical debt faster.

Senior operators direct multiple agents in parallel across code generation, migration, testing and documentation, every output passing a human gate before production. We do not cite an operator-to-agent ratio as a deliverable; the number rises with tooling and falls under the gate, and what we underwrite is the release cadence the fleet produces. Five disciplines govern the gate: a curated golden-path infrastructure context; model tiering with spend visibility; enforced component reuse; an AI-mapped codebase where every change carries its plan; and one managed control surface for models, prompts and tools.

Without the gate, agentic development produces faster slop. Build the gate first, then scale the fleet.

Three Levers, in Order of Execution Certainty

Three levers · in order of execution certainty

L1Cost arbitrage + modernisationLow risk · requiredSenior capability at 40–55% of tier-1 cost; the core modernised in place. The base case stands or falls here.

L2Revenue expansionMedium risk · optionalityCross-sell and expansion once delivery trust is restored. Upside, never load-bearing.

L3Platform integrationHigh risk · laterIntegration across assets at a premium exit multiple. Deliberately sequenced last.

Nothing in the engagement economics depends on L2 or L3 arriving.

For a sponsor, the model decomposes into three levers, and it is built so the case stands on the first. L1, cost arbitrage plus AI modernisation: low risk, required. The base case stands or falls here. L2, revenue expansion and cross-sell once delivery trust is restored: medium risk, upside optionality. L3, platform integration across assets at a premium exit multiple: high risk, and deliberately later. Nothing in the engagement economics depends on L2 or L3 arriving.

Phase 0: The First 100 Days

Phase 0 is the standardised diagnostic entry point. We do not begin by guessing at the timeline, the team shape or the financial upside. We map it, on a fixed cadence with named outputs.

Window	Activity	Output
Days 1–7 · Listen	Cross-function sessions, leadership 1:1s, customer calls, P&L baseline	Listening report and scorecard
Days 8–30 · Diagnose	Four-function diagnostic; redirect envelope quantified; knowledge-holder map built	EBITDA plan, draft
Days 31–60 · Validate	Plan refined against board constraints; sponsor reviews; Phase 1 scoped	Phase 1 proposal and hub plan
Days 61–100 · Mobilise	First 3–5 senior architectural hires; product cadence installed	Phase 1 begins

The measurement baseline set in these hundred days is the instrument that converts operational work into an attributable value-creation narrative at exit (Chapter 6). Three things we deliberately do not do: no mass restructure announced on day one, which signals the sponsor has given up and craters engineering retention; no premature mass hiring ahead of the organisation’s capacity to absorb it; no public commitment of the plan before the diagnostic is complete, costed and defensible.

Week One, and What the Read Actually Is

The mandate is week one. The design is not week two. Week one is deliberately narrow: meet the executive team individually, meet the top five customers’ relationship holders, read the last four board packs and the last two quarters of delivery data, and change nothing. The first artefact is a one-page statement of what will be measured and against what baseline — not a plan.

The read that follows asks the same six questions at every team and every product line, so the answers can be compared. In a group built by acquisition, the differences are the finding.

Question	What it exposes
Walk me through one real enterprise commitment, from the sales call to the release that carried it	The actual route, not the policy. Ask for a named one
Who decides what enters the core product — and who has to agree before they can?	Whether the boundary has an owner or a committee
Why is this work done by this team and not another?	Structure inherited from deals rather than designed
What gets rekeyed, and into what?	Rekeying marks every seam the systems do not join
What is bought in, and who manages that relationship?	Spend outside the delivery pool, and who controls it
How do you know, inside the month, whether you are on plan?	Whether anything is measured at a useful frequency

Watch one real transaction end to end before believing any documentation. Baseline cost by activity, not by department — departments hide duplication — and do it before anything moves, because once change starts the baseline cannot be rebuilt.

The read ends in three to five hypotheses, each with what would disprove it. Findings make a report. Hypotheses start a design.

The 24-Month Shape, and the Gates

Phase 0 earns the right to Phase 1; a delivered EBITDA flip earns the right to scale. Months 0–3, the diagnostic. Months 3–18, the operating engagement: the model installed, the core modernised, the flip delivered. Months 18–24 and beyond, exit or expansion. Three milestones gate the scale: validation, when the first production-ready module ships and holds under real load; benchmark, when the agentic delivery model runs across the full engineering lifecycle; expansion, when the same playbook is pointed at the next surface or the next asset.

The Transition: Changing the Engine in Mid-Air

The standard objection to any cost transition is the hollow-out: institutional knowledge walks out, service degrades, and the saving is consumed by the damage. The objection is legitimate — it describes most transitions. These mechanics exist to prevent it, and they are the most differentiated part of the model.

The transition, without the hollow-out

Hire ahead, then transition. Three waves, three-month overlaps, knowledge holders retained before anything moves.

M0M3M6M9M12M15M18M21M24

Phase 0 — audit, baseline, knowledge-holder map

Knowledge holders retained · written agreements, cash + equity

Maintenance team — v1 stable, renewals protected

Wave 1 · hire ahead

Wave 2 · 5–8 roles

Wave 3 · 5–8 roles

Rebuild team — modernisation under the gate

Client migrations

Hire Ahead, Then Hand Over

Migration runs in three waves of 5–8 roles, each with a three-month overlap: the successor is hired, onboarded and productive before the incumbent hands over. No customer experiences a service-level gap. The phasing deliberately trades some first-year saving for materially lower transition risk.

Retain the Knowledge Holders First

Phase 0 produces a knowledge-holder map: the specific 5–10 individuals who carry the legacy architecture in their heads. Before anything moves, they are retained under written agreement — cash or equity or both, vesting over 24 months, funded outside the savings budget. Where the engagement starts at acquisition, the packages fund from a pre-allocated portion of the purchase price; where the sponsor already owns the asset, from the Phase 1 budget or a sponsor equity top-up agreed in Phase 0; in the founder case, from the option pool or the round. Their departure mid-transition is how programmes of this kind actually fail, which makes this a week-4 action, not a contingency. The retention framework is a portable Phase 0 artefact: it arrives as a template and leaves as a signed set of agreements.

Two Teams: Maintain and Rebuild

From Day 60, once the map is complete, operations split. The maintenance team, built around the retained holders, keeps the existing product stable and protects renewals; that is its entire mandate. The rebuild team executes the modernisation under core-ring-edge as a greenfield-style unit. The resourcing shape is deliberately small: roughly six senior operators with the agent fleet on the rebuild, a compact maintenance layer with 3–5 retained holders on the legacy estate, replacing a conventional function several times the size.

Customer Continuity in the First 30 Days

Relationships at vendors of this profile are typically held by two to four senior individuals, personally rather than institutionally. Within the first 30 days, the senior decision-maker at each top-five customer is met directly, establishing relationships independent of any individual’s tenure. Renewal windows are mapped in Phase 0, and migration waves are sequenced so the highest-risk transitions never coincide with them.

Two Board Mandates: Freeze and Rationalise

1. Freeze legacy development. New feature work on the existing product stops. Without the freeze, maintenance load drags the rebuild team back into business-as-usual and starves the modernisation. That is how the previous attempts failed.
2. Rationalise the tail. The least profitable, most bespoke accounts are wound down under a disciplined framework: full-cost margin analysis in the first 90 days, strategic value assessed before any decision, a structured six-month transition for those confirmed. The objective is released capacity, not revenue reduction.

The Cash Shape of the Transition

These mechanics cost money before they release it. Phase 0 fees and retention packages land first; waves double-run for a quarter each; the tail winds down before its capacity is redeployed. EBITDA typically crosses zero around month 12 and builds to run-rate through months 24–36. The companion case study prints the half-year arithmetic; a board that has not seen the J-curve in advance treats month 6 as failure.

What Counts as Realised

Every number in a turnaround is contested at exit, so the discipline is set before the first benefit is claimed. It is the same three classes used throughout these documents — realised, operational, modelled — enforced at the point of measurement rather than at the point of writing.

A benefit is measured against a number taken before the change touched it, then measured again after bedding in. No baseline, no benefit: it is a forecast, and it is written as one. The failure mode is reversion — the programme leaves and the old way returns — which only repeat measurement catches.

Every benefit has one owner in the business, still in post after the engagement ends, who accepts it in writing. A benefit nobody will sign for is a forecast wearing a different label. That signature is also the handover: the successor named in casting is the person who inherits it.

Identified and realised are reported as two numbers, and the gap between them is the honest measure. Cash savings, efficiency gains and non-financial gains are never summed; a single blended figure is the first thing diligence takes apart.

The costs of the change are measured like the benefits — pruned accounts, authority taken from people who held it, the productivity lost while two teams run in parallel. A benefits case that shows only benefits is not a case.

Engagement Economics: Seats, Fees, Carry

The structure rests on a clean P&L separation, and on a distinction buyers reward at exit: what is permanent cost, and what is transformation cost.

Element	Shape
Operating seats	Fractional but titled, because titles carry the authority the mandate grants: a Chief Transformation Officer (org design, product operations, people) and a CTO for the new platform (architecture, rebuild team)
Incumbent leadership	Recast, not removed: the existing CTO owns the current platform, its stability and customer continuity — the part only they can run. Joint roles during the turnaround are the design, not a compromise
Reporting	Operating seats report to the CEO. A defined status-and-baseline reporting channel runs to the sponsor’s deal team: alignment, not authority. The CEO’s own reporting line to the board is untouched
Delivery team	Hired onto the company’s books: permanent payroll in R&D and cost of goods. Never an add-back, never a vendor line
Operating fee	A separate, clearly named professional-fees line, sized to fractional senior-partner time, structured as non-recurring transformation cost. A quality-of-earnings review adds it back: the fee is invisible to the exit multiple. A fee blended into run-rate OPEX is not
Carry	A pool tied to incremental EBITDA against the Phase 0 baseline, vesting over the engagement, paid at exit or refinancing
Founder variant	Fractional engagement under professional fees, with a success fee or equity participation at the raise or exit

Illustrative and subject to confirmation by counsel and the company’s CFO; tax and accounting treatment vary by jurisdiction.

Paid on the Increase

The engagement ties our wealth event to yours. A firm paid by the day carries no exposure to whether the discipline survives month 14; a firm that keeps the delivery team on its own books is selling an annuity that requires the core to remain absorbent. Ours requires the opposite: the fee is senior operating time, the delivery team is your payroll, and the upside is carry on EBITDA that did not exist.

The Baseline: Attribution and Exit Separability

What is rewarded inside a sponsor is attributable improvement. A gain nobody can trace to a decision funds nothing at the next investment committee; the Phase 0 baseline is what makes the improvement attributable. At exit it does a second job: once custom work lives outside the core, a buyer underwrites product and services each on its own terms rather than discounting the blend, and the perpetual maintenance obligation — usually unpriceable, therefore priced punitively — acquires a defined scope and a cap.

What We Don’t Offer, and What Breaks the Model

Scope, stated plainly, then the failure modes. Both belong in a document a sponsor will diligence.

Outside Our Scope

• We don’t sell AI transformation programmes. The phrase assumes the outcome is known before the work begins.
• We don’t bill for delivery hours. The fee is senior operating time.
• We don’t rebuild from scratch. We modernise the core. The data, the customers and the workflow stay.
• We don’t ship AI-written code without a human gate. Architecture decisions stay human.
• We don’t take engagements where speed trumps everything. That client cannot be served honestly.

Sales and Customer Success we treat differently, and precisely. We do not run the sales function, and we do not sub-contract it. We review it as an operation: is the process mapped and documented; is the funnel measured, with fixes feeding predictability and intelligence into the CRM; is the customer journey mapped, with A/B testing running on engagement, objection registers, show-up rates, close rates and budgeting tactics; is the offer articulated for every ICP, with marketing assets built for each buyer persona’s journey. Structurally, both functions are pulled into the product cadence: the forum that puts the right people in the room with the customer around an evidence-led roadmap, which is where delivery-driven churn actually gets fixed.

What Transfers, and What Is Rebuilt

A fair question in diligence is what, exactly, arrives with us. The method artefacts transfer directly: the product flow, the governance gate, the user-story format, the capacity model, the release calendar, the retention framework, the meeting discipline. The client-committee segmentation and squad shape transfer as method and are rebuilt against your customer base and product lines. Engineering frameworks and component libraries built inside prior engagements are the client’s IP: they do not transfer, and an equivalent is built per company. The agentic delivery patterns are our IP and transfer directly. Nothing arrives that a buyer would later have to license out.

Four registers are the durable artefacts, and they stay with the company. The decision-rights register: every boundary ruling with its owner and its reason. The baseline book: cost by activity, measured before anything moved. The benefits register: each benefit with its owner, its baseline and its measure. And the boundary register: what sits in the core, what sits in the ring, what was priced at the edge. A sponsor can ask to see all four at any point in the engagement, and at exit they are what makes the improvement attributable.

The Eight Failure Modes

The model is deterministic but not immune to physics. If the programme fails, it fails here:

1. Mandate not granted. The programme runs on influence and dies at the first renewal-pressure exception, before anything technical is attempted.
2. Boundary exceptions for material accounts. Custom code returns to the core. The architecture survives in name; the discount returns.
3. Customer concentration. A single account above 20% of revenue churning mid-programme takes the economics with it.
4. Knowledge-holder departure. If the 5–10 individuals who carry the legacy architecture are not retained under written agreement in week 4, the transition fails.
5. Ring efficiency disappoints. Configuration replaces custom code at lower throughput than assumed; gross margin lands high-70s rather than above 80. The thesis holds; the threshold claim does not.
6. Renewal repricing stalls. Conversion to clean subscription happens renewal by renewal over 18–24 months. A weak acceptance rate delays the re-rating.
7. Churn above 10%. This breaks the thesis outright. The model fixes margin, mix and predictability; it improves retention only where churn was delivery-driven. Where the product lacks market fit, every operating line can succeed and the asset still declines.
8. Mandate durability. Decision rights quietly stop being honoured after a sponsor reshuffle or board change, and everything below Step 1 reverts. The least discussed failure in this class of work, and the reason the mandate is papered by resolution rather than goodwill.

One scale statement belongs beside the failure modes. The four-function installation and the offshore transition are proven at product-team scale, inside a larger organisation, with the outcomes stated in the thesis document. Portfolio-company scale — multiple product lines, sales motion, finance, governance — is materially harder, and it is what the first mandate demonstrates. We would rather you hear that from us than find it in diligence.

01The ThesisWhy the discount exists, and how to detect itRead the report →02The Scaling LadderWhere it forms, stage by stageRead the report →

03 · this documentPortfolio OSThe operating model and the mandate

04The Modelled CaseThe arithmetic, entry to exitRequest the modelled case →

A Reinessence Group company.

Figures carry the classification scheme of the accompanying thesis report: realised, operational, modelled or externally verifiable, never summed. The 18–24 month window and the retention of the senior team are realised outcomes, and the 6x release cadence against the sector norm an operational one, all from one engagement, stated without naming the group, its sponsor or its customers; an evidence note with initiative-level bases is available on request, and fuller detail is disclosed under NDA. Compensation figures reference the public surveys at Exhibit 1. Engagement structure is illustrative and subject to counsel.