01 · Agreed before deployment
Baseline, metrics, evidence design and Stop / Modify / Scale thresholds are agreed with the client and written into the record first. Changes require documented joint approval.
Value Lab
Prove the value before scaling the solution.
Value Lab is BUSAC’s system for converting technology opportunities into attributable enterprise value: which interventions deserve capital, and which do not. It starts from an economic hypothesis, agrees the evidence plan and baseline before deployment, builds a bounded intervention in the real workflow, and lets the evidence decide whether capital should stop, modify or scale. After scale the loop keeps running: the intervention stays instrumented, is re-measured on the same baseline and is improved where the numbers say it can be.
01Identify
02Quantify
03Engineer
04Deploy
05Measure
06 · Decide
StopModifyScale
If scaleScale implementation
OngoingMonitor and improve
ENGINEER THE INTERVENTION. INSTRUMENT THE EVIDENCE. SCALE ONLY WHAT EARNS THE CAPITAL. THEN KEEP IMPROVING IT.
How It Works
Six stages, one record. The evidence decides.
The baseline and the evidence plan are agreed before anything is built, and scale is a decision the evidence has to earn. Below is the record itself, filled in for one order-to-cash hypothesis: every stage writes into it, and the decision at the end reads from it.
Value Evidence Record
VER-0001 · Order-to-cash · Discount leakage · PE-backed B2B software company
Evidence level V4 · Attributed
Decision: Scale, phased
Method walkthrough with illustrative figures. This is one internally consistent worked example of the record, not a client result.
Hypothesis
Discretionary discounting by sales reps is leaking margin. A guided discount authority (recommended discount plus approval above a threshold) will reduce discounts given without a corresponding change in win rate or churn.
- Value lever
- Pricing / discount leakage (order-to-cash)
- Mechanism
- Reps discount above policy to close faster; there is no recommendation at quote time and approval is after the fact
- Where it shows
- Realised price below list, net revenue per deal, credit notes
- Risks to the hypothesis
- Lower win rate, slower cycle time, churn at renewal
- Falsifier
- If leakage does not fall, or win rate falls more than 1 point, the hypothesis is rejected
Baseline and value range
- Baseline definition
- Discount leakage rate = discount given above policy as a share of gross bookings. Computed from CRM quotes and invoiced lines, 24 months, all regions, so that treated and comparison regions share one definition
- Scope in this record
- Region A, two product categories: €60m gross bookings annualised, out of €150m company-wide. Every value below is stated on this €60m; effects are given in thousands so the arithmetic can be followed
- Baseline value
- 4.1% of gross bookings in scope (illustrative: €2,460k on €60m)
- Locked
- Agreed with the CFO and frozen in this record before any build; the query, sources and exclusions are attached
- Value range (V2)
- €900k to €1,500k gross margin annualised on the €60m, if leakage falls 1.5 to 2.5 points with win rate held. Assumptions stated: policy-versus-actual analysis, no price elasticity beyond the current discount band
- Cost of Stage 2 (illustrative)
- €140k build, deploy and measure, plus €35k for Stage 1; run cost of the engine estimated at €60k a year
Evidence plan
- Design
- Phased rollout by region: Region A treated, Regions B and C untreated as comparison; difference-in-differences
- Primary metric
- Discount leakage rate, treated versus comparison
- Guardrails
- Win rate (tolerance minus 1 point), churn at renewal (no change, noting that eight weeks holds few renewals), quote-to-close cycle time
- Decision thresholds
- The threshold applies to the attributed point estimate; the 90% interval is reported beside it. Scale if the point estimate is at least 0.8 points and every guardrail is within tolerance; Modify if 0.3 to 0.8; Stop below 0.3 or if any guardrail breaches. If the interval lower bound is below the threshold, or a guardrail cannot be concluded inside the window, Scale is phased: one region at a time, with the open guardrail measured through the next renewal cycle before the next region
- Go / no-go
- Go. Value range clears the cost of Stage 2 several times over
Intervention
- Build
- Guided discount authority inside the quoting tool: a recommended discount per deal from a rules-plus-model engine, an approval step above threshold, and the reason shown to the rep
- Why a model
- Fixed rules cannot price the mix of segment, volume and renewal history. The engine recommends; the rep and the approver decide, and every override is logged with its reason.
- Instrumentation
- Quote-level event logging; measurement pipeline that recomputes the leakage rate daily on the locked baseline definition
- Scope
- Bounded: Region A, two product categories, 12 reps
- Controls
- Human approval above threshold; reps can override with a reason; overrides are logged and reviewed weekly
- Duration
- Six weeks from agreed plan to first live quote
Deployment
- Rollout
- Region A live in the existing quoting workflow; Regions B and C unchanged as comparison
- Test window
- Eight weeks, set by the evidence plan; long enough for two quote cycles
- Monitoring
- Engine uptime, recommendation latency, override rate, approval turnaround; alerts to the rep-ops lead
- Incidents
- One: recommendation service degraded for four hours in week 2; quotes fell back to manual. Kept in the headline measurement, because the outage is part of how the system performed; an ex-outage figure is reported alongside it
- Adoption
- Recommendation followed on 71% of quotes by week 4, 78% by week 8
Measurement
- Observed, treated
- Leakage 4.1% to 2.8% (Region A, eight weeks, outage included)
- Observed, comparison
- Leakage 4.0% to 3.9% (Regions B and C)
- Attributed effect
- 1.2 points, difference-in-differences; 90% interval 0.7 to 1.8 points. Excluding the four-hour outage: 1.3 points
- Guardrails
- Win rate minus 0.4 points and cycle time minus 6%: both within tolerance. Churn at renewal: unchanged on the 19 renewals that fell in the window, too few to conclude either way; recorded as open, not as intact
- Side effects
- Approval queue peaked in week 3; resolved by raising the threshold for repeat customers. Recorded, immaterial to value
- Model quality
- Recommendation acceptance, override reasons and the error between recommended and accepted discount, tracked weekly against the week-1 model.
- Reconciliation
- Invoice-level, tied to finance ledger for Region A; confounders stated: a price-list update in week 6 applied to all regions equally
- Attributed value, gross
- €720k gross margin annualised (1.2 points on €60m); range €420k to €1,080k on the interval
- Bridge to EBITDA
- €720k, less €60k a year to run the engine, less €100k margin foregone on the 0.4-point win-rate dip: €560k net EBITDA annualised; range €260k to €920k. The dip is inside tolerance and is still counted. Annualised means the eight-week run-rate extrapolated to a year, not a realised annual profit
Decision
- Rule, pre-agreed
- Scale if the attributed point estimate is at least 0.8 points and every guardrail is within tolerance; phased if the interval lower bound is below the threshold or a guardrail is open
- Result
- 1.2 points on the point estimate: clears the threshold. Interval lower bound 0.7: below it. Win rate and cycle time within tolerance; churn open. Two conditions for phasing are met
- Decision
- Scale, phased. Taken by the client on the record: Region B first, with churn at renewal measured through the next renewal cycle before Region C is decided
- Scale plan
- Region B, then C, all product categories: €90m further bookings. Modelled at V2 on the measured interval rather than the original range: 0.7 to 1.8 points on €90m is €630k to €1,620k gross, less €150k margin foregone at the same win-rate dip and €40k incremental run cost, so €440k to €1,430k net annualised. Measured before it is counted, and subject to the churn result from Region B
- Stage cost and payback
- €175k one-off for Stages 1 and 2, against €560k net annualised: simple payback 3.75 months on the point estimate, 8.1 months on the lower bound. Both extrapolate an eight-week window and exclude the fourteen weeks of build and measurement before it
- What stays
- The engine, the instrumentation, the baseline definition and this record, owned by the company
- What keeps running
- Monthly re-measurement on the same baseline, drift checks on the recommendation quality, and a retraining trigger agreed with the client. The record stays open: every re-measurement writes into it.
- Reviewable by
- Client finance, an independent reviewer, or a buyer’s diligence team, from the attached sources
Evidence Design
The strongest feasible counterfactual, chosen per intervention.
The method depends on feasibility, risk, data and the materiality of the decision. The standard does not: every claim carries its evidence level.
- Randomised or holdout designwhere feasible in the workflow
- Phased rolloutstaggered by region, team or segment
- Matched control or difference-in-differenceswhen a natural comparison group exists
- Interrupted time serieswhen only the treated population is available
- Finance-reconciled before/afterwith explicit confounders, when nothing stronger is possible
The V1–V4 Taxonomy
The technology is not the evidence. The change in outcome is.
A modelled €4m opportunity is not €4m of EBITDA. Every claim carries its evidence level: hypothesised, modelled, observed, attributed.
BaselineWhere are we today?
→
InterventionWhat did we change?
→
PerformanceDid the system perform better?
→
Operational impactWhat changed in the business?
→
AttributionWhat would have happened anyway?
→
Financial impactWhat value did that create?
V1Hypothesised
V2Modelled
V3Observed
V4Attributed
Hover a level to see which links of the chain it covers.
€3m of V2 ≠ €3m of V4. The unit is the same; the evidence is not.
The capital rule
Capital may scale once value is observed and the attribution is strong enough for the decision at hand. V4 is the standard for attributed value; not every engagement reaches it, and the record says so when it does not. Attributed means attributed by BUSAC, not independently audited: an independent review can be added when the decision warrants it.
The Referee Rule
We build the intervention. The evidence must stand on its own.
Before deployment, BUSAC and the client lock the baseline, metrics, evidence design and Stop / Modify / Scale thresholds. Calculation logic, source lineage, assumptions and approved changes remain reviewable. Our fees are not tied to the impact we measure, and every Value Evidence Record is structured for independent review. Where the decision warrants it, an independent reviewer can validate the evidence.
02 · Reviewable by design
Calculation logic, source lineage and assumptions remain available in the client’s environment, so a finance team or a buyer’s diligence team can retrace the figures.
03 · Fees not tied to measured impact
Our fees are not tied to the impact we measure. The next stage is a client decision, taken on the record.
04 · Structured for independent review
Every Value Evidence Record is structured so an independent reviewer can validate the evidence where the decision warrants it. The same standard applies to interventions BUSAC did not build: an Evidence Review is a standalone engagement.
Our Focus · Where We Start
Four levers where technology moves enterprise value.
Where we start: PE-backed software, tech-enabled services, financial services and insurance distribution. The discipline travels.
AI-Enabled Operating Leverage
Automation across the workflows that consume human capacity: intake, processing, servicing, sales support. Measured in cost-to-serve, throughput per FTE and conversion.
Platform & Integration
Fragmented systems and data, often the residue of acquisitions, turned into one operating platform that automation can run on. Measured in cycle time and capacity released.
Technology Cost Efficiency
Cloud, tooling, vendors and technical debt. Measured as attributed run-rate change against a locked cost baseline. Vendor-neutral: BUSAC sells no technology, resells nothing and receives no commission from any provider.
Profitable Growth
Pricing, retention, cross-sell and product enablement where operating data can support a testable hypothesis. Measured in margin, retention and conversion, with the same evidence gates.
Built for private equity
For investors and portfolio leaders making technology deliver.
We help mid-market private equity firms and their portfolio companies decide where technology deserves capital, improve what is already deployed, and measure the value it creates.
From diligence and the first 100 days through ownership and exit.
How Engagements Work
Staged commitment. Evidence unlocks the next stage.
Small capital buys information; information de-risks larger capital. You bring one place where you believe value is hiding. Each stage ends in a decision you can take on the record. When there is a deal, it starts earlier: Technology Value Diligence seeds the Opportunity Map from the memo.
Opportunity Map
The engagement is the Opportunity Stage; the Opportunity Map is what you receive from it. How the relevant processes run today, where technology changes the economics, and the candidate interventions ranked on value potential, feasibility, risk and evidence quality. For the selected opportunity: locked baseline, modelled value range (V2), intervention design and evidence plan. You learn whether the value justifies building, before anything is built.
Build & Measure
The intervention is engineered into a bounded part of the business and instrumented against the locked baseline. You receive working software, a Value Evidence Record and a Stop / Modify / Scale recommendation.
Scale implementation
Only after a Scale decision: full population, integrated, monitored, governed. The instrumentation stays with the company.
Portfolio Value Lab
A standing technology-capital capability across the portfolio. Each intervention produces a Value Evidence Record; together, those records create a live view of where capital is committed, what evidence it has earned and what should Stop, Modify or Scale. Client data remains client-controlled. Cross-company learning uses only approved, abstracted patterns.
BUSAC is at the start of its own track record and launches Value Lab with a small number of founding partners. Their records are the first in the library, and every pattern proven with them starts ahead for them.
Start a Value Lab
Which opportunity in your portfolio deserves the next euro of capital?
Bring one hypothesis, or one workflow that may be constraining value. One Opportunity Stage tells you whether the value is material, what it would take to prove, and whether it deserves the next euro of capital.