Our methodology
Every business is a profit machine - it turns inputs into profit. The question isn't whether the machine works. It's whether it works well, whether you can see it working, and whether it keeps working predictably.
The shape of it
A business only tells the whole truth when you read it through all three lenses at once. Each is run at full resolution - on the actual data, not category averages or a single interview. The operator names the question; data science answers it properly.
Is the profit real, repeatable and cash-backed?
The Profit Machine, below - unit economics, capital efficiency and margin consistency, read line by line.
Can the operation physically deliver the plan?
Capacity and the true constraint, computed from the operational log - the question a growth plan is really asking.
How much walks out if the owner does?
Founder and key-person dependency, measured from system exhaust - not the org chart or a self-report.
The first step
Every lens runs on your actual data, line by line - so the first job is always the same, and it's the one most projects skip: make the data readable. Messy, scattered data is the norm, not the exception, and it's exactly what we're built to handle.
Connect the systems - finance, ERP, billing, inventory, CRM - and pull the raw transaction data out. AI has made structuring messy, real-world data cheap: what used to need an analyst army now needs good architecture.
Join it into one clean transaction layer where every operational event carries its financial meaning - revenue, margin, cash-timing. This is where operations and finance stop being two separate records and become the same event, viewed two ways.
Tie the model back to the accounts, to the pound, and pressure-test it against the owner's own knowledge of the business - so the numbers are trustworthy before a single conclusion is drawn. A finding is only as honest as the baseline beneath it.
Only once this foundation holds do we read the three lenses. Everything else sits on top of it.
The financial lens
If a business is healthy across all three, the machine is working. If not, something is broken - and usually measurable.
Do you make money on every unit you sell? Revenue growth with declining unit economics is a leaky bucket - you're feeding the machine faster, but profit is falling out the bottom.
Does capital flow through the machine or get stuck? Strong earnings don't matter if working capital is invisible across five systems.
Can you predict your output from your input? When margin varies 30+ points within the same business, you don't have an industry ceiling - you have a consistency problem.
Layer 2
When a business underperforms, it's always one or more of these failure modes.
The machine is blind
Data exists across disconnected systems. Nobody has joined it up. Basic questions take a week to answer instead of an hour.
“A $400M company had $80M in working capital invisible because data lived in five systems. Leadership couldn't answer basic questions about cash conversion.”
The machine is misjudged
Assumptions that stopped being questioned. The machine is capable of more, but everyone believes the current output is the ceiling.
“Executives believed margins were capped by the industry. Disaggregation revealed 30-point variance - bottom at 15%, top at 45%.”
The machine is leaking
Value drains between functions. Order to ship to invoice to cash - each handoff is where revenue gets won but profit gets lost.
“Revenue leakage runs 1-5% of total revenue for mid-market companies. On $50M, that's $500K-$2.5M annually - often invisible across functional boundaries.”
The machine is protected
The people closest to the data know it's broken but won't say so. Problems are visible from the inside but invisible from the top.
“Visibility creates accountability. Accountability has owners. Owners have careers to protect. This is exactly why outside assessment matters.”
Layer 3
Diagnose before prescribe. Each stage has a clear purpose and deliverable.
See the machine
Connect data sources and make the machine visible. We map actual data flows, surface manual workarounds, and produce a first-pass view of all three cornerstones.
Size the gaps
Disaggregate every average, test inherited beliefs with data, and size each gap in pounds. Not “you have a pricing problem” - but “your bottom-quartile deals are 22 points below median margin, representing $1.8M in annual EBITDA.”
Fix the leaks
Build the specific measurement systems that close the diagnosed gaps. Automate manual workarounds, eliminate key-person dependencies, and create systems that trigger action - not just produce reports.
Keep it visible
Ongoing monitoring of all three cornerstones. Threshold alerts, leading indicators, and reporting that's management-grade and investor-grade at the same time. The machine stays visible permanently.
The operator lens
A deal - or a growth plan - is priced on volumes the business hasn't hit yet. Financial diligence verifies the past that produced the plan; almost nobody checks whether the operation can physically produce the future. We read capacity from the operational log, not a day's walk of the floor.
Nameplate (fiction), effective (the real ceiling), and demonstrated (the most it has ever actually done). The gap between demonstrated and what the plan needs is the risk being underwritten.
A business “at 70%” can have a single cell at 96% that caps the whole operation. Pour sales in above it and you buy queue and overtime, not output. Finding the real constraint is most of the job.
The last seasonal or promo peak already showed where the business bends under load. A plan that assumes 1.6x of a peak that broke at 1.0x is a re-price, not a growth story.
The leadership lens
Owner-dependent businesses look stable right up until the founder steps back - then the relationships walk out with them. You can't find it by asking the owner; they're the most conflicted witness in the room. So we don't ask. We reconstruct it from the fingerprints they left in the systems while running the business.
The revenue and margin on accounts the owner personally sources, prices or signs - read from quote authorship and account-owner fields, not from a claim.
The real org chart is the approval graph. It tells you whether the “#2 who handles pricing” approves anything, or just holds the title.
What happened to new quotes and collections the last three times the owner was away. You're not modelling the cliff - you're reading one that already happened at small scale.
A 30-minute call. Tell us what's keeping you up - tight cash, a growth plan you're not sure is deliverable, a business that leans too hard on one person - and we'll tell you honestly where to start, and whether we can help.