Healthcare data & analytics · for leadership teams, sponsors & the people who run the floor
One set of numbers for healthcare companies that grew by acquisition.
You bought five companies. Each one counts revenue differently. We build the unified reporting layer inside your own tenant, in weeks, not quarters, and prove every number reconciles to the report your leadership already uses.
The problem
We’ve bought five companies. Each one reports revenue differently. Every month my controller rebuilds the same spreadsheet, and when the board asks why a number moved, it takes us four days to answer.
Heard inside every healthcare roll-up, eventually
No single source
Each entity kept its own system and its own definitions. The same metric means something different in every one.
No trust
Numbers exist, but leadership keeps its own spreadsheets because the two never agree.
No speed
Answering a board question takes a person, a week, and a rebuilt spreadsheet.
Every acquisition adds a box on the left. The right side never changes.
How this is different
Three commitments. Each with receipts.
Numbers you can defend
Every build ships with a validation harness: the warehouse ties line-by-line to the reports leadership already relies on, and re-checks itself on every load. When a source system changes upstream, you find out before the board does.
Weeks, not quarters
Fixed-fee engagements with named deliverables and dates. No discovery phase that bills for three months. The first useful artifact lands in week two.
You own everything
Built in your tenant, in your repo, documented. No proprietary platform, no per-seat data tax. If you hire an internal team next year, they inherit a working system. That’s the plan, not a failure.
Services
Five rungs. One ladder.
Every step is fixed-fee, scoped in writing, and useful on its own. Most clients start with the assessment.
Data & KPI Assessment
2 weeksAn inventory of every source system, feed, and spreadsheet in your reporting path. A KPI definition audit: where the same metric means different things across entities. A data-quality profile, and a prioritized 90-day roadmap with effort and cost per item, delivered in a 60-minute readout to your leadership team. Genuinely useful whether or not you ever hire us for the build.
Rung 0 · The sponsor door
Diligence & 100-Day Data Sprint
2–3 weeksFor sponsors. Pre-close: the clean operational dataset the QoE team can’t get out of the target’s systems. Post-close: the 100-day reporting baseline: what the new owner measures from month one.
Rung 2 · The foundation
Data Platform Build
8–14 weeksIngestion from every source system. A conformed warehouse with one definition of provider, client, encounter, and revenue. Orchestrated daily loads with health monitoring, a validation harness reconciling everything to the reports you already trust. Documented and handed over.
Seen in production: five brands consolidated; monthly close assembly from days to zero.
Rung 3 · The visible layer
Executive & Operational Reporting
3–6 weeksThe dashboard the board actually reads (revenue, gross profit, pace-to-plan) plus per-function operational views and an automated weekly leadership digest. KPI definitions agreed in writing before anything is built.
Seen in production: board questions answered same-day, down from four days.
Rung 4 · The standing relationship
Fractional Analytics Leadership
monthlyAn analytics function without the $180K hire. Models and dashboards evolve as the business changes, new metrics turn around in days, monthly close and board-package support, standing time with your leadership team.
Natural-Language Insights Layer add-on
“Which brand drove the margin dip in March?” Asked in plain English, answered with the number, the chart, and the drill-down. It sits on the governed mart, so every answer inherits the same metric definitions as the dashboards and traces back to source. The difference between an answer you can take to the board and a guess with good grammar.
Targeted fixes
scoped individuallyNot everything needs the full ladder. We also take on surgical work: repairing a warehouse that exists but isn’t trusted, untangling one stubborn data-quality problem, standing up a single pipeline from a new source system, or fixing the Power BI and Tableau dashboards you already have rather than replacing them.
How pricing works
- Every engagement is a fixed fee, quoted before work begins. No hourly meter, no open-ended discovery.
- No change orders for doing what the scope says. The scope is written down; so are the deliverables and dates.
- Retainers name their hours in the SOW, so you always know what you’re buying.
How it works
A 20-minute call. Two weeks. A ranked plan.
The call 20 minutes
You describe how reporting works today. We tell you whether this is a problem we fix, and if it isn’t, who to talk to instead.
The assessment week 1–2
Week 1: systems access, source inventory, and the KPI definition audit: where the same metric diverges across entities. Week 2: data-quality profile, the ranked 90-day roadmap with effort and cost per item, and a 60-minute readout with your leadership team.
Your call after the readout
Execute the roadmap with your own team, with us, or not at all. The assessment stands on its own. No engagement continues by default.
One more thing no leverage-model firm can say: 100% of engagement time goes to your project. The person who scoped the work is the person doing it.
Results
Anonymized, but real.
The revenue nobody was counting
The situation. At a PE-backed workforce platform, permanent-placement fees lived entirely outside the reporting path: booked, banked, and invisible in every board metric.
The work. Modeling that fee stream into the warehouse and wiring it into executive reporting, a byproduct of building one governed definition of revenue, not a separate project.
The outcome. $559K of previously untracked revenue surfaced in a single line of business. The question it leaves behind is the one worth asking: what else isn’t being counted?
An acquisitive workforce platform ran five brands on four source systems across three SQL dialects. After the build: one revenue and gross-profit number across all brands, refreshed daily. Consolidation went from days to zero.
Leadership kept its own spreadsheets because the dashboards never agreed. A line-by-line validation harness fixed that, then caught a silent upstream extract change that swung two brands by +68% and −91%, before it reached a board report.
Who we work with
From the boardroom to the floor.
One system of numbers only works if everyone reads from it, so the work serves the whole company, not just finance.
The company
- PE-backed healthcare services: staffing & workforce, provider groups, behavioral health
- $20–250M revenue
- 2–8 acquired entities that were never integrated
- No internal data team, or one overwhelmed analyst
The executives
- Whoever inherited five versions of revenue and owns the board deck
- Leaders who want pace-to-plan without waiting a week
- Operating partners who want one reporting standard across the portfolio
The operational teams
- Per-function dashboards: sales pipeline, delivery, recruiting productivity, utilization
- Metrics with one agreed definition, so ops and finance stop arguing about whose number is right
- An automated weekly digest instead of a spreadsheet request queue
How we build
Familiar tools. The hard part is what the numbers mean.
Thousands of firms can stand up this stack. The scarce part is knowing what a working provider, a net booked week, or time-on-assignment actually means, and why every acquired entity computes them differently. That said, your IT team will recognize everything we bring:
- Conformed dimensional models: star schemas your analysts can actually query, not a lake of parquet nobody owns
- Orchestrated ELT with logging, health monitoring, and staleness alerts on every load
- The validation harness runs as automated tests, tied line-by-line to the reports leadership already uses
- Everything version-controlled in your repo, with documentation written as we go, not reconstructed at handoff
FAQ
The questions we always get.
Why not just hire an analyst?
A good healthcare analytics hire is $140–180K loaded, takes four months to find and six to ramp, and they’ll build on whatever foundation exists, which is the problem. We build the foundation in weeks for less than a year of salary, and you can hire onto it afterward. Many clients do; that’s a success.
Why not a big firm?
Big firms sell a partner and staff a first-year associate. You get the same person who scoped the work: someone who has personally built this stack in healthcare, not adapted a retail template. You’ll also pay 3–5x for the logo.
We already have Power BI.
Almost everyone does. The dashboard layer isn’t the problem. What’s underneath it is. If five systems disagree, five dashboards disagree faster. The assessment tells you in two weeks whether you have a reporting problem or a data problem. It’s usually the second one.
What’s your AI story?
A sober one. AI here means the natural-language layer on top of your governed warehouse: every answer resolves to a metric with an agreed, written definition, the same one the dashboards use, and traces back to source. If it can’t be traced to the warehouse, it isn’t answered. Your data stays in your tenant, is never used to train anyone’s model, and where PHI is in scope the same BAA and HIPAA-eligible services apply to the AI layer as to everything else.
Is our data safe? Is this HIPAA-compliant?
Most of this work is operational and financial data, not PHI. Where PHI is in scope, we sign a BAA and build only on HIPAA-eligible services, inside your tenant, under your access controls. Your data never leaves your environment.
What happens when you’re gone?
You own the tenant, the repo, the models, and the documentation from day one. There’s no proprietary layer to lose access to. That’s a deliberate design choice, and it’s in the contract.
Worth 20 minutes?
You describe how reporting works today. You leave knowing three things: whether you have a reporting problem or a data problem, the two fixes we’d do first, and whether we’re the right people to do them. If we can’t help, we’ll say so on the call.
Kaleb Lewis · Principal. The person who answers is the person who does the work.