A healthcare SaaS ROI model is a defended argument, not a spreadsheet. It compares everything a customer gains, in dollars, against everything the customer spends, over a fixed period, from a baseline the customer's own data confirms. The models that win contracts and funding show benefits a finance team can reconcile, costs that include the customer's internal effort, and a downside case that still holds. In our work as a healthcare advisory firm, the clinical credibility of the inputs decides whether the model survives the first review, and that is where a Fractional Chief Medical Officer earns their place.
This guide is for founders, healthtech executives and the investors backing them. It covers the formula, the inputs, a worked example, and how to test the model before a buyer or a diligence team does.
TL;DR
- ROI is total benefits minus total costs, divided by total costs, over a stated period. Above zero (or above 1.0x as a benefit-cost ratio) means the customer gets more than it spends.
- Anchor every benefit to a baseline in the customer's source systems. Unanchored benefits are the first thing a CFO discounts.
- Count the customer's full cost: license, integration, training, data preparation and internal staff time.
- Show conservative, base and optimistic cases, plus a stress test at half the projected benefit. Lead with payback when the audience is skeptical.
- Pair the financial case with 2 or 3 clinical outcomes, because payment is moving toward outcomes and a purely financial ROI loses force.
Table of Contents
- What ROI means for a healthcare SaaS product
- Why a purely financial ROI is not enough
- What goes into the model
- The formulas, in plain terms
- A worked example
- Assign confidence to each assumption
- Test the model, then measure the real result
- Where founders and investors go wrong
- Our perspective
- Getting the model reviewed
- FAQ
- Sources
What ROI means for a healthcare SaaS product
Return on investment compares what a buyer gets with what the buyer puts in. The federal Agency for Healthcare Research and Quality (AHRQ) defines ROI as net returns from improvement actions divided by the investment in those actions. Net returns come from changes in quality, efficiency and utilization of services, or in the payments for them. Investment covers all costs of developing and operating the change.
AHRQ also gives a plain way to read the number. A ratio above 1 is a gain: 1.8 means $1.80 back for every $1 spent. A ratio between 0 and 1 is still a loss, because costs are not recovered. A ratio below 0 is a net loss. A projected ROI, built before implementation, is used to estimate when the customer breaks even.

Who is reading your model
Three readers see it differently:
- The buyer's CFO wants dollars that reconcile to the general ledger, claims data or payroll.
- The buyer's clinical leader wants to know the workflow will not add burden and that outcomes will not suffer.
- Your investor or acquirer wants to see that your revenue rests on customer value that persists after the first renewal.
A model that serves only the first reader tends to stall with the second. Clinical leaders can end a pilot quietly, whatever the spreadsheet says.
ROI is not ROAS
Founders in digital health sometimes borrow marketing metrics. Return on ad spend divides attributed revenue by advertising spend. It says nothing about profit or about whether the buyer's investment pays back. A customer evaluating your software is asking an ROI question, not a ROAS question. Keep the two apart in investor decks, where they are often blended.
Why a purely financial ROI is not enough
The authors of a peer-reviewed Value of Investment framework for healthcare IT argue that as US payment shifts from fee-for-service toward outcomes, a strictly financial ROI loses usefulness. They recommend pairing ROI with clinical outcomes and nonfinancial value, and using net present value (NPV, the value today of future cash flows) and total economic impact alongside it.
This matters for how you build the model. In a fee-for-service setting, a tool that reduces readmissions may reduce revenue to the hospital. In a value-based contract, the same tool is worth money. A 2025 peer-reviewed article on selling the return on investment for digital health says a credible case must identify the institutional problem being solved, the payment model (value-based, fee-for-service or both), and how the product fits existing workflow, including EHR integration.
So before you write a single formula, answer three questions for each target customer:
- Which problem does the buyer feel in dollars today?
- Under which payment model does the benefit accrue to them?
- Which workflow change does your product require, and who absorbs it?
If you cannot answer the third, the model is incomplete. For the integration piece, our posts on EHR integration explained for investors and what healthtech teams should budget for EHR integration show how it shapes cost and timing.
What goes into the model
Benefits
Benefits fall into four groups. Pick the few your product moves most directly rather than listing everything.
- Financial: recovered or new revenue, prevented denials, lower vendor spend, cost avoidance.
- Operational: staff hours saved, turnaround times, throughput, documentation time.
- Clinical: readmissions, avoidable emergency visits, care-gap closure, control of chronic measures.
- Experience and retention: patient or clinician satisfaction that affects retention.
Bessemer Venture Partners notes that the most direct way to track financial ROI for healthcare SaaS is the incremental revenue customers expect from the product. Cost reductions also count, but they take longer to realize and attribution is harder. Flare Capital Partners likewise separates direct benefits from quantified indirect ones, such as engagement that lifts retention, and suggests a tech-enabled service should ideally return 3:1 to 5:1.
For the clinical group, resist the urge to track a dozen measures. We have written about choosing 2 or 3 clinical outcomes metrics that health systems and SaaS companies can both stand behind. The same discipline applies here: a few outcomes tied to the problem you solve beat a long dashboard.
Costs
The most common error is counting only your price. The customer's total cost includes:
- Licensing or subscription fees
- Implementation and integration work
- Training and change management
- Data preparation and security review
- Ongoing support
- Internal staff time, including clinicians pulled from patient care
Leaving internal time out makes the model look better and the buyer's finance team suspicious. If you price pilots, our guide to pilot pricing for healthcare SaaS shows how to keep the pilot cost visible in the same model.
Baseline and time horizon
Every benefit is a change from something. Record the baseline before go-live from the customer's source systems: claims, payroll, the scheduling system, the EHR. State it with a denominator (denials per 100 claims, minutes per note) so the change is measurable in both absolute and percentage terms.
Pick a horizon and keep it. Three years is common for software. Nucleus Research averages annual net benefits over three years against total initial cost and cautions that ROI is often miscalculated as total benefits over total costs, which is really a benefit-cost ratio. Decide which one you are showing, and label it.
The formulas, in plain terms
| Measure | Formula | What it tells the buyer |
|---|---|---|
| Net benefit | Benefits minus costs for the period | The dollars gained |
| ROI | Net benefit ÷ total cost, as a percentage | Return per dollar invested |
| Benefit-cost ratio | Total benefit ÷ total cost | Value relative to spend |
| Payback period | Months until cumulative net benefit covers cumulative cost | How soon the buyer's money comes back |
| NPV | Future net cash flows discounted to today | Whether the project beats the cost of capital |
Nucleus Research argues payback should lead a technology business case, because it is easier to grasp and signals risk: the longer the payback, the more likely actual returns drift from the estimate. Its example is that recovering the investment within 24 months lands better with a skeptical finance team than a bare percentage. We agree for early-stage products, where a buyer is weighing an unfamiliar vendor.
A worked example
The numbers below are hypothetical, built to show the mechanics. They are not client results.
A revenue cycle tool is sold to a health system. Assumptions:
- 60,000 claims a year, with a 10% denial rate today (6,000 denials).
- The tool is projected to cut the denial rate to 8%, or 1,200 fewer denials.
- Each avoided denial is worth $450 in recovered net revenue: $540,000.
- Staff time saved: 6,000 hours at $40 an hour: $240,000.
- Year 1 cost: $300,000 subscription plus $120,000 for implementation and internal staff time: $420,000.
Base case. Benefits are $780,000. Net benefit is $360,000. ROI is 360,000 ÷ 420,000, about 86%. The benefit-cost ratio is 1.86. If benefits arrive evenly, payback is roughly 6.5 months.
Conservative case. The finance team credits 75% of the denial benefit and 50% of the time savings, since saved hours rarely become cash unless headcount or overtime changes. Benefits fall to $525,000. ROI is 25%.
Stress test. Cut every benefit in half: $390,000 against $420,000 of cost. ROI is about negative 7%, and payback stretches to nearly 13 months.

The stress test is the useful result. It tells you that the case depends on achieving more than half the projected benefit in year one, and that year two, with implementation costs behind it, looks far better. A buyer who sees this will trust the base case more, not less.
Assign confidence to each assumption
Not every input deserves equal weight. A practical convention is to rate each benefit by how it was established:
- Reconciled cash from a prior pilot or the customer's ledger: full credit.
- Observed operational performance in similar deployments: partial credit.
- Modeled capacity, such as hours saved that may or may not be redeployed: low credit.
Whatever weights you choose, show them. One published framework uses 100%, 75% and 25% for these three tiers. Your own weights should come from your data, and you should be ready to explain them.
Three assumptions deserve extra scrutiny:
- Adoption. Benefits scale with the share of clinicians who actually use the product. Model a ramp, not day-one full use.
- Reimbursement. If the benefit depends on a payment rule or contract, name it and test what happens if it changes.
- Retention. Revisit renewal assumptions with cohort data past the first renewal, not with first-year enthusiasm.
Test the model, then measure the real result
Before sending the model to a buyer or investor:
- Build conservative, base and optimistic cases.
- Cut all benefits by half and see when the customer breaks even.
- Change one assumption at a time to find the one that moves the result most. That is your diligence risk.
- Have a clinician read the clinical inputs. Implausible effect sizes damage credibility fast.
After go-live, measure. The Value of Investment authors suggest assessing realized value once a trained, engaged user community has had enough time, typically three to six months after go-live, and then comparing it with the estimate: were the assumptions sound, and why did actual value differ? Feed the answer into the next model. Companies that do this build a library of real customer results, which is the strongest ROI evidence there is, and it matters in investor reviews. See our notes on healthcare SaaS due diligence for how those reviews examine customer value claims.
Where founders and investors go wrong
- Counting vendor price as total cost. Buyers count their own time, and so should the model.
- Mixing revenue and profit. A recovered claim is not all margin. State the margin you assume.
- Double counting. Staff hours saved and cost per case reduced can be the same dollars.
- Attributing seasonal change to the product. Separate the software's effect from seasonal or policy shifts by comparing against a control period or unit.
- Ignoring clinical workflow. A product that saves administrators time but adds clicks for physicians may never reach the usage the model assumes.
- Presenting ROI as certainty. A projection is a range with stated assumptions.
Our perspective
I have spent more than 25 years in medicine and business, and the ROI models that fail rarely fail on arithmetic. They fail on the inputs. A founder assumes a physician will change a workflow that, in real care, they will not. Or the model credits hours saved that a busy unit never converts into cash.
That is the gap a Fractional Chief Medical Officer is built to close. When a clinician who has practiced and also built businesses reads your model, the clinical assumptions get tested against how care is actually delivered: which metric a health system's quality team will accept, how much adoption is plausible, and where EHR integration will slow things down.
For investors, the same lens applies to portfolio companies. Before a diligence team challenges a customer value claim, it is better to challenge it ourselves, with a baseline, a downside case and two or three outcomes that matter to clinical leaders.
If a model feels more confident than the evidence behind it, I would welcome a conversation.
Paul Bergeron MD, MBA
Getting the model reviewed
The StartUp MD provides Fractional Chief Medical Officer leadership and advisory services for healthcare SaaS companies, including clinical strategy, product evaluation, market readiness assessment and investor support. Our founder, Paul Bergeron, MD, MBA, brings both clinical and business experience to each engagement. If you are preparing an ROI model for a first health system contract, a fundraise or a portfolio review, you can see how we work on our services page.

Requests range from a single advisory project to an ongoing fractional role, depending on how much medical leadership your team needs.
FAQ
What is a good ROI for a healthcare SaaS product?
Any ROI above zero means the customer recovers more than it spends, and a ratio above 1 on AHRQ's scale means more than a dollar back per dollar invested. Investors often look for more headroom: Flare Capital Partners suggests a tech-enabled service should ideally return 3:1 to 5:1. What counts as good also depends on how quickly the customer is paid back and how well the benefits are documented.
How is ROI different from payback period?
ROI describes the return over a chosen period, such as three years. Payback describes how many months pass before cumulative net benefit covers cumulative cost. Payback is easier for skeptical buyers to judge and reflects risk, so many teams show both, with payback first for early-stage products.
How long should the ROI horizon be?
Three years is a common choice for software, because implementation costs are front-loaded and benefits build as adoption grows. State the horizon in every figure you publish. Then compare realized results with your estimate about three to six months after go-live, and update the model.
Should clinical outcomes be in an ROI model?
Yes, when your product touches care. Researchers writing on the value of investment in healthcare IT argue that as payment moves toward outcomes, clinical and nonfinancial value must sit beside the financial return. Choose two or three outcomes tied to the problem you solve, with a baseline for each.
Who should review the model before it goes to buyers or investors?
A finance reader checks the arithmetic, and a clinical leader checks whether the assumptions match real practice. Many healthcare SaaS teams lack the second reader in-house, which is where a Fractional Chief Medical Officer fits.
Sources
- AHRQ: Return on Investment Estimation instructions
- Value of Investment as a Key Driver for Prioritization (PMC)
- Selling the Return on Investment for Digital Health (PMC)
- Bessemer Venture Partners: The ABCs of health tech
- Flare Capital Partners: Surviving Digital Health, Part 3
- Nucleus Research: Everything to know about ROI, TCO, NPV, and Payback
- How Do Healthcare SaaS Leaders Calculate a Defensible ROI
