Building the business case for Medical Affairs technology

A business case for Medical Affairs technology fails when it borrows the shape of a commercial one. Medical Affairs does not own revenue, cannot attribute it, and should not claim to — which means the case has to rest on what the function can actually defend, rather than on a number that will not survive the second meeting.

Most cases for this kind of technology get written in the familiar shape: cost in, revenue out, payback in some number of months. It is the shape a finance function recognises, which is exactly why people reach for it, and it is the shape that gets the proposal killed in the room that matters.

This piece is about constructing the case. Which vendor to choose is a separate question, covered in the questions worth putting to any vendor. Both sit inside the measurement problem behind all of this, which is what makes the business case hard in the first place.

You will notice there are no numbers anywhere in what follows. That is deliberate, and the reason is in the fourth section.

Why the commercial ROI argument fails here

Three reasons, and they get more serious as they go.

Attribution is not available. A conversation that changes how a clinician reads the evidence may show up in practice months later, mediated by a guideline, a colleague, a formulary decision, a conference. There is no clean causal line, and constructing one requires a chain of assumptions that a finance reviewer will take apart in the first meeting. You will be defending the method instead of discussing the decision.

Claiming it is worse than failing to claim it. If Medical Affairs asserts that it drove prescribing, it has just described itself as a commercial function. That sentence can be quoted back by a compliance officer, an auditor, or a regulator, and it damages the non-promotional standing the whole function rests on. This is the reason not to make the argument even in the cases where the data appears to support it.

A borrowed benchmark will not hold. Published ROI figures for this category are almost always drawn from a different function, a different size of organisation, or a vendor's own customer base. Quoting one stakes your credibility on somebody else's data, and the first person to ask where it came from wins the exchange.

None of this means the case cannot be made. It means the case cannot be a weaker version of the commercial argument. It has to be a different argument, and the good news is that the different argument is stronger — it is just less familiar.

It is worth noticing that the alternative most functions fall back on is worse still: volume. Meetings held, contacts reached, materials delivered. That is counting activity instead of substance, and a finance reviewer will spot immediately that it measures effort rather than effect.

What finance will actually ask

Five questions. Preparing these matters more than any modelling.

1. What does this replace, and what happens to that cost? If it replaces nothing, say so plainly. A case that presents itself as a consolidation when it is actually an addition gets discovered, and everything else in the document loses credibility with it.
2. What is the total cost over three years? Not the licence. Implementation, integration, training, internal time, and the cost of running the old process in parallel during transition — which is almost always underestimated and almost always noticed.
3. What happens if we do not do this? The most important question and the one people are least prepared for. If the honest answer is "not very much", the project is discretionary, and discretionary projects are cut in the next budget round whether or not they were approved in this one.
4. How will we know in twelve months whether this worked? If you cannot answer this before purchase, you will not be able to answer it afterwards, and the renewal conversation will be an argument rather than a review.
5. What is the exit? What happens to the data, in what format, at what notice, at what cost. This is standard due diligence and it belongs in the case rather than in the contract review, because the answer affects how much the commitment is really worth.

Notice that none of the five asks for a return figure. In practice finance functions are considerably more comfortable with "we cannot attribute this, and here is what we can defend instead" than people expect. It is the unconvincing number that attracts scrutiny, not the absence of one.

Three defensible arguments, and one that is not

Risk and defensibility. A structured, auditable account of what was discussed, with whom, under what consent, with safety signals identified and routed rather than sitting unread in a free-text field. This is the same argument that funds every other compliance system in the organisation, so the finance function already understands its shape. It is also the argument least dependent on the technology performing perfectly, which makes it the sturdiest thing in the case. What any vendor actually holds here is checkable — ours is set out at what we hold today, and what we do not — and a vendor who is vague about it is telling you something.

The asset that persists. Most of what a field team learns about clinician understanding evaporates at the next territory change. An organisation that retains it is buying something that accumulates rather than something it rents, and the distinction between an expense and an asset is one finance grasps faster than anyone else in the building. This is the only asset in the function that compounds, and it is the argument that improves the longer the system runs.

Decision quality upstream. Insights that actually reach evidence generation, publication planning and medical content change what those functions produce. That is a real effect on genuinely expensive work, and unlike prescribing it is traceable — you can point at a specific decision and the specific input that informed it. Most organisations have never checked whether insights reach anyone who can act, which is why this argument is usually available and usually unused.

And the one that is not: time saved.

Hours recovered from administration is a real effect. It is also the weakest argument on the table, and it is the one almost everybody leads with.

Three problems with it. It is small relative to the cost, so it invites the comparison you least want. It leads directly to a question you do not want to answer, which is whether you are therefore reducing headcount. And it anchors the entire discussion on cost — and a Medical Affairs system will lose a cost argument to a commercial system every time, because the commercial system can put a number on the other side of the equation.

Use it as a supporting detail if you use it at all. Never as the headline.

What to measure during a pilot so the case can be made afterwards

The common mistake is running a pilot to find out whether people like the software, and then trying to build a business case out of satisfaction data. Satisfaction is not evidence of value and a finance reviewer knows it.

Decide the questions before the pilot starts, and capture the baseline before anything changes. The baseline is the step that gets skipped, and it is the one thing that cannot be recovered afterwards.

Four things worth capturing.

What the current process produces. Take a sample of existing records and have them assessed by the people downstream who would actually use them — for whether they are specific enough to act on. Do this before the pilot begins. It is uncomfortable and it is the single most persuasive item most cases contain.

Whether insights reach the function that can act on them. Take a set of insights from the last quarter and trace each one: who received it, who acted, what changed. Most organisations have never run this exercise, and the result usually makes the rest of the argument unnecessary.

Adoption under realistic conditions. Not the enthusiasts. A pilot staffed entirely by volunteers measures volunteers, and the number it produces will not survive contact with the wider team.

Time to a usable record — measured rather than estimated, and measured on both paths.

Set the decision rule before you begin. What result would lead you not to proceed? If there is no such result, this is not a pilot, it is an installation with a trial period attached, and it should be described that way in the case.

Which brings us back to why there are no numbers in this article. Every figure in your business case should come from this exercise, inside your own organisation. A number you generated is defensible in the room, because you can explain how it was produced. A number from a vendor's website is not, and somebody will ask.

What to do when the honest answer is "we do not know yet"

Some of this genuinely cannot be known in advance. Saying so is a stronger position than manufacturing a figure, and there are three ways to write a case that survives not knowing.

Name the uncertainty and put evidence around it. "We do not know how much this will improve insight quality. We do know what our current insights look like, and here is a sample of them." A specific unknown, bounded by real evidence, reads as more rigorous than a confident projection — because it is.

Stage the commitment. Buy the pilot, not the three-year term. Structure the decision so that the expensive part follows the evidence rather than preceding it. This is also a useful test of the vendor: one who will only sell the full commitment is telling you how confident they are.

Separate what you know from what you are projecting, visibly. Finance reviewers are used to documents where certainty and speculation are blended, and they discount the entire thing accordingly. Separating them explicitly buys credibility for the parts you are actually sure of.

A case built this way takes longer to approve. It is also much harder to kill in year two, because nothing in it was overstated in year one. The business cases that die are almost always the ones that promised a number at the start.

If you are building this case now, we would rather help you build a defensible one than send you a figure you cannot stand behind. Talk to us before procurement does →