The 10 basis point problem

Since 2017, the medtech industry has expanded its EBITDA margin by 10 basis points. Technology expanded its margin by 460 over the same period, according to McKinsey. Medtech finished eighth of the sectors it tracked.

Shareholders have noticed. McKinsey also found that about 80% of public medtechs delivered zero to negative returns over the past four years.

Now every board has an AI programme, and every AI programme has a business case. Chamath Palihapitiya put the problem with most of those cases bluntly on the All-In podcast this year: “There is literally not a scintilla of evidence that AI has helped lift the operating margins of the S&P 500.”

There are two ways AI can change that. Shrink operating expense and hold revenue. Or hold operating expense and grow revenue. Most MedTech and Pharma boards are drifting towards the first path, because it is the one they know how to measure. Their own P&L says the second is the bigger prize.

Run the numbers: price beats headcount

Take two of the largest device companies in the world. Together they sell about $56bn a year. SG&A runs at 32% and 34% of revenue. Operating margin sits at about 18%.

Cut SG&A by 1% and operating profit rises by about a third of a percent of sales. Raise realised price by 1%, with volume held, and the full 1% of sales drops to operating profit. The price lever is worth three times the cost lever.

Scale it up and the gap stays. A 10% SG&A cut, which means hundreds of roles at this size, buys about 3.3% of sales in operating profit. Once. A little over 3% more realised price buys the same, and keeps paying every year the contract runs.

Pharma tilts further. NYU Stern’s January 2026 industry data puts pharmaceutical SG&A at 22% of sales and R&D at 21%, with a 72% gross margin. A 1% SG&A cut is worth 0.22% of sales. A 1% price gain is worth 1%. That is a ratio of about 4.5 to one, and it comes straight from the size of the SG&A line: the smaller the cost base you are cutting, the less each percentage point of cutting returns.

The timing matters for Pharma in particular. IQVIA expects the coming patent expiry wave to wipe out nearly $200bn of brand sales, a wave larger than the cliff of 2011 and 2012. Once exclusivity goes, much of Europe sets price by tender. The commercial question shifts from “what can we charge?” to “what will win?”

One caveat, stated openly. The price lever assumes volume holds. Bid 3% higher without knowing where the market sits and you lose the contract, and the maths runs in reverse. That is precisely why pricing intelligence, not pricing conviction, is the lever. More on that below.

Where revenue leaks: three places between lead and cash

If revenue is the bigger prize, the next question is where it escapes. In MedTech and Pharma it leaks at three points: before the bid, at the bid, and after the award.

Leak one: before the bid

Demand arrives in public. Across Europe, Asia and Latin America it comes as public tenders. In the US it comes as RFPs from group purchasing organisations and health systems. Either way, someone publishes a requirement and a clock starts.

The first leak is the tender you never saw, or saw too late to staff. A notice posted on a regional portal in a language your team does not monitor. A framework renewal spotted with eight days left on a twenty-day window. Each one is revenue a competitor bids for unopposed.

The leak is rarely a lack of effort. Bid teams are sized for the volume they can see, and they spend much of their week reading notices that turn out to be irrelevant. Every hour spent qualifying the wrong tender is an hour not spent on the right one. The result is a pipeline set by capacity rather than by the market.

The fix is coverage and speed. Wider discovery across portals and countries, faster qualification of what fits the portfolio, and less manual effort assembling the response, so the same team answers more bids. This is path two in its purest form: operating expense stays flat, revenue grows.

Vamstar’s Tender AI handles discovery, qualification and bid preparation across public and private procurement.

Leak two: at the bid

In European medtech, the price on the bid usually decides it. MedTech Europe reports that 61.8% of European medtech public tenders are awarded on price alone. Its CEO, Oliver Bisazza, argued this year that price-only procurement undermines innovation and outcomes. He is right, and suppliers should keep pushing for value-based criteria. In the meantime, most contracts are still won on a number.

You cannot hold price by conviction. You can know where the winning price sits.

Without that knowledge there are two ways to fail. Bid above the winning price and you lose the contract, often by a margin you never learn. Bid far below it and you win, but give away margin the buyer would never have asked for. Both look like reasonable decisions inside the bid team. Neither is visible on the P&L as a loss.

Vamstar’s award data puts the average gap between first and second place in MedTech tenders at 26%. On a price-led award, that is how far the winner could have bid higher and still won. It is margin left on the table every day, by companies that won. That gap is the price of bidding blind.

Award and price benchmarking closes it. Vamstar tracks awarded prices across buyers, countries and product categories, so a bid team can see what won the last comparable contract, at what price, and against whom, before setting its own number. The price lever in the previous section only works if you know how much room you have. This is how you find out.

Leak three: after the award

A signed contract is a forecast. It is not revenue.

Between signature and cash, value erodes quietly. Awarded volumes are never fully ordered. Indexation clauses that allow price rises go uninvoked. Rebates are paid on the wrong tier. Renewals lapse because nobody tracked the date. World Commerce & Contracting puts average contract value erosion at 8.6%, with the worst performers losing more than 20%.

The discipline that stops it is revenue assurance: checking what was invoiced against what was signed, line by line, contract by contract. Across a portfolio of thousands of tender awards, with different terms in every country, that is not a job a spreadsheet does well.

It is the least glamorous of the three leaks. It is also the only one where the revenue is already won. No new customer, no new bid, no price negotiation. Just collecting what the contract says you are owed.

Vamstar’s Polaris platform connects contract terms to orders and invoices, flagging the gaps between what was agreed and what was billed.

Why the headcount case wins anyway

If the revenue case is bigger, why do headcount cases get signed off first?

Because they are easiest to count. A role removed is a line on a budget. A finance team can verify it in a quarter. Revenue cases feel softer: they depend on markets, competitors and buyers.

They are not soft. Each of the three leaks has a measurable number attached. Bids answered against bids published. Win rate. Price realised against the winning-price benchmark. Revenue invoiced against revenue contracted. These are as countable as headcount. Most companies simply do not count them yet.

There is a second reason. A headcount case has one owner: the function losing the roles. A revenue case crosses tendering, pricing, market access, finance and sales operations, and nobody owns the whole of it. The leaks sit in the handovers between those teams, which is exactly why they persist.

Easiest to count does not mean biggest.

A four-question audit for your AI budget

Before the next AI business case reaches the investment committee, ask four questions.

1. Fewer people. What share of your AI spend targets SG&A reduction? If the answer is most of it, you are pulling the smaller lever.
2. More bids. How many relevant tenders did you miss or decline last year? If you do not know, that is the answer.
3. Better prices. Do you know where the winning price sat on your last 20 losses? And how far below it you were on your last 20 wins?
4. Revenue already won. When did you last reconcile invoices against contract terms, across the whole portfolio rather than a sample?

A company that can answer all four with numbers has a revenue case ready to fund.

Back to 10 basis points

Nine years of margin expansion in medtech add up to a tenth of a point. AI will change that number. The question is which way boards point it.

The fork is coming, and the P&L already says which path to take. Cost is the lever everyone can see. Revenue is the lever that pays three to four and a half times as much.

Sources:

McKinsey & Company, “Medtech trails market”, chart, 28 April 2026. (Link)
McKinsey & Company, “Fundamentals first: Closing medtech’s value creation gap”, 27 February 2026. (Link)
Chamath Palihapitiya, All-In podcast, as reported by 24/7 Wall St., “Chamath Warns Companies Must Prove ROI from AI Within ‘500 Days’”, 12 May 2026.
A. Damodaran, NYU Stern, “Operating and Net Margins” dataset, Drugs (Pharmaceutical), data as of January 2026.
IQVIA Institute for Human Data Science, Global Medicine Use Trends 2026, March 2026.
MedTech Europe, CEO Oliver Bisazza, “Buying European is not enough: Europe must buy better”, Euractiv, May/June 2026.
World Commerce & Contracting, The ROI of Contracting Excellence; cited in “From value leakage to better outcomes: why contracting needs integration”, June 2026.
Disclosure: tender award and price data cited in this article come from Vamstar’s own platform.