Hanwen XueChief Information Officer · Essays & perspective
← All essays

IT economics5 min read

Narrower, not thinner: the IT budget after volume-based procurement

Volume-based procurement has cut distributors' margins on the medicines it covers. The usual response, an even cut across the IT budget, weakens the systems the new model depends on. Retiring whole applications protects them.

China's volume-based procurement began in 2018 with a pilot covering eleven cities, Chengdu and Chongqing among them, and has since extended through successive rounds to several hundred medicines and many categories of high-value consumables. Manufacturers bid for a committed share of public-hospital volume, and winning bids have often been far below the prices they replaced. For patients and the insurance funds the reform has saved a great deal of money. For distributors it has done something less discussed: it has shrunk the value of the goods on which our fees are calculated, while leaving most of our work unchanged.

What procurement reform did to a distributor's margin

A distributor is paid, in effect, a percentage of what it delivers. When the price of a common medicine falls by half or more, the cartons do not get lighter, the hospitals do not move closer to our warehouses, and the delivery note still has to match the hospital's drug dictionary. Our revenue on that product falls; our cost of handling it does not. Procurement rules have also shortened the time hospitals take to pay for centrally procured medicines, which helps our working capital, but faster payment does not restore the margin.

At Shuheng the gross margin on hospital distribution fell from 6.4 per cent in 2019 to 5.6 per cent in 2022, and net profit fell 14 per cent in 2022. The two-invoice rules introduced from 2017 had already consolidated the industry by limiting the invoices between manufacturer and public hospital, and many of the regional distributors we acquired in those years arrived with their own systems. When procurement reform squeezed the margin, we were carrying an IT estate built for a larger margin and a more fragmented industry.

The finance response was predictable and, from finance's position, reasonable. Every function was asked to cut 15 per cent. For IT that meant taking roughly RMB 37 million out of a budget of RMB 248 million.

Why an even cut is the expensive choice

An even cut looks fair because everyone shares the pain. In an IT budget it is rarely fair in its effects. Most of a distributor's IT spending is maintenance of applications already in service: licences, support contracts, contractor hours and infrastructure. A 15 per cent cut spread across 438 applications reduces the care given to all of them, including the order platform, the warehouse system and the settlement interfaces on which every delivery depends. Meanwhile it preserves every report, portal and customised interface that someone once asked for, each a little less well looked after.

The result is a thinner estate. Every system survives, and every system becomes more fragile. The costs of that fragility arrive later and elsewhere: an outage at a logistics centre, a delivery-note format that falls behind a hospital's upgrade, a security patch applied a month late. They seldom appear in the IT budget, which is why the even cut looks cheaper than it is.

The alternative is a narrower estate. Instead of reducing everything, you decide what to stop doing. In 2023 we began retiring 162 of our 438 applications over eighteen months. Many were reports built for a single hospital or a single manufacturer; some duplicated functions after acquisitions; three were separate warehouse systems doing the same job in different provinces. We froze a RMB 38 million data-platform programme that our commercial division wanted, and we increased security spending.

The objections were real. Our commercial leadership saw the data platform as future revenue from reporting services to manufacturers, and they may yet be proved right about its value; it was deferred, not cancelled. Regional managers lost reports their customers used, and one hospital group complained to our chairman. A narrower estate is a political cost paid immediately in exchange for a technical benefit realised slowly. That is why even cuts are so common, and why boards rarely see the alternative presented.

Narrowing also needs a rule for what may come back. We agreed that a retired application could return only if its business owner found the money inside the discretionary budget and named the people who would use it each week. In the two years that followed, owners asked for eleven to return. Nine were rebuilt as standard reports on the common platform, and two stayed retired.

Systems that matter more after reform

Procurement reform does not only reduce margin. It changes which systems carry the business, and three have become more important rather than less.

The first is order and delivery accuracy. When a product's margin is thin, one error that leads to a return or a second delivery can erase the profit on many correct ones. Systems that keep our product master aligned with hospitals' drug dictionaries now pay for themselves faster than they did.

The second is settlement and payment data. Centrally procured products carry their own payment arrangements and reporting, and a distributor that cannot reconcile what was delivered, accepted and paid will watch its working-capital gain disappear into disputes.

The third is traceability. When we had to verify manufacturers' code data from 2024, we changed a single warehouse system instead of three, because of a consolidation decided for cost reasons a year earlier. I did not foresee that benefit when I argued for consolidation. It has proved worth more than the savings.

By 2025 the run cost of our estate was RMB 131 million, against RMB 172 million in 2022, and security spending had roughly doubled. Total IT spending has since risen again, to RMB 236 million, but it now funds fewer systems more thoroughly.

For boards, the lesson is a question to ask whenever margins fall. Do not ask only whether IT has taken its share of the cut. Ask which systems the business has decided to stop running, and whether the systems it cannot operate without are funded better or worse than before. If the answer is that nothing has stopped and everything has been trimmed, the company has chosen the expensive option and deferred the bill to a year when nobody will connect the outage to the budget meeting.