Our annual planning process is well run and I have no complaint about the people in it. Departments submit capital bids, each with a case, a benefit and a payback period. The bids are ranked and funded down to a line. Alongside that sits the operating budget, a mostly flat number labelled IT operations, which is expected to fall by a small percentage each year as evidence of efficiency.
Everything about that structure funds change and nothing about it funds continuity.
A storage platform reaching the end of its supported life is not a project. It has no benefit statement, because replacing it delivers exactly what the organisation already has. It cannot be argued for against a customer-facing initiative, and it will lose every time it is tried, because the honest case for it is that something will eventually stop working. So it does not get replaced on a plan. It gets replaced after an incident, when urgency produces money that a forecast never could, and at a worse price with less choice, in a hurry.
The second effect is subtler and now larger. Most of what we buy has moved from a purchase to a subscription. That spend lands in the operating line, which grows because consumption grows, while the process reads any growth in that line as a failure of discipline. Two years running I have had to explain that an increase was the direct consequence of a capital project the same committee approved, whose running cost nobody had asked about.
The changes we asked for were modest. Every project case now carries three years of run cost, agreed with finance, and that figure moves into the operating budget when the project closes. We publish a five year refresh schedule by asset class, with dates and prices, so that replacement is a known commitment rather than an annual argument. End of support dates go to the committee as a register. And the operating budget is now reported in three parts, so that growth in consumption is not mistaken for waste.
A budget process that only funds new things will, in time, own a great many old ones. Ours had been doing it for a decade.
– Serguey Shinder
