Straight answer

Why does IT always cost more than planned?

Usually because the first number was set when the least was known, and approved as if it were firm. Data, integrations and business change are understood later. The overrun is often the true cost arriving late. Ask for a range, not a single figure, and fund the uncertainty openly.

How can the board tell whether it is under control?

Signs it is under control

  • Cost estimates come as a range, with the assumptions that drive the top of the range written down.
  • Contingency is held visibly and its use is reported to the board.
  • Business costs (people released from day jobs, training, temporary staff) are in the budget, not left out.

Signs it is not

  • The approved budget was a single figure with no range.
  • Change requests are approved one by one, but nobody reports their total.
  • Running costs after go-live were never estimated, or are reported only once they are due.

What is the one question to ask?

Ask at the next meeting

What are the three assumptions in this budget most likely to be wrong, and what would each cost if it were?

A strong answer

Named assumptions (for example the state of the data, the number of integrations, the availability of business staff), a cost for each if it proves wrong, and how much contingency covers them.

A weak answer

“The budget has been through procurement and the supplier has given us a fixed price.”

Where does the extra cost come from?

Technology projects are costed at the point where the least is known. The supplier prices the work it can see. The parts that are hard to see, such as poor data, old systems that need connecting, and the time the business must give, are underestimated or left out. A fixed price from a supplier usually fixes only the supplier’s own share, and anything outside the agreed scope comes back as a change request.

Running costs are the other gap. Licences, support and the internal team needed after go-live often grow quietly once the project closes. They then sit in the IT budget for years, where nobody connects them to the original decision.

What might the board be getting wrong?

Many overruns start in the approval meeting. A board that asks for a single number gets one, and it is usually the lowest number that can be defended. A board that challenges contingency as padding gets a budget without any. Directors sometimes approve the capital cost and never ask about the cost of running the result. If the executive has learnt that an honest range will not get through, the board has taught it to give a hopeful figure. That is a board habit to change, not a management failing to punish.

What should the board ask for instead?

Ask for a range, with the assumptions behind it, and approve against the realistic figure rather than the lowest. Ask for contingency as a separate line, released by a named person, with its use reported each quarter. Ask for the full cost of ownership over several years, including the business’s own time and the running cost after the project ends. Then judge management on how well they explain movements in the forecast, not on whether the first number held. A team that reports early that costs are rising is doing its job.

Where to go from here