“How long will it take?” A question full of risk, yet, one that gets asked of IT leaders frequently. IT leaders also ask it of their team, so it can be a double-edged sword. How can we deal with these requests for estimates?

We dislike giving estimates because they are wrong so often. Or there are caveats that get forgotten. We might say that it will take three months to upgrade system A, but only if we don’t need to update system B as part of the effort. People will remember the three months, but they won’t remember the system B caveat.

The optimists among us (often management) tend to underestimate. I am especially guilty of this one. The pessimists (often the people doing the actual work) tend to overestimate if they are willing to give an estimate at all.

Sometimes you need a general idea, an order of magnitude, of how long something will take. Sometimes you need a definitive schedule that won’t change. Sometimes you need an estimate that is in between to plan your resources.

One way through this is to think about three types of estimates. The name is the amount of time that you spend to make the estimate. There is a direct tradeoff between the time spent on an estimate and the accuracy so be careful how you ask for each.

• The Five Second Estimate
• The Thirty Minute Estimate
• The Four Hour Estimate

The Five Second Estimate

We need these estimates in the idea stage such as a conversation about possibilities. We are in a meeting and someone throws out an idea and you need a sense of timing. The wrong question is “How long would that take?” The correct question is “Will this take days, weeks, months, or years?” Correct answers should not include numbers. There will be many assumptions that are subconsciously made in a five second estimate. Those assumptions don’t usually get communicated, so be careful about spreading Five Second estimates.

The Thirty Minute Estimate

Simply listing the major tasks, risks, assumptions, and interfaces allows you to get to the Thirty Minute estimate. This level of estimate is usually sufficient for prioritization and resource allocation. There is some supporting documentation that you can include if you need to communicate with others. The Thirty Minute estimate is usually sufficient for small to medium-sized projects.

The Four Hour Estimate

Large projects need the Four Hour estimate. The output is a standard project plan with resources and a mostly complete task list, full risk assessment, and critical dependencies. Of course, it may take longer than four hours for major projects.

Bing clear about which of these three estimates you need will help set expectations. If you can teach those outside IT about these three, communications can be clearer. Good Luck!