Gut Feel, Past Averages Or Breaking It Down: Three Ways To Size Work

There are three ways to size work, and most teams use one of them for everything without ever asking if it fits. Gut feel is fast and wrong in one direction. Past averages are cheap and only work if the new job resembles the old ones. Breaking the work into tasks is the most accurate and costs the most time to do. Which one you should use depends on how much the answer matters and how much your team already knows about this kind of job.

Here is the short version. Use gut feel for work under a day. Use past averages for work you have done shapes of before. Break it down when someone outside the team is going to hold you to the number.

Gut feel

Somebody who has done the job looks at it and says a number. That is the whole method. It takes a minute and it costs nothing.

It works when the person estimating has personally done that kind of work recently. It falls apart when the work touches other people, other systems or anyone's approval. Gut feel almost always misses the waiting: the review that sits for two days, the credential nobody has, the question that needs an answer from someone on leave.

Suits: solo work, small changes, anything under about a day. Any team size, because only one person is involved.

Cost to produce: a minute.

Past averages

You look at how long similar work took before and you use that. If the last four onboarding builds took roughly a fortnight each, the fifth is probably a fortnight.

This is better than gut feel for one reason: the past number already includes the waiting, the rework and the meetings. Nobody has to remember to add them, because they happened.

It needs two things. You need a record of what things actually took, not what you said they would take. And the new work has to genuinely resemble the old work. A client with a different approval chain is a different job even if the build is identical.

Suits: teams of any size doing repeatable kinds of work. Agencies, service teams, anyone with a project shape they keep selling.

Cost to produce: a few minutes, if the record exists. Weeks, if you have to start keeping one.

Breaking it down

You split the work into pieces small enough that a person can picture doing each one, then size each piece and add them up. Pieces of a day or two are about right. Anything you estimate as a week is really a piece you have not understood yet. Finishing work is the piece missed most often, which is why the last ten percent eats a third of the time.

Breaking down is the only method that surfaces the parts nobody had thought about. That is its real value. The number at the end is useful, but the argument you have while producing it is more useful, because that is when someone says "wait, who is writing the migration script".

The catch is the time. A proper breakdown of a two month project is most of a day for two or three people. That is a real cost, and it is the reason teams skip it and then miss the date.

Suits: work over a few weeks, work with a contract attached, work where a miss costs money or a relationship. Needs at least two or three people who know different parts of the job.

Cost to produce: hours to a day.

Picking one

A rough rule that holds up: spend about one percent of the work sizing it. A two day job gets a few minutes of thought. A three month project earns a day of proper breakdown. Anything in between takes past averages if you have them.

The mistake worth avoiding is using gut feel on something big because a breakdown felt like a delay. It is not a delay. It is the cheapest place to find out that the job is twice the size you sold. The expensive place to find out is halfway through, when you have to tell a client the date moved.

Mixing them on purpose

The strongest approach uses two methods and compares. Break the work down, add it up, then check that total against what similar work actually took. If the breakdown says three weeks and history says six, you have missed something, and it is usually the waiting rather than the building.

When the two numbers disagree, trust history. Breakdowns are produced by optimistic people on a good day, sitting in a room with no interruptions. History includes the bad days.

What to do this week

Pick your next piece of work over two weeks long. Size it with a breakdown, write the number down, and write down what similar work took last time. Then, when it finishes, record what it actually took. Do that three times and you stop arguing about estimates from memory, because you will have a record instead. Keeping that record for eight weeks turns it into something you can plan from.