Why estimates are always low
Underestimation is not carelessness. Every force acting on the person producing the number pushes in the same direction.
Studios and freelancers underestimate consistently, across decades, across industries, and by broadly similar margins. That consistency is the interesting part: something producing an error in one direction every time is structural rather than personal.
The forces, named
- You estimate the work you can imagine. The work you cannot imagine is not zero, and it is most of the overrun.
- You estimate the task, not the visit. Briefing, waiting for assets, three rounds of feedback, the call about the call — none of it is the work and all of it is on the clock.
- You estimate the good version of the project. The one where the client answers on Tuesday and the brief does not change.
- You want the job. Nobody estimates high when a competitor is quoting.
- You have no record of what similar work actually took, so the estimate is anchored on how long it feels like it should take.
Optimism is a permanent feature of human estimation. An honest record of what previous work consumed is the only correction that reliably works, and it is the one most studios do not have. For a related discussion of how activity signals can be manipulated, this guide provides useful context.
Estimate the elapsed project, not the hours of work
A piece of work containing sixteen hours of effort does not take two days. It takes three weeks, because it waits for a review, then for content, then for a decision by someone on holiday.
Those waits cost money even when nobody is working: the project stays open, it is re-entered mentally each time, and it occupies a slot that could have held something else. Estimating in effort hours while quoting a delivery date is how studios end up committed to both a low price and an impossible schedule.
Count the rounds
Feedback rounds are the single largest uncounted item in creative and development work. An estimate that assumes two rounds and meets four has doubled the review effort, and review effort is rarely trivial. For a broader framework, PMI's overview of project management summarises the standard concepts behind planning and delivery.
State the number of rounds included, in the estimate, as a number. It costs nothing, it is the most commonly exceeded assumption in the whole document, and it converts an argument into a change request.
Break it down until the pieces are checkable
A single number for a whole project is unfalsifiable and therefore unimprovable. Broken into eight or ten items, each one is small enough that you can say whether it is plausible, and afterwards you can see exactly which item was wrong.
The useful side effect is that decomposition itself surfaces work nobody had thought about. Half the value of the exercise is the items you write down and had not been planning to do.
Apply the multiplier honestly
Many people know their estimates run low and quietly add a margin. That is reasonable, and it works better when it is explicit and derived rather than instinctive.
If your last ten projects came in at an average of 1.6 times the estimate, the multiplier is 1.6, not 1.2 because that is easier to justify. Padding applied without measurement gets negotiated away first, because nobody — including you — can defend a number that came from nowhere.