Estimation

A four-part series that acknowledges all the challenges of software project estimation and offers actionable advice for pursuing excellence this notoriously hard skill.

Estimates and Deadlines are often controversial in startups, with developers disliking them. However, estimates are simply educated guesses that help communicate the difficulty of a task. They are not commitments or Deadlines. Real Deadlines arise from external factors while artificial Deadlines create stress without value. Estimates can help identify complex issues, find quick wins, and motivate developers, though they do require some time investment. If used properly, estimates and Deadlines can improve communication and productivity instead of causing toxicity.

While estimating timelines for software projects is difficult, Gergely argues that it comes with benefits. Giving estimates helps set Deadlines that focus teams and drive faster progress. Missing a deadline is usually not a big issue, as long as delays are communicated well in advance. Gergely's own experience shows that teams tend to work faster and more productively as Deadlines approach. Estimates also force teams to reflect on what went wrong and improve for the next time. Over time, engineers can get better at estimating through practice and learning from mistakes.

Rod discusses four reasons why estimation is so important, and shares some tactical tips around estimate granularity, including bugs in sprints, Retrospectives and more. 1. Estimation forces us to think about what we’re going to do before we start writing code 2. Estimation enables meaningful discussion between engineers pre-development 3. Estimation is valuable input for informed product Scoping 4. Estimation gathers data we can measure against at the end of the project

The Transferwise team asserts that estimates aren't required and pushes back on some common reasons for estimations.

Kiryl takes a look at pros and cons of the planning poker estimation technique, and suggests magic estimation as a fast and effective alternative.

Jacob proposes an estimation technique that captures both time and uncertainty. He recommends breaking down work into smaller tasks, estimating uncertainty levels, calculating expected and worst case times, refining estimates as needed, and tracking actual times to improve future estimates. The key is to choose a system and stick with it so you can calibrate estimates over time. While there are other techniques, capturing uncertainty is critical for effective estimates. Evidence-based scheduling, where you track actual times from the start, can be highly accurate but requires stable teams and long-lived projects.