Agile 101: From Waterfall to Iterative Planning, Splitting with INVEST, and Estimating by Ranking

In the first stage titled “Agile 101 based on continuous integration”, the team shifts from a traditional waterfall development model to a lightweight agile approach, introducing one‑to‑four‑week time‑boxed iterations that include requirement analysis, coding, testing, and demo acceptance.

The team discusses the need for a reliable project plan to answer when the product can go live, and decides to adopt a new planning method that focuses on three steps: splitting, estimating, and scheduling.

Step 1: Splitting (INVEST principle) – User stories are broken down into small, independent pieces that can be completed within half a day to two days, allowing early testing and feedback. The INVEST acronym stands for Independent, Negotiable, Valuable, Estimable, Small, and Testable, each described in detail.

After splitting, the development lead presents the refined backlog to the tester, who raises concerns about large items that may be hard to verify, leading to a discussion on partial verification through logs or database checks.

Step 2: Estimating (Ranking Method) – The team uses a relative sizing technique originally described on InfoQ, placing printed cards with requirements on a wall and ordering them by perceived effort using three rules: place similar‑size items below, larger items to the right, and smaller items to the left. The method relies on group comparison, a Fibonacci‑like scale (0, 1, 2, 3, 5, 8, 13,…), and consensus after a brief discussion.

“Take a requirement card, stick it on the wall, and then each person adds their cards following the rules: no discussion before permission, place cards after the previous person, and align them based on relative size.”

Images illustrate the wall‑sorting activity and the final layout of the cards.

The estimation approach assumes that at least two people understand each requirement, testers do not participate in sizing, all items are already split into similarly sized pieces, and the total number of items is relatively large.

Step 3: Planning – With relative sizes assigned, the team still lacks an absolute time unit. They discuss converting the relative points into person‑days to forecast overall project duration, proposing a meeting with all stakeholders to finalize the schedule.

The article ends with a note that the discussion will continue in the next installment.