Posts

“Continuous Delivery” has become a catchy term over the years. The real way to make this real for your team is to focus on shortening all of the cycles before it. And then naturally, the delivery will become continuous too. Years ago “Small Releases” meant monthly drops; then weekly, and today our goal is Continuous Delivery: releasing to production after each stable change. Continuous delivery isn’t only about pushing more code more often. It’s about discipline and striving for clearer scope, faster feedback, and shrinking the design, review, test, and deployment cycles. ...

Golden Stories are a small set of well-understood baseline stories that the team agrees to peg with certain point values. These are used as references for future pointing. As the team estimates new stories, they can compare them to the golden stories which act as an established agreed upon baseline. So when a story “Fix a typo in the main menu” arrives the team can compare the new story directly to the golden Stories: “Is this typo harder or easier than the “Login” golden story?”. The team calibrates the estimate and explores and hidden scope or assumptions. ...

A falling velocity usually signals code quality problems - rot - not laziness. Builders avoid refactoring when the test suite is weak, because without tests they fear changing the working code. And that fear slows down delivery. And this mounting pressure leads to story point inflation which props up the recorded velocity score to hide the actual velocity decline. Encourage the team to treat testing and refactoring as first-class work. ...

A rising velocity line doesn’t always mean the team is actually faster. When managers push for speed, builders will often unconsciously de-value points to make the numbers look better. Points become a currency, and inflation follows. That upward slope can be pressure, not progress. Velocity is a measurement, not an objective. Don’t force the thing you’re measuring. We estimate work during the planning meeting so the stakeholders can choose stories and plan for the next iteration. And that estimate is a forecast, not a promise. ...

We update the burn‑down and velocity charts only with points for stories that have passed their acceptance tests. Recording only done stories prevents optimistic noise and makes the charts honest signals. After several iterations both charts start to reveal a slope. And the burn‑down slope can help to predict the date for the next major milestone. The velocity slope helps to show how effective and consistent the team is running (early iterations will be noisy; but expect stabilization after a few sprints). ...