One year of shaping up

This is a long post. TL;DR; - Shape Up has been great for the Humly Product development team, and we do not miss scrum. It's not all roses though, and you still need to run your retrospectives, and iteratively improve and adapt to your reality.

In software development teams, agile feels like a given nowadays. At the same time, I never felt that scrum or kanban resulted in a great output or a pleasant work environment. Too much ceremony, and often focus on process over outcome. A lot of meeting overhead, silly titles like “scrum master” and difficulty in getting other stakeholders in the company involved. When joining Humly last year, I saw the same symptoms I’d experienced in the past and felt like enough was enough. Having recently read https://basecamp.com/shapeup and resonating with many things in the book I and team team decided to go all in, ditching the traditional two-week sprints.

One year in, and 6 successful cycles completed, we’re not looking back. Of course, it's been a journey with bumps along the road, but with productive retrospectives, we have clearly identified the good and the bad and iteratively improved how we work. Some high-level aspects that we have experienced in 2023:

The Good Parts

- Two-week “Cooldowns” - This gives us a breather, allows us to tie up loose ends, and lets the team regroup and recharge.

- Ownership - the format of pitches has put the right level of fidelity on project definitions and has created much stronger ownership throughout the team.

- Longer cycles (sprints) - Having a full 6 weeks to immerse yourself in a project avoids the stress of going from sprint planning to sprint demo in two weeks. This lowers the cognitive load and allows us to get in the flow.

- Easier to look further ahead - The ownership and longer cycles have given product owners and leads more room to lift their gaze and look further into the future.?

- Less process overhead - More time to create value, and less time sitting in meetings.

The Hard Parts

- Circuit breaker - A key aspect of Shape Up is to kill projects that don’t reach the definition of done at the end of the cycle. With scrum sprints, it's common to say “Just one more sprint”, and projects get dragged out into infinity. This makes it difficult to plan and lowers the morale of the team. Sticking to circuit breakers, and not saying “Just one more cycle” is easier said than done. What we have learned throughout the year is to make sure we improve our estimations, get better at defining our “definition of done” and force ourselves to make the hard call and stick to the circuit breaker.

- Setting the right appetite - One of the hardest things in software development is making good estimations. We are more often than not building things that have never been built before, and the unknown unknowns are always there. Shape Up does not solve this, it has some general ideas about how to mitigate, but finding the right scope is still difficult.

- Finding a home for all the small stuff - We’ve ditched the giant central backlog, and are betting on the top priority pitches each cycle. But as with any product in production, there are a myriad of smaller fixes that in aggregate create a lot of value. We’re not able to deal with all of them in the cooldown, and it's been tricky to find a good way to bundle these smaller things into bigger projects.

- Planning for the unplanned - In the early days of Shap Up we maximized our appetite to get as many projects done as possible. But we soon realized that we needed to lower this appetite to make room for the unplanned. Maybe a new release in previous cycles causes an influx of support tickets, or maybe a change in volume or traffic forces us to address some infrastructure problems.

- Ensuring time for testing - We always strive for “done means deployed”. But we found ourselves working into the final days of the cycle, doing the finishing touches on the last day. This meant that we rarely had time to do proper testing. Even if 6-week cycles are three times longer than 2-week sprints, 6 weeks is still a short time, and it is important to set the appetite correctly, ensuring that there is enough time to have working solutions ready early enough for testing.

The Future

We will continue to run Shape Up and learn as we go. It is not a fix-all solution, and the process comes with its own unique challenges. We are looking forward to improve the way we work with functioning slices of work, and better QA process.

If you are curious to try it out with your team, but want to discuss the pros and cons, don't hesitate to DM!