Story points: the final chapter?
I've helped a lot of teams over the years trying to get a handle on the concept of estimating using story points, and I've gotten very good at explaining away the apparent arbitrariness that they are infused with. And whilst I can't always remove that skepticism, I can get teams to see their value.
Coming from a waterfall delivery origin, and having witnessed countless short-lived Gantt charts due to a reluctance to accept the unpredictable nature of the universe, the concept of estimating with story points resonated with me immediately. But that was a long time ago, and now being older and slightly more wizened, I have a question:
Do we need to bring Story Points in at all?
I've seen multiple suggestions of replacing story points as the unit of velocity with simply measuring throughput - how many issues do you get through in a sprint? Proposals include temporary removal whilst teams become used to operating with new processes, and complete removal - doing away with Story Points altogether.
The more I think about it, the more the unthinkable removal of story points appeals to me: What value do they add that simply measuring throughput doesn't provide? Knowing how many issues we generally get through in a period still allows us to plan work – so why do we use them?
Is it about getting more granular than the sprint time-box to know when to expect issues? But why does anyone was to know that? If it's about trusting your team to deliver - just switch back to hours, minutes and seconds. Sometimes, specifically in Agencies, it is about clients trusting you to deliver. Fair enough, but a better way to build trust would be to work smaller, and consequently have something to show clients faster.
What about knowing the size of issues so we can establish if we need to break them down further? Be honest, you don't need a planning poker session to work that out.
You could address both of these contentions and use throughput to gauge velocity by breaking stories down to their smallest form. This creates similarly-sized chunks of work, making work more predictable enough to plan into your timebox of choice.
Anything that helps teams plan is valuable. However, anything that is there to give a granular overview of how far along you are toward your next big goal does not support delivering value faster. It's akin to looking at a Gantt chart. Does that mean story points therefore just a hangover from waterfall? There to make Agile (big A) methodologies more palatable to the discerning business owner?
Story points were a step in the right direction - they placed our innate human inability to estimate accurately front and center, accepting uncertainty. But they did not go far enough; they did not remove the overseeing - the granular reporting – and most importantly they did not remove the hard-and-fast-deadline. They merely offered a level of abstraction and allowed everyone to pretend those deadlines weren't still looming there.
I think it's time to put away those planning poker cards. It's hard to believe that I may never get to explain the Fibonacci sequence to nonplussed teammates ever again.