TBM 259: The Predictability Trap

Predicability, predictability, predictability…

If the growth of your business relies on the predictable delivery of new features, then you will be concerned about teams meeting delivery commitments and finishing things when they said they would. If you run a consulting business, and your profits rely on charging enough and limiting how much time you invest, then you will be concerned about hitting estimates and predictable delivery. Your earnings depend on it.

But suppose your business relies on continuously innovating on behalf of your customers, keeping the complexity of your product in check, and figuring out a path to sustainable and differentiated growth. In that case, you will be seeking a different type of predictability. 

You will want your team to innovate predictably. You will want your team to predictably determine what's truly valuable and ship that thing. You will want your team to predictably tame the complexity of the product so it can keep moving and innovate quickly. You will want your team to predictably trim away the parts of your product that are not delivering value to customers or that can't sustainably provide value to customers in a profitable way to your business.

An uncomfortable percentage of things we try to deliver miss the mark. Sometimes, that is obvious, and the feature falls flat. But sometimes, it's a lot more subtle. Teams ship features that each missed the mark by just a bit, but together, collectively, they really miss the mark. The’ve added a lot of complexity to the product, and are not earning enough money for that complexity. Instead of unit economics being favorable as they add more functionality, it's costing them increasingly more to make the product valuable to customers.

Over the years, I've rarely observed a team that is predictably delivering value for a company get hassled about predictable delivery or meeting deadlines. People generally stop worrying about those things once they've established that trust and confidence.

However, when a team is struggling and losing the confidence of its managers and leaders, the conversation often shifts to one of predictability. "We want this team to deliver what they said they would deliver, and when they said they would deliver it," we say. Success becomes doing what you promised to do in the allotted time. And this makes perfect sense in many ways. We think, "This team needs to take care of the basics before it can do any of that fancy stuff." I'll always remember a manager saying, "I don't care if this takes three times as long, as long as it takes predictably three times as long."

But I would like to challenge this idea for a moment.

Regarding delivery, you only have so many levers to deliver more predictably. You can reduce batch size, minimize work in progress, collaborate more closely, work in shorter time boxes, focus more, start together/work together/finish together, limit the reactive tasks you get pulled into, and limit multitasking and context switching. 

We think that better plans and better estimates make us deliver more predictably. But in my experience, that's not really what does it. All this is to say that if you want to deliver more predictably, there's a straightforward playbook. Most people don't want to run that playbook because it would mean saying no, pushing back, and potentially appearing to do less. Still, if your goal was to deliver more predictably, there's a handful of things that you would do immediately. I'm not saying it's an easy playbook. But there is a playbook that outputs "more predictable delivery."

However, this exposes the real underlying problem. Many teams that struggle with predictable delivery were, at some point, highly predictable. They delivered predictably—so many products and features—that it has left them in a situation where they have their back against a corner. The business can't afford to let them slow down to speed up. They can't afford to let them run the playbook I mentioned above. They want them to deliver predictably and deal with all the debt the company has accrued without slowing down! 

Yes, the team may need to do a lot of revamping of how they work to deliver. You can't score on shots that you don't take, so the ability to get something smoothly into production, instrumented, and measuring to see if it works is an extremely important muscle that you'll have to get right.

But my key argument is that you cannot make predictable delivery the goal. You have to put value front and center and put sustainable and differentiated growth front and center. Predictable delivery needs to be in service to that, not the end goal.

Because if you don’t, then one day you wake up and predictable delivery is impossible.