Deciding what features to work on as a Product Manager

I dont have all the answers. Don’t trust anyone who says that that they do. Feature polls and customer centric product development are cliches. Be careful when you hear these words from product managers. The reality is different and difficult.

Figuring out what to work on is hard and changes often. It does not matter whether you are on new or established products. New shit always comes to light. So.. how do you pick what you work on? Here is what I’ve learnt.

You can work on features that will appeal to the largest set of customers or you can chase new customers by building features that get new customers to use your product. Doing the former is easy. And, it is easy to explain to the team and really hard to pull off.

In the former category, things like performance and simplification always appeal to existing customers and build a larger gap between you and the competition. New product managers never focus on these features. These features are not sexy. They dont motivate the engineering team since they are hard. You do not get rewarded for picking these features by execs and you do not hear the cheers from the users until the improvements are significant. So.. you spend a of time in anxious anticipation. Questioning whether this will work or not. Thus, releases planned around enhancing existing features generally die on the vine.

Feature Name Who will use it
(Segment or Persona)
What % of user base is this persona
Performance improvements All 100%
Usability enhancements All? 100%
<New Feature name> Existing Segment 20%
<New Feature name> New Segment  0%


When you work on new or forward looking features, your team is going to give you a hard time. It should because you can’t rationally explain how you are trying to predict the future and why your are building features for people who don’t exist and are not your customers today. So, you have to tell a story. A believable story. A story that you need to back up with numbers. You need data on the size of this new segment of customers. Their proclivity for change. Why your 1.0 solution will be better for them than what they use today? This is tough too. But, if you tell your story right, your execs will love you… but mostly for your story telling ability.¬†Most PMs do not know how many users will really use the new features consistently. Most of us do not have reliable past metrics to predict adoption numbers for the future.

Try to fill up this matrix for every release and see if your predictions match up. This exercise gets significantly harder as your product grows and is adopted by lots of users that can’t be put in clearly defined categories.

Feature Name Who will use it
(Segment or Persona)
What % of user base is this persona Actual adoption
Performance improvements All 100% 100%
Usability enhancements All? 100% 100%
<New Feature name> Existing Segment 20% ?
<New Feature name> New Segment ?

My team and I have just started to do this now. I’ll keep you posted on how this goes for us. Do read my post on “developing world class products“. I still believe that starting with the business goals for each release and managing for outcomes is very important. The features you build have to get you there.