Skip to main content

Craig Kerstiens

A product management blueprint

I find myself having more conversations with startups – both small and large – about product management. I’ve blogged about some of the tools in my chest here but I haven’t talked much about my “blueprint” for product management, which I find myself laying out in many conversations over coffee. What follows is this process I’ve used a few times over with new teams to get product and engineering moving together, shipping in a predictable manner, and tackling bigger and more strategic projects.


I need to know how to work with my team, what their working styles are, and how we interact. This starts by simply interacting – specifically, outside of the office. I heard a similar opinion recently from Chris Fry (who ran engineering at Salesforce and Twitter) when he remarked something to the effect of: “you can tell a good PM from a bad one based on if he goes to drinks with his team.” Without getting hung up on whether it’s beers or coffee, it’s more about socialization with your team and time outside the office. My personal approach: expect a dinner invite over to my place when I take on running product for a new team.


Once you’ve started to build some rapport, it’s time to get down to business. If being able to quickly commit and ship something isn’t a problem for you, then it’s easy to just assume this is working. In reality most teams I encounter that need PM support don’t have shipping nailed down. You probably already know if you fall into that category of feeling like you can commit and ship vs. not, so if you’re not able to do that a few tips:

  • There’s some projects that everyone wants to ship that’s been tried over and over, don’t tackle that first.
  • Shipping something is better than nothing. It doesn’t have to be the right thing.
  • Sometimes you don’t have to ship something to get velocity, you can launch things you already have
  • Kill scope Test things earlier and more iteratively, the more you can validate or try something without requiring a large investment the more everyone feels better about the direction you’re heading.

The key here is to commit to projects, deliver, and move on. Your velocity depends solely on delivery, not tasks, not sprints, not projects, etc. If you haven’t shipped anything in a year, then your velocity for the year is zero. At a later point you should move from the focus on shipping anything to shipping the right things, it’s more important to ship 1 thing that moves the needle than 10 that don’t, but that’s a later concern.

Killing things

On the note of killing scope… I’ve heard it articulated at times, that some engineers are happy when certain PMs show up because it means less work for them. When you go over to an engineer’s desk are you creating more or less work? The answer should be less some large percentage of the time. If you can find a way to accomplish your goals with less effort, it’s always a win. Every project everywhere always needs more time or money, what’s more innovative is how you can help a project to ship without one of those two.

At a broader perspective than just scope – one of the biggest ways product can help engineering is by pushing harder for killing off features and the scope of a product. There’s a good test on if something is ready to ship: if you tell beta users you’re killing it and they yell at you that you shouldn’t kill it, then it’s ready to ship.

If you’ve already shipped things, but they’re not delivering value or not being used, kill them. It’s that simple, it may have been a great idea at the time, but either invest in making sure it’s used or kill it so you don’t have to maintain it.

Roadmap planning

Usually getting velocity and killing things takes 3-6 months to really take full effect. At this point a team feels like they’re not under a pile of technical debt, and they can commit to shipping projects. This is the point when product and engineering are melding and you can really start to have fun about where you’re headed. At this point I’ve seen a huge mix of where engineers are more actively or less actively engaged in this process. And the reality is this is everyone’s job to be thinking about where you’re headed as a company – at least that’s the case for any company that classifies itself as a startup.

My favorite tool for this is a team gridding exercise, you can read more about this here and here. This is often best conducted at an off-site where you have an opportunity for casual conversation which can foster broader thinking beyond the obvious bug fixes or smaller product improvements.

One item of note I’ve heard from teams that have done this or similar exercises is they still have trouble deciding what to do after the fact. The role of product is to get to that decision. The most important part is getting to a decision and not the perfect one, gather data, decide, revisit as you go along. All of this isn’t to say that it’s an arbitrary decision, customers, data all inform that as well as the effort to impact matrix exercise, but in the end a clear direction isn’t executed on consensus.

In conclusion

There’s really no end or done when it comes to the role and the work.

There’s always another milestone and the market is always moving around you. But once you’re able to execute predictably and think in an ordered sense about your roadmap, you’re in a position to be able to monitor and adapt to the market, and even more so experiment and shape the market yourself. At that point you have to keep doing it and then the hard part becomes finding ways of keeping a fresh perspective protip: customers are an important part of that equation

Have tips/tricks/practices that I completely missed here or that you disagree with? I’m always happy to talk with others so drop me a note