Managing Startups: Objective-Key-Results (OKRs)

Having spent the past year working with and incredible team on a product I love, I thought I’d share a management tool that we adopted from Google called Objective-Key-Results (OKRs). OKRs are a quarterly panning tool that help you and your team focus priorities and align around a common set of objectives.

Before I lay this out, I want to point out that this gets trickier if your company is pre product-market fit (defining P/M fit is all over the web, so I’m not going to get into it here). Until you have P/M fit, your only objective is likely to find that. Everything else is such a distant second that your process should probably be a little more organic and less structured on quarterly goals.

With that caveat, here’s how to manage with Objective-Key-Results. To illustrate, I’ll use a baseball team as an example startup:

1. Define, at a high level, what you want to accomplish over the next three months

This should be qualitative and narrative based. If you’re in between financing rounds, your goals should help you get to where you need to be in order to raise your next round. I always like to look at runway, subtract six months to raise and decide what the company should look like at that point. How many quarters you have until that milestone date should inform this process. For our baseball team startup, let’s define this as ‘build a world class team with a rabid fan-base and huge profits’.

2. Define three high-level objectives that support your goals

Some people recommend that these objectives are narrative based, e.g. ‘improve our first-time user retention’, or ‘surprise and delight our customers’. We tried that at first but we ended up making our objectives measurable outcomes that required results from multiple teams. Coming back to our team analogy, one of our objectives might be ‘win the World Series’. Winning the World Series is measurable — it’s a binary outcome, but it requires results from multiple teams: pitching, batting, coaching and owners.

3. Define three measurable results that support each objective

These should be very specifically measurable and, to the extent that’s possible, owned mostly by a single team. They should also be results of actions, not action items. Lastly, they should directly support the objective. An example result that would support winning the World Series might be ‘sign a five-tool player with an OBP over .325′ or ‘get a team-wide OBP of .250′. These would be a result of scouting, budgeting, deal making and finally signing a player or set of players who can get you the result you’re looking for.

Make sense?

A few other notes:

- get team input and buy-in. The important work in defining OKRs is the process of defining and prioritizing objectives. Handing these down from above doesn’t include the team in the processes, and if there’s no buy in then no one feels accountable for the results.

- these should be stretch objectives. You should expect that you’ll hit 60-70% of these.

- notice the use of three (three months, three objectives, three results-per objective). Stick to this. If you have three objectives and three results for each, you’re basically asking people to internalize twelve points. This is hard enough. If you start throwing a few bonus results in there the process will become unwieldily and ineffective. Keep it simple.

- try to create objectives that will actually drive your business. This sounds obvious, but on our first shot at this we set vanity metrics as goals. The problem with vanity metrics isn’t so much that they don’t drive your business, but they really don’t tend to inspire your team. As an example, try to focus on a metric that supports both growth and engagement — returning users might be one. It requires both growth and retention, so it’s both more challenging and more valuable to hit.

Have fun and happy managing!