OKRs are outcome-oriented and I’ve found that they work particularly well for product teams. There are, however, some cautions and updates from common practice required in a healthy product culture. I’d even say there are some things venture capitalist John Doerr got wrong in his widely read book on OKRs, Measure What Matters.
Objectives and Key Results (OKRs) are a simple system for aligning teams around common objectives. Invented by Intel, made famous by Google and Zynga, OKRs are a huge phenomenon among tech companies and, increasingly, in any organization that wants to operate like one.
OKRs in a Nutshell
Objective: Get real traction for our app
Key Result 1: Increase paying customers to x per month
Key Result 2: Increase monthly active users (MAU) by y%
Key Result 3: Increase revenue to $z MRR
A team (or a whole organization) can have 3-5 top-level OKRs that together reflect the most important outcomes the company is pursuing. In larger organizations, OKRs may “cascade” from these top few down to child OKRs that divide up the focus across teams and levels of hierarchy. Key Result 1 from the above Objective could become a Parent Objective with its own KR children below.
Objective: Increase paying customers to x per month
Key Result 1: Increase self-sign-up to x per month
Key Result 2: Increase free-trial conversion by y%
Key Result 3: Reduce churn to < z% per month
In theory there is no limit to the size of this tree. I say in theory because the most common rookie mistake I see in adopting OKRs is making the forest of interlocking and interdependent goals too dense for anyone to understand. When this happens, the tool that was intended to create alignment in the organization actually undermines it because everyone is focused only on their tiny twig.
A lot has been written about OKRs generally. This series of articles will focus on using OKRs for product teams and product-driven organizations. The principles and examples come from lessons learned in implementing OKRs in product teams and product-led organizations around the world.
I plan to discuss:
Why OKRs Should Start With Why
Outcomes, Not Outputs: Don’t Focus on How
Measure Twice, then Measure Again
Commit Hard but Recalibrate When Needed
Forget Cascading, Build Teams
OKRs Fit Neatly With Agile
What NOT to Do with OKRs
Let’s start with why. No irony intended.
Why OKRs Should Start With Why
Why do so many have a hard time believing Mark Zuckerberg when he says that at Facebook “the future is private?” Maybe it’s because it has been glaringly obvious for some time that Facebook’s mission is, in fact, to sell people’s attention to advertisers. A couple of years ago, the company updated its mission statement to be more empowering of users, but this too seemed more like a reaction to public backlash than an intrinsic motivation.
Facebook could be in the dictionary as an example of a successful growth company, but I’d argue they could have avoided the last few years of scandals with a clear focus on the value they wanted to create in the world (and if management were held to that, of course). The company uses OKRs internally and is well-known for focusing exclusively on measuring engagement. Perhaps, though, they should have stopped to ask themselves ‘why are we doing this?’ What good do we hope will result for our users? Will we improve their lives in some way? Will they feel more supported? More connected with friends and family? Better informed? Happier?
Perhaps this sounds less like business and more like ethics, but motivation -- why you do things -- is important to customers and to employees. As business guru Simon Sinek says, “People don’t buy what you do, they buy why you do it.” Engagement strikes me as a very internally-focused metric. It’s about users paying attention to our product. What they get out of that engagement seems entirely left out of the equation. So the clear-eyed reason for focusing on user benefit is the realization that eventually people figure out when they are being used.
Product OKRs Should Start with the Customer
Asking what is our product vision or team mission is also really helpful as a starting place for OKRs. When asked to write objectives, lots of executives default to “grow by x%” or “generate y revenue.” Then they generally run out of ideas for objectives or start quickly thinking in terms of deliverables like “ship the new thingie by Q1” or “launch the whackamole initiative.”
If, instead, you have a clear product vision focused on your customers like “help [type of customer] [improve in some way]” then you have a much more robust place to start in creating goals. Sure, you probably still have some growth metrics in mind, but you could start by measuring the improvement you are seeking in customers’ lives. Measuring the value you are generating as directly as you can is a great way to align your team on what matters.
Let’s test how Facebook’s new mission statement might work here. "To give people the power to build community and bring the world closer together" starts with a bang but kind of ends with a whimper. ‘Closeness’ seems more feel-good than do-good, but you could probably find ways to measure ‘building community.’ Dictionary.com’s second, less geographic definition of community is “a feeling of fellowship with others, as a result of sharing common attitudes, interests, and goals.”
Facebook already measures the number of ‘friends’ you have, but if they want to focus on building community they might consider focusing on the number of two-way interactions people have with those friends or even the feelings those interactions generate.
Social media is often cited as a cause of social isolation and mental health issues. Facebook has been shown to undermine life satisfaction. Imagine if some sort of happiness metrics became OKRs at the company and they actively sought to tune their algorithms to achieve it for their users. Maybe social media could become a virtuous cycle that drove healthy outcomes for users, advertisers, and the company as well.
Here is a hypothetical revision of Facebook’s mission statement and one key OKR tree to support it.
Mission: To give people the power to build community and make the world a happier place
Objective A: Help users support each other
Key Result 1: Increase the number of two-way message exchanges per user by x
Key Result 2: Increase the percentage of messages with positive sentiment by y%
Key Result 3: Increase self-reported happiness levels of engaged users by z points
There could also be OKRs about forging new connections, length of exchanges, asking for and receiving favors, or even fostering real-time or real-life interactions.
None of this would mean they could not also have OKRs around revenue and user growth as well. A healthy business looks out for all of its stakeholders (users, buyers, partners, employees, and the community as well).
Product Vision Is Your Highest Objective
To summarize, OKRs can quickly drive you to a narrow fixation on business results, unethical behavior, and short-term outcomes that eventually undermine your very business. But if your OKRs begin with an inspiring vision of how awesome it will be for your customers when you succeed, then you can break that down into whatever is required to achieve that vision or carry out that mission.
In fact, you could view a good mission statement or product vision as your highest-level, longest-term Objective. Its KRs would then cascade down and form the top-level Objectives for your team. As a generic framework, OKRs can really be about any goals you wish to pursue. The most successful product teams, though, start with the customer and their success. They start with a compelling why.
For examples and an in-depth explanation of product vision statements, check out chapter 4 of my book, Product Roadmaps Relaunched.
In the next installment, we will discuss John Doerr’s biggest f*** up in Measure What Matters (which is worth reading in any case), why outcomes are more important to align on than outputs, and when it is OK to add tasks to OKRs.
Facebook Icon made by Freepik from www.flaticon.com