What OKRs are for
OKRs are not a task list. They're a focus and alignment tool. A well-written OKR tells every team member: here's what matters this quarter, here's how we'll know if we succeeded, and here's how your work connects to the company's direction. If an OKR doesn't change what the team works on, it's not doing its job.
Writing good OKRs
Objective: qualitative, inspiring, directional. 'Become the go-to platform for cross-chain settlement in Japan.' Key Result: quantitative, binary at the end of the quarter (achieved / not achieved), and leading (within team control). Bad KR: 'Improve user satisfaction.' Good KR: 'Increase 30-day activation rate from 42% to 60% by end of Q3.'
The 70% rule
OKRs should be set at a level where 70% achievement is a success. If a team consistently hits 100%, the goals are too conservative. If they consistently hit 40%, the goals are either too ambitious or team capacity is the real problem. I review OKR score trends over time - not just the number, but the direction.
Connecting team to company
I run a quarterly OKR cascade: company OKRs first, then department, then team. Each team OKR must map to at least one company-level Key Result. Any team OKR that can't map to the company level is a signal that either the work shouldn't be done this quarter, or the company OKRs are missing something important.
Real example
At SmartNews, I introduced OKRs across the Ads product team as part of a broader PDLC rollout. We defined 3 company-level OKRs tied to Ads revenue and set cascaded team OKRs for each of the 5 sub-teams. Weekly check-ins replaced monthly reporting - teams could self-correct within the quarter rather than discovering misses at the end.