Key Result

The easiest sentence to write when setting a quarterly goal is something like "improve the user experience." Everyone nods, nobody objects, and it takes under five minutes. The trouble starts when the quarter ends.

Pagedeck's first quarter went exactly that way. The goal read "improve the onboarding experience," and when Grace asked three months later whether it had happened, the answers split. Noa had redrawn the first screen and rewritten every piece of guidance copy, so yes. Mia pointed out that the share of teams creating a document within their first week hadn't moved, so no. Both were right by their own standard, because no standard had been written down at the start.

A sentence describing a direction always has this hole in it. It tells people which way to go and nothing about whether they arrived. So a few checkable items get attached underneath, and at the end of the quarter only those items are read to make the call. Numbers agreed on in week one stand in for the argument in week thirteen.

Two or three key results hang under a single objective. With one, gaming that number quietly becomes the goal; past five, nobody can tell which one matters.

Four rules govern how they're written.

1. Put the starting value, the target value, and the deadline in the same line. Not "raise the paid conversion rate" but "raise paid conversion from 4 percent to 7 percent this quarter." Without the starting value there's no way to judge the ending one.
2. Write the change the work produced, not the work. "Onboarding redesign shipped" is not a key result. The redesign can ship and conversion can sit exactly where it was, leaving a changed screen and an unmet goal. The redesign is what you do to reach the result; it isn't the result.
3. Make the two or three items look at different faces of the same thing. Conversion alone can be raised by locking features out of the free plan. Pair it with "keep three-month retention of converted teams above 80 percent" and that shortcut closes. A quantity item next to a quality item is the usual pairing.
4. Pin down what gets counted and how, in advance. Whether the denominator of conversion is every team that signed up or only teams that opened the product can put a factor of two between two readings of the same quarter.

Scoring usually runs on a scale from 0 to 1.0. A target of 4 to 7 percent that ended at 6.1 scores 0.7. A team posting a column of 1.0s reads as a team that set easy targets rather than a team that did well, and 0.6 to 0.8 is the normal band. That reading only holds if the number was set as a stretch goal in the first place.

The most common home is OKR, the practice of hanging two or three key results under one objective, running it by the quarter, and keeping what every team wrote visible to everyone. Key results are the back half of that format. The sentence supplies direction; the items supply the verdict.

They also get read throughout the quarter, not just at the end. Pagedeck updates the current value of each item every Monday. A conversion rate meant to go from 4 to 7 that reads 4.3 in week six won't get there at that pace, so the team decides right then whether to change the plan or adjust the target. Opening the document for the first time in week thirteen leaves no time to do either.

They act as the joint in cascading goals, where company goals are carried down to teams and individuals. If the company key result is paid conversion, engineering might take payment failure rate and design might take first-week activation. Each step down drifts further from the original number, so many companies ask teams to write one line naming which item above theirs their item moves.

Teams that use SMART goals, the five-part test asking whether a goal is specific, measurable, achievable, relevant, and time-bound, tend to apply the same test here, since those questions are the ones a key result has to survive anyway.

"That's a task, not a key result. What number moves when it's finished?"

A request to convert an activity into the outcome it's meant to produce.

"Item three is sitting at 0.4 and four weeks won't close that. Do we add people or land it at 0.6?"

Mid-quarter reallocation driven by where an item actually stands.

"Conversion alone can be gamed by locking free features. Put retention next to it."

Closing a shortcut with a paired item rather than with a policy.

Pagedeck's goal last quarter was one line: make teams on the free plan pay and stay. Three items hung under it — paid conversion from 4 to 7 percent, three-month retention of converted teams held above 80 percent, payment failure rate from 3 percent to under 1. Mia updated the first two weekly and Theo updated the third.

The quarter closed at 6.1 percent, 83 percent, and 0.9 percent. The first item scored 0.7, and Grace still read the quarter as a win. Seen together, the three numbers said conversion moved less than hoped while the teams that did convert stayed, and that told everyone what to work on next without further debate.

The most frequent mistake is using the slot as a to-do list. It looks like the box where you write what the quarter contains, so plenty of teams fill it with "ship feature A, ship feature B." Written that way, a quarter where all three shipped and nothing changed for users gets recorded as a success. After two or three of those, goal-setting turns into paperwork and nobody opens the document at quarter's end.

The other is wiring attainment straight into individual performance reviews. Once the score becomes the rating, people write numbers they're sure to clear. A team capable of going from 4 to 7 writes 4 to 5 and collects a 1.0, because that's the safe move. A year of that leaves every team with high attainment and the company's numbers flat. Many companies therefore use scores as material for a conversation about what was learned and keep them out of the compensation formula.