Baseline Key Result
A baseline Key Result is used when you genuinely cannot set a credible target yet: this cycle you measure to establish the number, and next cycle you set a real from-X-to-Y target. It is a legitimate move exactly once per metric, and a warning sign if it recurs.
When it is the right answer
You are measuring something for the first time, or the metric has just been redefined, and any target you set would be invented. Picking a number out of the air and defending it for three months helps nobody.
In that situation, "establish the baseline for activation rate and agree a target for next quarter" is an honest, checkable Key Result with a clear definition of done.
What stops it becoming an excuse
A baseline Key Result has an expiry. It runs for one cycle, and the deliverable is a number plus an agreed target, not a dashboard.
If the same metric needs a baseline twice, the problem is not measurement. It is that nobody owns the metric, and that is a different conversation to have.
How to write one
Name the metric precisely, including how it is calculated, because half the value of a baseline cycle is discovering that three people define it differently.
State what you will do with the number: "and agree a target with the leadership team by week eleven". Without that clause it is a measurement task, and it will quietly become permanent.
What is a baseline Key Result? +
A Key Result used when no credible target can be set yet: you measure to establish the number this cycle, then set a real from-X-to-Y target next cycle.
Can you have a Key Result with no target? +
For one cycle, yes, if the honest position is that any target would be invented. It should come with a commitment to produce the number and an agreed target by a stated date.