There is a particular kind of product number that nobody can do anything with. It arrives alone, somewhere between 0 and 100, and the only two possible responses are to believe it or not. Influence scores are usually that number.
So ours never travels alone. Every answer that carries a score carries the four parts it was built from and the raw figures each part was computed from, and the weights are published rather than tuned in private.
What that buys
It buys disagreement, which is the point. If somebody with 200,000 followers scores below somebody with 12,000, the parts say why: the big account posts twice a quarter and gets taps, the small one posts twice a week and gets arguments. You may think the weights are wrong. Now you can say so about something specific.
It also makes the number useful for the thing people actually want, which is not a league table but a decision: who is worth answering this week. A profile that is big and silent is a poor place to leave a comment. The parts tell you that; the score alone does not.
The third value
LinkedIn's own Top Voice badge is reported next to our score, and the field has three values rather than two. True and false are what the profile said. Null means the profile did not say.
Collapsing that third value into false would be convenient and would be a lie about four times a week. Data we did not receive is not data that says no, and a field that cannot tell the difference will eventually tell somebody that a Top Voice is not one.
The same rule runs through the rest of this API. An account with no status is not an account that failed, and a post with no timestamp is treated as new rather than quietly skipped.
Questions people ask
Can I weight the parts differently?
Not in the score itself, which is one published formula so that two people comparing notes are comparing the same thing. The parts come back separately, so sorting by the one you care about takes a line of your own code.
Does a high score mean a comment will be seen?
It means more people are reading that account than are reading a low-scoring one. Whether your comment is read depends on how quickly it arrives and whether it says anything, which is why the campaign only answers posts that are still new.
Where does the measurement come from?
From the profile and the recent posts, read as one of your own connected accounts. Nothing is bought from a third party and nothing is inferred from data we were not shown.