A Top Voice score is a number out of 100 that says how much a LinkedIn account is actually read. It is ours. We compute it from what the profile and the recent posts show, and it exists because follower counts on their own are a poor guide to whether answering somebody will be seen by anyone.
The four parts
- reach — how many people follow them. Weighted 0.35, with full marks at 100,000.
- resonance — what one of their posts earns in reactions. Weighted 0.35, full marks at 1,000 a post.
- conversation — how much of that is comments rather than taps. Weighted 0.15.
- cadence — how regularly they publish, measured over the span the posts themselves cover. Weighted 0.15, full marks at five a week.
The parts always travel with the score, and so do the figures each was computed from: followers, posts read, reactions, comments, reposts, reactions per post, posts per week and the number of days those posts spanned. You can take the number apart.
What it is not
It is not LinkedIn's Top Voice badge. That badge is LinkedIn's, awarded by LinkedIn, and we report it separately in a field called badge. The field has three values and the third one matters: true and false are what the profile said, and null means the profile did not say — never that there is no badge.
A score is a measure of a moment. Someone measured the week they went viral scores higher than they will next month, which is why the measurement time comes back with it.
Questions people ask
Can I see where my own account stands?
Yes. GET /linkedin/topvoices/standing measures your account on the same four parts as the people you watch, and answers the share of that field you are above on each part, the middle of it to aim at, and the single part you are furthest behind on.
Why is the score different from the one I saw last week?
Because it is read fresh from what the account has published since. A quiet fortnight lowers cadence; one post that travels raises resonance. Watching somebody stores the measurement, and the campaign refreshes it as it works through the list.
Can an AI agent read these scores?
Yes. Every one of these calls is an MCP tool, so an assistant connected to the MakeHybrid API, or your own agent in a project room, can ask who the voices on a topic are and what they score.