Value
What the price buys: how much is behind it, and whether the pricing holds up over time rather than swinging around a permanent sale.
How the LUVS Score works
Six public signals, blended into one number from 0 to 10, the same way for every creator.
6 public signalsRecomputed live
Example read Must subscribe
LUVS score8.9/10
0 to 10 scale
- Skip0-3.9
- Only if it is cheap4-5.4
- Only on a deal5.5-6.9
- Worth it7-8.4
- Must subscribe8.5-10
The inputs
The six signals, in plain words
Every creator gets one number from 0 to 10. It is a weighted blend of six public signals, scored the same way for everyone and recomputed as the data moves.
Reactions
How much a creator interacts with fans rather than only broadcasting at them.
Output
How much content exists and whether it keeps coming, so a live catalogue reads differently from a stale one.
Active
How alive the profile is right now. Recency of posting and last seen signals. A creator who vanished scores lower.
Steady
Whether posting is steady or bursty. A reliable cadence beats a flood followed by weeks of silence.
Variety
Breadth and richness of the catalogue. Media variety, length, and the substance behind the numbers.
From data to number
How it is built
- 1
Gather
Pull the public signals: profiles, counts, pricing, activity, socials. Nothing from behind a paywall.
- 2
Normalize
Put every signal on the same scale, judged against the whole pool, so a number means the same thing for everyone.
- 3
Blend
Combine the signals into one 0 to 10 score with a fixed weighting, identical for every creator, applied automatically.
- 4
Refresh
Recompute as new public data lands. The number tracks the creator today, not a one off snapshot.
The index
How the price index is computed
Each month we average what a subscription actually costs across every creator we publish, after discounts, using the last price observed rather than an average of the month. Once a month closes it is computed once and never revised.
The one thing we keep private
We show you the ingredients, not the exact recipe.
You have just seen everything that goes in: six public signals, blended into one number, the same way for every creator. What we keep to ourselves is the precise weighting, and that is on purpose. A score you can reverse engineer is a score you can game: farm a metric, buy a placement, juice a number. Ours cannot be. Creators cannot pay for it, and they cannot fake their way to it. Keeping the exact formula private is what keeps the number honest.
The promise
What the score is, and what it isn’t
A comparison aid
A fast read on whether a profile is active, fairly priced and worth a look.
Not an endorsement
No moral judgement and no claim about a person. Only a read of public signals.
Not for sale
Creators cannot pay to rank, and referral links never move the number.
Honest about the edges
Known limitations
Every dataset has edges. Naming ours is not a disclaimer, it is part of the method: a number you can trust is one whose limits you were told about.
Partial market coverage
We publish a curated catalogue, not the entire platform. Nothing here is a census: a figure describes the creators listed on LuvsOne, and the catalogue grows as we publish more.
Public signals only
Everything is what a logged-out visitor can see. Nothing is taken from behind a paywall, which also means private engagement is invisible to us and no amount of it can raise or lower a score.
An update lag
Profiles are re-checked on a rota rather than continuously, so a number can trail reality by days. Recency is itself one of the signals, which is why a profile that went quiet reads as quiet.
Labels are current, not historical
Niche and country come from the labels a creator carries today, applied to past months as well. They are stable attributes, but a creator who changed direction is described by where she is now.
The number answers to the data. You can too.
Creators can claim their profile to correct the facts, or just reach out. We read every message. The score itself stays the same: claiming fixes the record, it never buys a better number.