How Your 0-100 Health Score Is Calculated in 1 Minute

--
canonical_url: https://digital-footprint-health.shop/blog/health-score-calculated-1min
title: How Your 0-100 Health Score Is Calculated in 1 Minute
tags:

• twitter
• privacy

- security

Everybody reads the number first. 79. Then the question: is that good? Why does my friend have 91 and I have 79? The short answer is that the score is a weighted sum of four dimensions, mapped onto 0-100. Here's each one, roughly in weight order.

Share of risky tweets. Risky tweets divided by total tweets. Heaviest component by far. A 20,000-tweet account with 200 high-risk posts gets dragged down more by this number than by anything else, because it's a ratio over a big base.

Category weight. Not all risk is equal. A tweet tagged phone, email, or identity costs more than one tagged sensitive. The more directly exploitable the category, the heavier the single-tweet penalty. This is the difference between "a number that reaches you" and "a joke that might embarrass you."

Time decay. A phone number tweeted in 2016 is less dangerous than the same tweet from 2024, because the number has probably been deactivated. Older tweets lose fewer points. Not zero, since screenshots stick around in other people's hands, but lighter. This is computed from the post's publish year and stays fixed until you rerun the report.

Quantity effect. Nonlinear. One phone number tweet versus ten is not a tenfold difference. It's the gap between "an accident" and "a habit," and the marginal harm grows as the count climbs.

Those four get weighted, summed, and mapped to 0-100. Bands in practice: 90-100 is clean, 70-89 is mostly healthy with scattered exposure, 40-69 is risk building up, under 40 is urgent. The bands are observational, not a precise scale, but the direction holds. Below 40 means P0 and P1 tweets were never cleaned.

The point I keep repeating: low score doesn't mean delete everything. Emotional-value tweets aren't on the penalty list unless they carry a label. A low score made entirely of P3 embarrassment is cringe but not dangerous. And some exposure you never controlled, screenshots already out there, deleting your account is just the first step of damage control.

A worked example makes the weights concrete. Take a decade-old account with 200 risky posts out of 20,000. The share term is small, 1 percent, but it sits over a huge base, so it still drags the score noticeably. Within those 200, one tweet carries an identity label from last year and costs more than twenty sensitive-labeled posts from 2015, because category weight beats both time decay and sheer count on a per-tweet basis. And a dozen old location check-ins barely move the number, which is why someone can have a low score while their actual exposure is concentrated in a handful of recent posts.

The bands make it actionable. Above 90 you're mostly maintaining. In the 70s and 80s you have scattered exposure worth clearing P1 for. The 40s and 60s mean years of accumulated risk, start at P0. Under 40, contact details and identity are likely out there, handle the top two tiers now. The bands are observational rather than a precise scale, but the direction holds, and a rerun after a tier cleanup is the fastest way to see the model respond to actual behavior.

One misconception worth clearing up: 100 doesn't mean permanently safe. It means the report found no recognizable risk in the current archive. Post a phone number tomorrow and a rerun drops the score. That's the property that makes the score honest, it tracks current exposure, not reputation, so it's best read as a before-and-after measurement around cleanups rather than a standing grade.

Practical use: remove an identity-level tweet and the number moves more than it would for ten selfies. Clear one tier at a time, rerun, watch the score recover. Scores aren't comparable across tools, but your own number over time is a useful signal.

The dimension table and FAQ: https://digital-footprint-health.shop/blog/health-score-calculated-1min