How the Digital Footprint Score Is Weighted: Inside the 0-100 Model

--
title: "How the Digital Footprint Score Is Weighted: Inside the 0-100 Model"
description: "Why the same X archive scores differently across tools, and how four weighted dimensions decide which items actually move your number."
tags: ["privacy", "digitalfootprint", "security", "twitter"]

canonical_url: https://digital-footprint-health.shop/blog/footprint-score-weighting-explained

Footprint scores are not counting your bad posts. They are measuring how much exposure a small number of weighted categories carries, which is a narrower question than most people assume they are asking.

That distinction explains nearly every confusing result people report about these tools. It is also why two products can read the same archive and disagree by twenty points without either being wrong.

1. The four bands, and why two of them dominate

Four dimensions carry the score, and the two heaviest are the ones you can act on directly: direct identity data at 35% and location at 25%.

Identity data takes the largest share because a phone number or an email address can be used without you present. Someone holding your number can call it, trigger a verification code, or start a password reset elsewhere. Since email is often both the login name and the recovery channel, it is worse than it looks.

ID numbers, street addresses and photos of shipping labels sit in the same band. They show up far less often, and when they do, one post carries most of the exposure.

2. Weight sits on categories, not on post counts

Weights apply to categories rather than item counts, which is why clearing three hundred ordinary posts can move the number by a single point.

This is the part that surprises people. Scoring models take the highest-risk entry in a category as the main input rather than averaging across entries. A category with one flagged item therefore behaves almost identically to a category with forty.

Practical consequence: if your plan is to reduce the number, deleting low-risk volume is the least efficient route available to you.

3. Why the same archive scores differently in two tools

The same archive scores differently across tools because of three design choices: how location is counted, how topic detection is granularised, and whether the baseline starts at zero or at a hundred.

Location handling is the biggest source of cross-tool disagreement. Some tools count any city name in the text, others only count posts that carry coordinate tags. The first approach produces a hit list several times longer on the same archive, which changes both your score and your reading of how exposed you are.

Baseline design matters the same way. A model that starts at a hundred and subtracts reads high on small archives. One that starts at zero and adds reads low on a thin posting history. Two reports are only comparable if you know which design each one uses.

4. Turning the weights into a work order

Reading the weights backwards produces a work order: identity band first, repeated locations second, sensitive topics third, linkability last.

Start with the identity band. It usually holds the fewest items and produces the largest movement, so it has the best return on effort of anything on the list.

Then handle locations that repeat. A home address mentioned three times, a gym you check into weekly, posts near a child's school. One-off travel posts can wait, because a single check-in far from home is not the same signal as a pattern.

Sensitive topics need a split between stated positions and ordinary venting. The first group usually needs action; the second depends on your own tolerance, and no tool can make that call for you. Linkability goes last but should not be skipped, since aligning handles and avatars across platforms is often less work than deleting dozens of posts and the effect lasts longer.

5. What the number does not tell you

Three things the model does not measure at all: legal risk, interpersonal risk, and future risk.

Legal risk depends on jurisdiction, and a scoring model cannot reason about the rules where you live. Two people can post the same sentence with very different consequences.

Interpersonal risk is invisible to the model. A post that names a colleague may score low while being the only item that actually causes trouble, because the tool has no access to your workplace relationships.

Future risk cannot be measured at all. Content that reads as low risk today can become high risk after a job change or a move. Treat the number as a sort order, not a verdict.

Practical takeaways

• Four dimensions carry the score, and the two heaviest are the ones you can act on directly
• Weights apply to categories rather than item counts, which is why clearing three hundred ordinary posts can move the number by a single point.
• The same archive scores differently across tools because of three design choices
• Reading the weights backwards produces a work order
• Three things the model does not measure at all

If you take one thing away from this, make it the flagged count per band rather than the total. A total that moves four points while the identity band drops from nine items to two is a far better outcome than the headline number suggests.

The longer version with the reference detail is here: https://digital-footprint-health.shop/blog/footprint-score-weighting-explained