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

Four dimensions and how they are weighted

Dimension	Weight	What gets scanned	Common false positive
Direct identity data	35%	Phone numbers, email addresses, ID numbers, street addresses	Reading a product model number as a phone number
Location and movement	25%	Geotags, city and venue names, itinerary posts, timezone clues	Treating one travel check-in as a home address leak
Sensitive topics	25%	Health, finances, workplace disputes, relationships, stated positions	Treating a review-style complaint as a real stance
Linkability	15%	Reused handles, avatars, profile links, distinctive phrasing	Treating a generic handle as strong evidence of a link

The four dimensions add up to a hundred. The first two take sixty points, and the reason is practical: a callable phone number creates real-world trouble faster than a stated opinion, and location data lets someone map your posts onto a person and a place.

Linkability sits last but is not negligible. It measures whether someone can decide two accounts belong to the same person. That risk does not cause immediate harm, but it amplifies the other three. One post in isolation is unremarkable; the same post next to an account under your real name is a different situation.

Why identity data sits in the top band

Phone numbers and email addresses can be used directly. Someone who has your number can call it, trigger verification codes, and walk through a password reset on another platform without you being present. Email is trickier, since it is often both the login name and the recovery channel for an account.

ID numbers, street addresses and photos of shipping labels belong in the same band. They appear far less often than phone numbers, but when they do appear the exposure from a single post is high. Scoring models generally do not average across items; they take the highest-risk entry in a category as the main input.

That explains a counterintuitive result. Clearing three hundred daily posts that mention a city may move the score by one point. Removing a single post containing a full phone number can move it by more than ten. Weights apply to categories, not to counts.

Why the same archive scores differently

Comparing how several tools describe their scoring, the differences cluster in three places. The first is how location is handled: some count any city name, while others only count posts with coordinate tags, and the first approach produces a noticeably longer list of hits. The second is the granularity of topic detection, where keyword matching files a post about working until you collapse under health, while semantic matching may pass it.

The third is the design of the baseline. Some models start at a hundred and subtract, others start at zero and add. The first reads high when your archive is small; the second reads low when your posting history is thin. Before comparing two reports, confirm which approach each one uses, or the comparison carries no information.

One more difference is easy to miss: the time window. Scoring the last year against scoring a full decade produces clearly different results. Reports usually state the range they scanned, so check that line before reading the number.

Working backwards from the weights to a plan

1. Start with the identity band. It usually holds the fewest items and delivers the largest score movement, which makes it the best return on effort.
2. Then handle repeated locations. Pull out the home address, the gym you visit weekly, the check-ins near a child's school. One-off travel posts can wait.
3. Split the sensitive topics band into stated positions and venting. The first group needs action; the second depends on your own tolerance. This judgement depends on your situation, and no tool can make it for you.
4. Finish with linkability, but do not skip it. Aligning handles and avatars across platforms is often less work than deleting dozens of posts, and the effect lasts longer.

Work in that order and the first two bands will move most people into a steadier range. Everything after that is maintenance rather than a one-time cleanup. For pacing, see a cleanup schedule that holds up.

Three things the weights do not measure

First, they do not measure legal risk. The same stated position carries very different consequences across jurisdictions, and a scoring model cannot reason about the rules where you live.

Second, they do not measure interpersonal risk. A post that names a colleague may score low while being the one item that actually causes trouble. That risk needs your knowledge of your own workplace, which the tool does not have.

Third, they do not measure future risk. Content that reads as low risk today can become high risk after a job change, a move, or a shift in public status. A score explains your archive as it stands; it is not a guarantee going forward.

Read the number as a sort order rather than a verdict. It tells you where to look first, not where you are allowed to stop. For what the bands mean, see what your footprint health score means.

Reading a score against its own baseline

Comparing your score to someone else's tells you almost nothing useful. Archives differ in length, in what a person posts about, and in how many years of history they cover, so a 62 and an 81 do not describe the same set of choices. What does carry information is your own number over time.

Record the score and the flagged count for each band on the day you run a check. Come back after your next round of cleanup and run it again. If the identity band dropped from nine items to two while the total moved by four points, the total understates what you achieved, because the category carrying the most weight is now nearly clear.

Two runs on different days also expose drift that has nothing to do with your actions. Post a location-tagged photo between the two checks and the location band will pick it up. Reading the two reports side by side tells you which change came from your cleaning and which came from new activity.

A one-week review you can actually finish

If you would rather not turn this into an ongoing project, run it across a single week. Day one: get a baseline report and record the flagged count in each band. Days two and three: work through identity data. Day four: handle the repeated locations. Day five: align handles and avatars across platforms. Day six: re-run the check and compare the two band counts. Day seven: store the deletion log and an archive snapshot, following how to snapshot before you clean.

The point of the sequence is ending up with two comparable reports. A single score on its own says little; the difference between the two is what shows whether your actions did anything.

About Digital Footprint Health

Digital Footprint Health (digital-footprint-health.shop) lays the four dimensions out in the open. You upload your X data archive, it parses every tweet on your own device, lists flagged items by category and produces a score from 0 to 100. The check is free and read-only, and it never asks for account access. Once you can see which items move the number, you decide whether anything gets deleted; scope and pricing are on the pricing page. You can run a check from the homepage, and the rest of the scoring material lives on the blog.