Methodology

How a phone becomes a reliability sensor.

Digital phenotyping models internal states — motivation, regularity, energy — from standard sensor signals, without ever prompting the user. Djob-Index applies it to a single question: will this candidate be reliable at work?

01

The worker installs the app

No test, no questionnaire. The analysis runs in the background, with explicit consent, from Day 0 — before the first shift.

02

Passive signals are measured

Sleep-cycle regularity, punctuality of movements, responsiveness to notifications, daily activity. Never the content of messages or calls.

03

The Index is calibrated in the field

Each signal is correlated with real ratings from warehouse and site managers on the Djob marketplace — our validation lab.

04

You check the score in one call

Integrated with your ATS or HCM, or via our portal. $15 per lookup when you fill the shift. The score belongs to the worker and follows them everywhere.

The four signal families

Time discipline

Stability of wake cycles, snooze frequency, location entropy.

Responsiveness

Response time to notifications on Day 0 — a proxy for engagement.

Energy

Daily walking speed, correlated with physical energy at work.

Proactivity

Completion of preparation micro-tasks and uptake of instructions.

What we never collect: the content of messages, calls, photos or browsing. The microphone is not used. Location is reduced to regularity measures, never stored as a route.

Versus the alternatives

Agencies
Active tests
Djob-Index
Method
Résumé, manual judgment
30–45 min tests / games
Passive phenotyping, Day 0
Candidate effort
Medium
High — 25–34% drop-out
None
Cost
50–100% of hourly rate in commission
Annual SaaS licence
$15 / call or $2 / month
Who owns the score
The agency
The vendor
The worker
Join the interest list Read the FAQ