Skills Data, Explained
Why skills data goes stale (and what actually fixes it)
Every platform's skills data eventually goes stale. The difference is what happens next — whether it quietly rots, or someone has a reason to keep it current.
The problem: all skills data decays. The question is why.
Ask any HR or L&D leader how confident they are in their skills data, and you'll get a pause before the answer. That's not a data-entry problem. It's a design problem — most approaches don't give anyone a reason to keep the data honest after the moment it's collected.
- Test fatigue — a skills assessment is accurate for exactly as long as it takes to close the browser tab. Nobody retakes a test to update a score nobody's using.
- Document lag — a qualification matrix or org chart is a snapshot. The day a role changes, a team reorganizes, or someone picks up a new responsibility, the document is already behind.
- No reason to update — even systems that let people edit their own profile rarely get used, because there's nothing in it for the person doing the updating.
Three ways skills data gets collected — and why two of them decay
Every skills platform makes a choice about where the data comes from. That choice determines whether it stays accurate or starts decaying the moment it's collected.
The Workera approach
Verified / tested
A skills test produces a precise, defensible score — at the moment it's taken. But a score is a snapshot, not a subscription: nobody re-tests routinely, so the data is accurate exactly once and stale from then on.
The Cobrainer approach
Document-extracted
Pulling skills from qualification matrices, PDFs, and org charts gives you a governed starting point fast. But documents don't update themselves — the data is only as fresh as the last time someone remembered to re-upload.
The Gloat / Phenom / Cornerstone approach
Inferred / admin-entered
Skills inferred from resumes and job history, or entered by a manager, were never confirmed by the person they're supposed to describe — so there's no telling how close they are to reality on day one, let alone a year later.
The fix: give people a reason to keep their own data honest
Mentessa asks people to declare their own skills — what they have, and what they want to deploy next. The data stays accurate because it's the direct path to something they actually want: a better match with the right mentor, mentee, training, or internal role. Not a compliance checkbox. Not an assessment score. A reason that's actually theirs.
This doesn't replace verification or document ingestion where those make sense — Mentessa's own Skills Agent still builds a governed baseline from your existing documents. It's what keeps that baseline from going stale the moment it's built.
Where the evidence stands today
We're not going to hand you a headline stat we can't back up. The honest answer: Mentessa is built on the premise that self-declared, incentive-aligned data holds up better over time than data nobody has a reason to update — and we're actively tracking profile-update frequency and match-acceptance rates as our self-declaration model scales, so we can publish the real numbers rather than a plausible-sounding one. If you want the current picture for your organization's size and industry, ask us directly — we'll tell you what we actually know.
Questions?
Does self-declaration mean skills data is unverified?
Not necessarily. Mentessa's Skills Agent still builds a governed baseline from your existing documents — qualification matrices, PDFs, org charts. Self-declaration is what keeps that baseline from going stale after it's built, not a replacement for governance.
Isn't self-reported data just as unreliable as any other kind?
Self-reported data is unreliable when there's no reason to keep it accurate. Mentessa ties it directly to something the person wants — a better match — so keeping it honest is in their own interest, not an administrative chore.
How is this different from letting employees edit their own HR profile?
Most systems that let people edit their profile see almost no voluntary updates, because nothing changes for the person who bothers. Mentessa's matching only works as well as the declared skills, so accuracy is rewarded immediately, not eventually.
Does this replace testing or assessments entirely?
No — assessments are still useful for a point-in-time, defensible score. What they don't do on their own is stay current. Self-declaration is what keeps the picture accurate between assessments, or where a formal test isn't practical at scale.
See the mechanism in your own skills data
Talk to us about what a governed, self-declared skills architecture would look like inside your organization.