Workforce Intelligence from Public Profile Data
By Kooperativa Engineering
Workforce intelligence tooling is usually sold to HR and talent teams as a view into their own organisation, built from HRIS data they already own. The more interesting and less served case is external: understanding the shape of a workforce you have no internal access to, a competitor, an acquisition target, a market you are considering hiring into.
Public professional profile data is the only practical source for that, and it supports a narrower set of questions than the category marketing implies. Being precise about which questions is the difference between a defensible analysis and a confidently wrong one.
What it answers well: structure
The strongest use of external profile data is organisational shape rather than headcount. A single employee number is nearly useless for comparison, since two companies at the same size can be structured completely differently. A seniority breakdown turns that number into something closer to an org chart.
{
"company_id": "2135371",
"total_indexed": 2847,
"breakdown": {
"c-level": 38,
"vp": 6,
"director": 28,
"manager": 432,
"individual": 2343
}
}Run that across a competitive set and the comparison becomes concrete: manager-to-IC ratio, how many layers sit between a VP and an individual contributor, whether a company is top-heavy relative to its peers. Those are structural facts about how an organisation is built, and they are visible from outside without any internal access.
What it answers well: direction
The second thing external data does well is movement over time. Hiring activity, filtered by seniority, distinguishes between two situations that look identical as a headcount trend line: a company adding individual contributors is building capacity, a company adding directors and VPs is building leadership ahead of a broader push.
The same applies in reverse. A cluster of departures from one function, visible through past-employer records, is a signal that no internal dashboard outside that company would surface.
curl "https://kooperativa.io/api/v1/company/hiring-signals?company_id=2135371&days=30" \
-H "Authorization: Bearer kk_live_..."The three questions it cannot answer
Each of these gets attempted with profile data and each produces a number that looks credible and is not.
- Compensation. Nothing in a public profile carries pay data. Any salary figure attached to external workforce analysis is modelled from title and location benchmarks, which is a legitimate technique and a completely different kind of claim from a measured one.
- Actual attrition rate. You can observe that people left. You cannot see the denominator reliably enough to compute a rate, because indexed count is not true headcount and departures surface at whatever lag the dataset re-indexes at.
- Current skills for a current role. Self-listed skills are cumulative across an entire career, not scoped to a present job. A profile still lists tooling from a role a decade ago, which is why generic skills dominate any frequency count and why "what does this team actually work with today" is not answerable this way.
The indexed-count caveat, in the one place it matters most
Every figure above is derived from profiles present in a dataset, not from an audited employee roster. For workforce intelligence specifically this caveat carries more weight than it does in a sales context, because the analysis is usually comparative, and indexing coverage is not uniform across companies.
A large, digitally visible tech company will be indexed far more completely than a similarly sized manufacturing business, which means a direct headcount comparison between the two is measuring visibility as much as size. Ratios within one company are considerably more robust than absolute counts across companies, and structuring the analysis around ratios is how you avoid the trap.
Keeping it current without a polling job
Workforce analysis has an unusual freshness profile: the underlying structure changes slowly, but the individual events that matter, a key departure, a leadership hire, happen at unpredictable moments. Polling on a schedule is a poor fit for that shape, since a monthly job is either too slow for the events or too expensive for the structure.
Subscribing to specific companies or people and receiving a signed webhook when a tracked field changes fits it better. A monitor on a set of competitor companies for `company.staff_changed`, plus monitors on named senior people for `person.job_changed`, covers the event side without re-checking records that have not moved.
Get started
Try Kooperativa
One API key. Person and company enrichment, structured search, and monitors under one flat license.