Skip to content
Jobsearch.ing

Data

Job-Search Data and Benchmarks

Benchmarks are only useful when you know what they counted. Every figure here is published with its definition, population, period, method, and limits attached.

External benchmarks

Figures published by someone else

Government statistics, official datasets, academic studies, and industry reports. Jobsearch.ing selects them, states what they actually measured, and shows where credible sources disagree. The value added is interpretation and provenance — never a new number invented by averaging old ones.

Original measurement

Figures Jobsearch.ing produced itself

Data collected or analysed here, published with its full methodology and a citable version. There is none of this yet, and nothing will be labelled original research unless it genuinely is.

Research standards

Publication status

No data pages are published yet.

This is the section where care matters most. Job-search statistics circulate widely with their origins stripped off: a figure measured on one platform’s users in one year becomes “the average” everywhere, and a survey of a few hundred people becomes a law of hiring.

A figure appears here only when every field below can be filled honestly. Where two credible sources disagree, both are shown with the reason for the gap. Nothing is resolved into a single tidy number that no source supports.

How sources are selected and weighed

Required on every data page

Eight fields, none optional.

These are enforced in the codebase: a data page missing its population, geography, time period, methodology, limitations, or sources fails the build rather than shipping incomplete.

Definition
What the metric counts, in words precise enough that two people would count the same thing.
Population
Whose applications, searches, or outcomes the figure describes — and, just as importantly, whose it does not.
Geography
The labour market the figure applies to.
Time period
When the data was collected. A hiring statistic from a different market cycle is a historical fact, not a current benchmark.
Methodology
How the number was produced: sampling, instrument, response rate, and any adjustment applied.
Sample size
How many observations sit behind the figure, stated for each breakdown, not only for the headline.
Limitations
What the figure cannot support. Written before the interpretation, not appended as a disclaimer.
Source and access date
The original publisher, a link to the primary document, and the date it was last checked.

Planned metrics

What the data library will cover, and why each one is hard.

The difficulty listed beside each metric is the reason it has not simply been published already. Naming it in advance is part of the standard.

  • Application-to-interview rate

    The hard part: Almost every published figure comes from a single platform's users. The population has to be stated, not averaged away.

  • Interview-to-offer rate

    The hard part: “Interview” means a screening call in one dataset and a final round in another.

  • Applications per offer

    The hard part: Usually a derived figure. Publishing it means showing which two rates it was derived from.

  • Job-search duration

    The hard part: Official unemployment-duration statistics measure something specific and are routinely quoted as if they measured everyone.

  • Employer response time after applying

    The hard part: Non-response is the majority case, which makes any average of responses misleading on its own.

  • Response rate by posting age

    The hard part: Requires distinguishing a posting's age from whether the role is still genuinely open.

  • Remote versus hybrid versus onsite competition

    The hard part: Applicant counts per posting vary enormously by platform, so cross-source comparison is rarely valid.

  • Referral versus cold application outcomes

    The hard part: Referred candidates differ systematically from cold applicants, so the raw gap overstates the effect of the referral itself.

Related