← all playbooks

Size a talent pool before you pitch for the role

A small table you can show a client: how many people hold the title, how many sit inside the radius, how many list the must-have skill, and which employers they work for today. It comes with the plain caveat that database counts are a floor, not a census.

use when
a client is about to brief you, or you are about to quote a fee, and you want to know how many people actually exist before you promise anything
starts from
A job spec

The prompt

Paste it into Claude Code or the Claude desktop app with Hyreflow connected. The first line tells your agent to use Hyreflow, so it reads the play, asks before it spends anything, and hands the work back to you.

paste this into Claude
Use hyreflow and tell me how many people actually exist for this role
before I pitch for it.

<PASTE THE JOB SPEC, OR DESCRIBE THE ROLE>

Location: <CITY>, within <KM> km. Must-have: <ONE_OR_TWO_MUST_HAVE_SKILLS>.

- Count first, do not pull people. Read the total each database reports
  on the smallest page you can ask for.
- Give me the funnel: everyone with the title in <COUNTRY>, then inside
  the radius, then with the must-have skill, then at <SENIORITY>.
- Widen the title to the variants people really use, and show me which
  ones you included.
- Tell me which number came from which source. If two sources disagree,
  show both. Do not average them.
- Then pull a small sample of real profiles, only so you can tell me
  which employers these people work for today.

Tell me what the sample costs before you pull it. Finish with a table I
can show a client, and say plainly that the counts are a floor.

Replace every <PLACEHOLDER> with your own detail. Everything else can stay as written.

What you need first

  • The job spec, or a plain description: title, location with a radius, and the one or two skills that are truly must-have
  • A Hyreflow workspace with credits. The counts cost next to nothing and the employer sample is a small paid pull
  • Optional: your own Apollo key for one more count. Apollo is bring-your-own-key and uses no Hyreflow credits

Tools it can reach for

The agent picks per step from what your workspace has. Nothing here is required by name.

What happens when you run it

Free steps are marked free. Anything that spends credits is marked, and the agent asks before the first paid run of any size.

  1. 1

    Turn the spec into countable terms

    free

    The title is widened into the variants people really carry, the radius becomes the towns inside it or a metro area, and the must-haves are cut to the one or two that truly rule people out. Every extra must-have shrinks the pool, so the agent asks before it adds one.

  2. 2

    Read the counts, not the people

    credits

    People databases report how many profiles match a search alongside whatever page you ask for. One source previews a search free and gives its total. The others return the total with the smallest page they sell, so a count costs one tiny page and not the pool.

  3. 3

    Build the funnel one filter at a time

    credits

    Title across the country, then inside the radius, then with the must-have skill, then at your seniority. Adding filters singly shows which requirement empties the pool, and a count that fails to drop exposes a filter the database ignored.

  4. 4

    Cross-check between sources

    credits

    The same cut is counted in more than one database. They index different people and match titles differently, so the numbers differ, and the table shows each one with its source. With your own Apollo key connected, its total joins the comparison.

  5. 5

    Sample the employers

    credits

    Once you approve the cost, a small page of real profiles is pulled and rolled up by current employer. That shows where the pool sits today: which companies hold a cluster and which hold one person each.

  6. 6

    Count the competing demand

    credits

    Live ads for the same title in the same area are pulled from the job boards, kept small because you pay per ad returned. A thin pool with a dozen employers advertising into it is a different pitch from a thin pool nobody else is fishing in.

  7. 7

    Build the client table

    free

    One small table: each cut, its count and its source, then the main employers and the live ads. Underneath sit the caveats in plain words, so nobody quotes the number as a census.

What a run costs

Credits are spent per role the play actually works, and a lookup that finds nothing usually costs nothing. The two figures are the run where the first provider answers and the run where every lookup walks its full chain.

rolesif the first provider answersif every lookup walks the chain
25$10100 credits$10100 credits
100$41410 credits$41410 credits
500$2032030 credits$2032030 credits
1,000$4054050 credits$4054050 credits

Free before anything is charged

  • Turn the spec into countable terms
  • Build the client table

What moves the number

  • How many roles survive the free filters. Everything dropped before the paid steps costs nothing.
  • Providers you connect with your own key. Those calls bill your account, not your credits.

An estimate, not a quote, priced at the volume credit rate. Your agent sizes the run against your own workspace and tells you what it will cost before it spends anything.

The number that changes the fee conversation

A client says the role is hard to fill, and the conversation about terms starts from there, on feel. A count changes it. "This many people hold the title inside your radius, this many list the skill you insist on, and most of them sit at four employers" is a sentence that justifies a retained fee, a wider radius or a dropped requirement before anybody has started searching.

It is also cheap to find out. People databases report how many profiles match a search whether or not you buy the profiles, so the play reads the totals and buys almost nothing. The only real purchase is a small sample of profiles, which shows where these people work today.

What you get back

A table with one row per cut: the title nationally, inside the radius, with the must-have skill, at your seniority. Each row carries the count and the source it came from, and where two sources were asked, both numbers. Below it, the main employers found in the sample, the live ads competing for the same people, the title variants and places that were included, and the caveats.

Variations worth knowing

Test a requirement. Count with and without one must-have. If a single skill removes most of the pool, that is a conversation to have with the client before the search, not after it.

Compare two locations. Run the same funnel for a second city or a wider radius. It shows what a hybrid arrangement or a relocation package would actually buy.

Turn the count into the search. Once the role is yours, turn a job spec into a ranked shortlist starts from the same filters. For a pool that turns out to be very thin, map a whole talent pool, including the half that is not looking covers the angles a title search misses.

Where this goes wrong

A filter the database ignored. A total the size of the whole database means a filter was not applied, not that the market is huge. Each added filter has to lower the count, and the agent checks that it does.

Skills are under-reported. Not every database can filter on skills, and many people list none. The skill row comes from fewer sources and undercounts, sometimes by a wide margin.

A radius is an approximation. Place names stand in for distance. Someone who lists only the country falls outside every radius, however close they live.

The employer list comes from a sample. It shows where the clusters are, not the full distribution.

Quoting the count as a census. It is a floor, drawn from sources that each see part of the market. Presented that way it builds trust. Presented as a fact, it invites the client to check it.

Questions

Is the count really free?

Close to it. One source previews a search without charge and reports the total. The others bill by the page or by the row and report the total on any page, so the agent asks for the smallest page there is and reads the number. A full funnel is a handful of those pages. The employer sample is the only part that buys real rows, and the agent quotes it before pulling it.

Why are database counts a floor and not the real number?

Because a database only counts the people it holds, under the title and the skills they chose to list. People with no public profile, an unusual job title or an empty skills section are missing, so the real pool is usually larger than the count. Some profiles are also out of date. Read the figure as the right order of magnitude and as a number you can defend, not as a headcount.

Two sources gave me very different numbers. Which is right?

Neither on its own. Each database indexes a different slice of the market and matches titles in its own way, so the table shows each count beside its source and does not blend them. For a pitch, the larger credible number is your floor, and the spread between them shows the client how much of the market sits outside any single tool.

How does a radius work when databases do not store distances?

They match on place names, so the agent turns your radius into the towns inside it, or uses a metro area where the database offers one. Most people list the metro and not their suburb, so the place list is built wide. It approximates a commute belt, and the table names the places that were included.

What do I do with the number?

Set expectations before a fee is agreed. A pool in the thousands supports a contingent search and a quick shortlist. A pool in the dozens, mostly sitting at three employers, is an argument for a retained search, a wider radius or a relaxed must-have, and the table is the evidence. If you win the role, the same filters become the search.