The economics of a narrow desk are unintuitive until you have run one. A smaller market should mean less work. In practice it means better work, because everything you learn stays relevant.
The compounding argument
On a generalist desk, the knowledge from a search is mostly spent. Different sector next week, different vocabulary, different decision makers.
On a specialist desk, every conversation adds to a stock you will use again. The same forty companies. The same credential landscape. The same twelve people who decide things. Year three is not year one repeated, it is year one plus two years of context.
That is why narrow desks eventually out-bill broad ones despite having a fraction of the addressable market.
What "niche" should actually mean
Not just a sector. A workable niche has three edges:
Vocabulary. You use the words the market uses, including the ones outsiders get wrong. This is the fastest credibility test there is.
Decision map. You know who signs, who blocks, and who is consulted, without asking.
Supply picture. You know roughly how many credible people exist and where they sit. Broad desks guess at this. Narrow desks know.
If you have all three, you can charge for judgement. If you have none, you are competing on turnaround time.
The risk everyone raises
Concentration. A narrow desk is exposed if its market contracts.
The mitigation is not to widen, it is to hold adjacency. Two or three neighbouring markets that share vocabulary and people, so a downturn in one is absorbed rather than fatal. Wide enough to survive, narrow enough to still know things.
Where the tooling usually fails specialists
Most systems built for recruitment are built for volume markets. They optimise for pipeline stages and throughput, and they treat your accumulated market knowledge as unstructured leftovers.
That is the gap Holi was built for: a partner that holds niche vocabulary, remembers who told you what, keeps a source against every claim, and gets sharper the longer you work one market. The narrow desk's advantage is memory. Memory should be the thing your systems are best at.
How to narrow without losing this year
You do not have to refuse work. Narrow the investment rather than the intake:
- Capture in depth for one market only, at first
- Publish or share only in that market
- Take the other work, but stop trying to accumulate in it
Within two quarters the specialist half of the desk will be producing shortlists faster than the general half, and the decision will make itself.