The Cost of Normal

Hiring treats every vacancy as a single moment. A person applies, a recruiter makes a call, and the relationship ends. That works for the one person hired. It fails everyone else, including the employer who will need those same people again next month.

We started Levlr because we kept seeing the same pattern in hospitality. Employers attract more applicants than almost any other sector, then pay to attract them all over again. The people were never the scarce part. The memory was.

How normal took hold

Job boards made advertising almost free. Bank of England researchers found the real cost of advertising a UK vacancy fell by more than 80% after 2000 (Bank Underground).

When an advert costs little, reposting becomes the answer to every vacancy. Post, collect, filter, reject, repeat.

That era is ending. Indeed now limits employers to three free job posts a month, and since March 2026 jobs sent through a single feed only appear if the employer pays to sponsor them. For one of our clients, losing that free traffic adds around £90,000 a year. Reposting was always wasteful. Now it costs real money too.

What normal produces

Globally, hospitality roles attract 117 applicants per hire, and only 0.6% of applicants receive an offer (SmartRecruiters). Most of the rest get a rejection email and nothing else.

Then the cycle restarts. Hospitality turnover reached 38.7% in 2024, the highest of any sector in data from over 4,000 UK employers (RotaCloud). Across the UK, 78% of hiring replaces people in roles that already exist (Jobtrain).

So employers keep hiring for the same roles, from the same pool, as if each time were the first.

The missing memory

The applicant tracking system still holds last month's applicants. But it judged each one against a single role. A strong candidate who missed that role never hears about the one next door.

Where employers keep that memory, it pays. Across Gem's customers, 46% of sourced hires now come from people already in the company's own database, up from 26% in 2021 (Gem).

Why hospitality missed it

The tools that rediscover candidates were built for large corporate employers. Beamery typically serves organisations of 1,000 or more employees (Edvenity). Third-party estimates put Eightfold at $150,000 or more a year (Voyse).

Hospitality looks nothing like that. 99.6% of UK hospitality businesses are SMEs (House of Commons Library). Multi-site operators hire too often to track candidates by hand, yet sit far below the size these platforms target. The enterprise tools also only look inside one employer's database.

Why we built Levlr

We are in business to repair recruitment. Hiring should work as a relationship, not a single moment.

The cost was never the advert (though with Indeed's prices rising, that is changing fast). At most clients, job board spend works out at £13 to £20 per hire. The real cost sits in the good people an employer already attracted, then let go.

So we built the memory hospitality never had. Intelligent matching ranks and engages every applicant inside the employer's existing system and brings past applicants back into view. Everyone who misses a role gets a next step. Through Community, a good person missed by one employer stays visible to the next.

The early results point the same way. At Champneys, applications from people already in the talent pool rose from 9% to an average of 35%. At Malmaison they rose from 10% to 21%, with no extra advertising. Candidates who re-engaged went on to receive offers at 2.8 times the rate of those who did not at Champneys, and 2.5 times at Malmaison.

The cost of normal

Normal hiring treats each vacancy as new and each applicant as used up. Hospitality pays twice: once to attract people, and again to attract them back or just conduct the same search again.

The applicants are already there. The question is whether your system remembers them.

See how Levlr works with your existing hiring system at levlr.io.

Research credit. This paper draws on research by University of Nottingham students (Team 8, Practera, April 2026): Aung Myat, Erraime Mitchell, Jerusha Daram, Sebastian Latif and Yohji Jones. Further research by Yohji Jones (May and June 2026).

Next
Next

Building Levlr: Why This Problem Is Worth Solving