Insights

Field notes on hiring and AI, from someone who gets paid to decide who's worth the call.

A bad hire cost $15,000 in Texas and up to $500,000 in India

A fractional CTO in India, a founder in Texas and a founder in Bangkok agree AI should cut the cost of sourcing, then show where the bigger bill actually lands.

A bad hire cost $15,000 in Texas and up to $500,000 in India

Most businesses want to haggle over the recruiting fee. Three founders would rather you look at the bill that comes after.

A fractional CTO in India puts one bad Staff or Principal hire at roughly $350,000 to $500,000, "dwarfing any upfront 20% recruiter discount."

A Texas founder says three unproductive months on a $60,000 role is already $15,000 gone.

A Bangkok founder says one rushed hire cost him roughly three times a typical fee. Three countries. Three price tags. None of them is the fee.

We asked the people who pay recruiting fees: if AI is making professional services cheaper, should recruiting follow, and what are you really paying for? Contingency fees often run 20-30% of first-year salary, so the answer moves real money. These three answered by pricing the other side of the ledger: the hire that goes wrong.

Cheap search, cheaper fee. A bad hire still isn't cheap.

Hussain Abdul Rauf Jatoi, Founder of Zahvox (Texas, United States)

“I would not pay a traditional 20 to 25 percent fee when most of the work is database search and automated outreach.”

Hussain Abdul Rauf Jatoi

“On a $60,000 role, three unproductive months already represent $15,000 in salary before considering management time, delayed work, replacement costs, and possible client damage.”

Hussain Abdul Rauf Jatoi

For a clearly defined web, marketing or operations role, he'd rather pay a fixed fee or 10 to 12 percent of first-year salary. Then he reminds you that $15,000 is only the salary line, with management time and client damage still unpriced.

One rushed hire. Three fees gone.

Nick Peplow, Founder of Thailand Care (Bangkok City, Thailand)

“One coordinator I hired in a hurry left after four months. Between salary, my training time and work I had to redo, it cost me roughly three times a typical fee.”

Nick Peplow

Notice his tab includes his own training time, not just salary. He also goes after the pricing model: a percentage rewards the recruiter for a bigger salary, not a better match. What he'd pay for is judgment about people, not résumés an AI tool can pile up in an afternoon.

Sourcing is automated. Staff-level mistakes are not.

Dr. Shashwat Bishwen, Solution Assembler & Fractional CTO of shashwat.si (India)

“Commodity resume aggregation and outbound LinkedIn filtering are 100% automatable with LLMs-I won't pay a 25% contingency fee for keyword-matched profiles anymore.”

Dr. Shashwat Bishwen

“A single bad hire at the Staff/Principal level bleeds roughly $350,000 to $500,000 in unraveled architecture, stalled sprints, and team attrition within 9 months-dwarfing any upfront 20% recruiter discount.”

Dr. Shashwat Bishwen

He is the strongest here on what AI can take, and just as firm on what it can't. Sourcing is cheap, he says, while technical due diligence is where the risk lives. Automate the search. Pay for the vetting.

Where they split

Where they agree: AI can do the search. The India-based CTO won't pay a 25% contingency fee for keyword-matched profiles. The Texas founder won't pay 20 to 25 percent when the work is database search and outreach. The Bangkok founder would pay nothing for a pile of resumes.

Where they split: what replaces the percentage. Texas wants a fixed fee or 10 to 12 percent for defined roles. Bangkok questions the percentage model itself. India keeps the premium for technical vetting.

Your turn

What did your worst hire actually cost?

Next: a Greek founder's two-monthly-salaries split, a Los Angeles firm's $1,600 monthly flat rate, and a Hamburg IT founder's year-one retention tie show three very different ways to stop pricing recruiting as a percentage of salary.

Tags
Was this useful?

Know before you hire.

One email when a new field note is out.