Free research · 3 working papers
The numbers nearshore vendors don't publish. We did.
How fast our side of a vetted hire moves, who's still on the team a year later, and which vetting scores actually predict anything. From 83,000 candidates and 15,000 check-ins.
- candidate profiles
- 83,215
- structured check-ins
- 14,874
- CVs presented to U.S. clients
- 2,099
- client positions
- 625
What's inside
Three questions every CTO asks a nearshore partner. Answered with data, not a sales deck.
RecruitingPaper 01
How long should a nearshore hire take?
First CV in 4 daysThe profile of the engineer you end up hiring in 7.5. Medians for roles filled in 2026.
- How long each step we control takes, from the opening to finding your perfect fit
- Why 3 in 4 engineers we placed in 2026 were vetted before the role opened
- How many CVs and interviews a filled role really takes
RetentionPaper 02
Would you see an exit coming?
92% at 12 monthsWhen the role continues. 63% of recorded endings were started by the client, mostly for budget.
- 12-month retention: 78% from any cause, 92% when the role continues
- The check-in signals that precede an ending, and what the manager does with them
- Validated on 2025–26 data the model never saw (AUC 0.76)
VettingPaper 03
Does a vendor's vetting score mean anything?
6 in 10of the engineers we score highest get a client interview. Our technical score predicts who you'll want to meet.
- The technical score predicts who you'll want to interview
- Soft-skills screening works as a gate: fewer than 1 in 3 candidates pass
- No pre-hire score predicted retention. We say so.
Including what didn't hold up
Every vendor says it's selective and that its engineers stay. We had seven years of our own records to check whether that was true for us, so we checked, and published the parts that didn't flatter us too.
- Data
- BEON's own operating records: recruiting pipeline, client decisions, placement histories, the structured check-ins run by Talent Experience Managers and the recorded reason each engagement ended, read against the written procedures each paper describes.
- Validation
- The early-warning model in Paper 02 was fitted on earlier placements and tested on later ones it had not seen.
- Limits
- One firm, one region, mostly U.S. clients. The papers describe what happened inside our process, not the nearshore market as a whole. They show what moves together, not proof that one causes the other.
3 working papers · PDF · free