PROFESSIONAL WORK · Recruiting data & AI
Provider data strategy
Benchmarked six provider-data vendors on the same 250-provider sample, renewed the incumbent with no increase against a proposed 25–53% rise, and moved enrichment to licensed data first.
LedCompleted · renewal signed July 2026Documented June–September 2026
The business problem
Sourcing depends on licensed provider data. At renewal, the incumbent vendor proposed a 25–53% increase that bundled add-ons into the renewal.
The call I made
Rather than renew as offered, I benchmarked first: every candidate vendor tested on the same 250-provider sample, a recorded reason for each one ruled out, and a renewal that pays only for the capabilities the pipeline actually uses.
What I led
The vendor benchmark, a data “waterfall” that uses licensed data first and paid research only when needed, and the renewal negotiation, signed off with the CEO.
Verified results
- Six vendors tested on one 250-provider sample, with a documented reason for each one ruled out (one returned no phone numbers on 500 records).
- Incumbent renewal held flat, with no increase, keeping the core datasets and unlimited exports (signed July 2026).
- A second data vendor signed about 19% below list price.
- Using licensed data first cut the direct cost per served lead by about half (September 2026).
Why it matters
The core data contract stayed flat while sourcing volume grew several times over, and every vendor decision has a written rationale leadership can revisit.
THE DETAIL
How it works and how it’s governed
My contribution
Ran the vendor benchmark, designed the data waterfall, and negotiated the renewal, with sign-off from the CEO.
Approach
Shared-sample vendor benchmark · Licensed provider data · Data waterfall design · Contract negotiation
How it works
- InputShared sample250 providers, tested against every candidate vendor
- CheckRule out with evidenceCoverage, phone-number yield and data quality, written down
No phone numbers on 500 records → ruled out
- StepDesign the waterfallLicensed data first; paid research only when needed
- PersonNegotiate the renewalKeep what the pipeline uses; drop add-ons it doesn’t
- OutputRenewal signedNo increase; core datasets and exports kept
Decisions that mattered
Same sample for every vendor
Every vendor was scored on the same 250 providers, so coverage and quality differences were real, not artifacts of different test sets.
Rule out with evidence
Each vendor that didn’t make the cut has a recorded reason, from missing phone numbers to weak coverage.
Pay for what the pipeline uses
The renewal dropped an API add-on the pipeline didn’t need and kept unlimited exports, which it did.
Licensed data before paid research
Checking licensed data before paying for contact research roughly halved the direct cost per served lead.
Scope & limits
This is a proposed increase avoided, not a reduction from what was paid before. Contract amounts and vendor names are left out.