We tested 12 data vendors on one job change.
Ten got it wrong.
Job-change signals drive a huge share of B2B outbound. So we ran one recent, public move through every job-change data source we have. Two weeks later, most tools still returned the wrong company — including one built specifically to detect job changes.
Everyone runs on the same signal. Almost no one checks if it's right.
A job change is the highest-value trigger in outbound. New leader, new budget, warm reason to reach out. Every GTM stack now buys it, and most teams treat whatever their data vendor returns as fact. We wanted a simple answer: when someone changes jobs, how quickly and how accurately do these tools actually know?
So we picked one move we could verify with certainty and graded every source against it.
Mikayla Hopkins left Tracksuit to become VP Marketing at Confido around July 14, 2026, and announced it publicly. She had also just posted about how fast the outbound machine reacted to her move:
"One job-change signal, and every GTM stack now thinks I'm the most qualified lead they'll touch all quarter. 32+ SDR calls. 53+ blank 'quick question' emails… The tools are the same for everyone now."
— Mikayla Hopkins on LinkedIn. Her move fired 80+ touches in days — so the tools clearly noticed. We tested whether they were also correct.
The job-change signal is free, and missing from most of what you pay for. The tools that flooded her inbox were reading LinkedIn live. That layer caught the move in days.
The paid databases — the ones you buy because they promise verified, accurate data — got it wrong. Two weeks after a public announcement, ZoomInfo, People Data Labs, Ocean.io, Crustdata, ContactOut, LeadMagic and four more still return the old company. ZoomInfo does it with a 93/100 confidence score attached.
So the signal everyone runs on is a commodity anyone can pull for free, and the "system of record" you pay for is often the slowest, most confident way to be wrong.
Every source, graded against one known answer
Ground truth: VP Marketing at Confido (confidotech.com), started around July 14, 2026, announced publicly. Every source was queried on July 28. "Correct" means it returned Confido.
| Source | What it returned on Jul 28 | Freshness it showed | Verdict |
|---|---|---|---|
| Correct — both read LinkedIn live at query time | |||
| Wiza Live lookup | Vice President Marketing, Confido Correct title and employer. | Resolved at query time | ✔ Correct |
| Findymail Live lookup | Vice President Marketing, Confido Correct title and employer. | Resolved at query time | ✔ Correct |
| Wrong — cached databases still serving the previous role | |||
| ZoomInfo | Head of Marketing, Tracksuit No record of the new company at all. | Confidence 93/100 · marked "valid" Jul 1 | ✘ Stale |
| People Data Labs | Head of Marketing, Tracksuit Re-checked the old role three weeks before the move, and still missed it. | job_last_verified: 2026-06-25 | ✘ Stale |
| Crustdata | Head of Marketing, Tracksuit Saw the job end. Never added the new one. | Logged the old role ending 2026-06-01 | ✘ Stale |
| Ocean.io | Head of Marketing, Tracksuit Its own role-change check returned zero changes. | Profile "updated" 2026-06-30 | ✘ Stale |
| ContactOut | Old role + an old board seat Returned a board seat as the current company. | Both marked current | ✘ Stale |
| Datagma | Head of Marketing, company: blank Sure about the person, blank and wrong on the company. | confidence_score: 1 | ✘ Stale |
| LeadMagic | Tracksuit, "Present" No way to tell how old the answer is. | No freshness field | ✘ Stale |
| DiscoLike | Head of Marketing, Tracksuit Still the previous role. | No freshness field | ✘ Stale |
| Prospeo | Board Member; old role still "current" Picked a board seat as the current role. | last_job_change_detected_at: null | ✘ Stale |
| Icypeas | Head of Marketing, Tracksuit Knows she left. Has no record of where she went. | Logged the old role ending 2026-06-01 | ✘ Stale |
Scored set = 12 tools that returned a current-employer answer (Wiza and Findymail, plus the 10 that were wrong). A few other tools were queried but didn't return a gradeable current-employer answer, so they're excluded from the 12.
"Verified" and "job-change detector" both got it wrong
A job-change detector that missed the job change
Datagma's dedicated job-change tool returned Tracksuit as the current employer, with endTime: "Present" and duration: "3 years 5 months." The one tool built to catch this event reported the old role as ongoing.
It re-verified the wrong answer
People Data Labs stamped job_last_verified: 2026-06-25 on the old role — a check dated three weeks before the move. A freshness date tells you when a source last looked, not whether it's right.
Two failure modes, and one misleading number
The misses weren't random. They fell into two patterns, and the "confidence" scores pointed the wrong way in both.
Old data with a fresh-looking date
ZoomInfo ("valid Jul 1"), Ocean ("updated Jun 30") and PDL ("verified Jun 25") all carry recent-looking dates on a two-year-old role. Crustdata and Icypeas even recorded that the old job ended, then never added a new one. They know something changed. They just don't know to what.
The wrong current role
ContactOut and Prospeo returned an old board seat as the current company, because two roles were both flagged current. Datagma returned the title with the company left blank. Even the tools closest to the answer picked the wrong row.
The misleading number: confidence scores measured whether the tool found the right person, not whether the job was current. Datagma returned confidence_score: 1 and ZoomInfo 93/100 while both were wrong about where she works. Nine of the tools showed no freshness field at all.
Signal is instant. The database of record lags.
The three sources that were right share one trait, and the ten that were wrong share the opposite one. That split is the whole story.
Reads LinkedIn live
- LinkedIn itself (Sales Navigator alerts)
- Wiza — live lookup, returned Confido
- Findymail — live lookup, returned Confido
Serves a cached record
- ZoomInfo, People Data Labs, and similar databases
- Ocean.io, Crustdata, ContactOut, LeadMagic
- Datagma, DiscoLike, Prospeo, Icypeas
- Sold on "verified accuracy." Still wrong.
If your stack pulls job-change signals from a database of record, you're working from a snapshot that can be weeks old, with a confidence score that won't warn you. The move isn't hard to detect — it was on a public profile and it triggered eighty touches. The gap is that the layer most teams pay for and trust is the slow one. For freshness, a live lookup beats a cached database.
How we ran it
- One clean case: a VP Marketing who publicly moved companies around July 14, 2026 (previously Head of Marketing at the prior company). Ground truth confirmed from the person's own announcement and live LinkedIn profile.
- Queried July 28, 2026, about 14 days after the public move. One lookup per source, by LinkedIn URL. Read-only — nothing was pushed, no outreach sent.
- Every result was re-checked against each vendor's raw response before publishing. The new company appeared in exactly two paid-provider responses: Wiza and Findymail.
- The scored set counts only tools that returned a clear current-employer answer. A few others were queried but didn't return a gradeable result, so we left them out rather than guess.
- One case is one case. This isn't a coverage rate — it's a worked example of a pattern any GTM team can reproduce on their own recent job-changers.
We benchmark the tools instead of reselling them.
If you want to know which data actually moves your pipeline — and which "verified" record is quietly weeks out of date — that's the kind of work we do.
Talk to 42 Agency →