The price of a miss should be zero

Classic data vendors sell you a file. You pay for the file. Then you discover that a slice of the emails bounce, a slice are catch-alls, and a slice might be real. You already paid for all three. The pitch that justifies the invoice is usually a round number: contact data decays at 30% a year, so buy the refresh. It appears in decks across this category. It has been repeated until it feels like physics.

It is not physics. Lusha went looking for the study behind the 30% line, could not find one, and measured the thing instead — twice, on the same cohort of US sales leaders, VP and C-suite, about 140,000 records. In the run they published for 12 September 2026, 12.25% had changed roles over twelve months. Over twenty-four months it was 25.67%. A month earlier the same cohort was 12.6% and 25.7%. UK sales leaders moved faster: 21.40% in a year. Their own conclusion, which is the part worth stealing, is that the famous 30% is closer to a two-year rate being sold as an annual one. Employment changes are not the only way a record goes wrong, so the true “this row is now a lie” rate is higher than 12%. It is not a license to invent 30. The write-up is on their site, method and both runs included. Cite it. Do not cite the ghost.

Here is the controversial reading. A smaller, real number is worse for a vendor’s urgency and better for your plan. You do not need a fake 30% to justify enrichment. You need to stop paying for the rows that are already wrong, and stop paying again when a provider cannot find the right one. Twelve percent of the leaders you emailed last September are not in that chair. Twenty-six percent over two years is a file that has quietly become a different company. If your contract bills you for the attempt, you are funding the decay.

Credits move when there is a match

LeadHootz runs a multi-provider waterfall for work email, verified email, and mobile. Providers are tried in order and stop on the first quality match. You are not “checking everyone twice” for sport. You are covering the shortfall. Credits move only when a match is found. A no-match is not a line item. That sounds like a billing footnote. It is the strategy. The honest pitch is not “we find everyone.” It is “you do not pay for the ones we cannot find.” That is the only enrichment math that survives a finance review once someone asks why last year’s file is being purchased again.

Verified email is the safer default for outbound. If a row already has an address from last year, re-verify before you enroll. Old data is how bounce rates surprise you on day one, and bounces are how you walk toward the spam line Gmail published — stay under 0.10% user-reported spam, and do not touch 0.30%. A decayed list is not a neutral asset. It is a deliverability decision you already made. See midnight is not a send window.

Do not enrich the ocean

Sample a batch. Read the hit rate in Live Activity. Then waterfall the rest. If the sample is embarrassing, the ICP is the problem and more credits will not repent it. Mobiles cost more. Add them on engaged or high-intent rows, not on the whole TAM. Phone-first on 20,000 cold names is how credits vanish and calling still feels empty. The operating version is the credit-only enrichment playbook.

The public API uses the same people and company enrich, on the same credit pool. If marketing wants enrichment in the CRM, they should not buy a second vendor that disagrees with the app and bills the misses the first vendor refused. See the API docs. Two truths about the same person is how a buying group — already inclined to conflict — gets two stories. Gartner found 69% of buyers already see the website and the seller saying different things. Do not add a third story from a second database.

Refresh because people move, on a cadence that matches the territory, and verify at the moment of send. Do not refresh because a slogan said a third of the file died in its sleep. Pay for the match. Let the miss cost nothing. That is the whole argument, and it does not need a fake statistic to be urgent.