Data at cost

Contact data has collapsed in price. Specialist finders sell a verified work email for about one to two cents: LeadMagic advertises "from $0.007 a credit" with failed lookups free, and Findymail claims 2.08¢ per usable email (vendor claim). Verification costs a fraction of that, with MillionVerifier at $449 per million at volume. The only player still charging a platform premium is Clay, whose March 2026 plans start at $185 and $495 a month. A comparison blog estimates Clay waterfalls at $0.15–0.25 per valid email, against ~$0.06–0.08 for FullEnrich and BetterContact (third-party estimates). See Contact data and enrichment and Email verification.

The obvious move is to resell a waterfall at near cost inside the sequencer. The problem is that everyone has already made it.

#The opening

What is already sold, as of October 2026:

Saleshandy publishing its supplier list shows how commoditised the waterfall has become.

The real gap is catch-all. Catch-all domains accept every address at the SMTP check, so ordinary verification returns "unknown". Specialists now verify them: BounceBan claims 97%+ accuracy on catch-alls, FullEnrich claims an "80% success rate", and Scrubby sends real emails and observes for 72 hours. Reoon advises sending catch-alls "with caution", meaning as a separate segment. Yet no sequencer checked in this research routes catch-all addresses to their own mailbox pool with their own bounce budget automatically (Email verification, by absence in the docs reviewed).

Bad data is still a top complaint. Apollo reviewers report "Well over 65% bounce and failed rate" (Sep 2026, single review). A directory profile of PlusVibe contrasts its native verification at "~4% bounce rates versus ~0.8% with dedicated tools". Instantly's benchmark says bounces should stay below 2%.

#Who pays and how much

Everyone who sends pays for data, but they pay through bundles:

• Instantly credit plans start at $47/month for 1,500 credits.
• lemlist's enrichment add-on is $20/month for 2,000 credits, about 400 emails, roughly $0.05 an email.
• Apollo seats run $49–149/user/month (third-party), and active teams reach "$150–400 per user per month" by the same source's estimate.

A wholesale waterfall at $0.03–0.04 per valid email undercuts Clay by 4–6x (my arithmetic on the figures above). It does not undercut Instantly or Saleshandy credits enough to make anyone switch sequencers.

#Why incumbents have not closed it

They mostly have. The waterfall is built and bundled. What they have not done:

• Make catch-all a policy rather than a status. Their bounce protection is campaign-level (Bounce protection), so a catch-all-heavy list burns the same mailboxes as the clean list.
• Re-verify at send time. Lists sit for weeks between import and the third follow-up (Built-in email verification).
• Refund on bounce. Findymail sells a <5% bounce guarantee with credit refund. The sequencers sell accuracy claims instead.

#What you would build first

1. License two or three finders and two verifiers behind one pay-on-valid credit. Do not build an SMTP verifier. Microsoft's postmaster policy forbids "namespace mining … verifying email addresses without sending", so probing Outlook.com from your own IPs puts your sending reputation at risk.
2. A catch-all lane. Catch-all and "risky" contacts go to a dedicated mailbox pool, sent last, at lower volume, stopped when that pool's bounce rate crosses about 2%. This is a Mailbox health engine feature as much as a data one.
3. Re-verification before step one, priced into the send.
4. A bounce refund, to put your money where the accuracy claim is.
5. Provenance per record: supplier, date and legal basis. In the EU you become the controller for data you sell. KASPR was fined €240,000 partly for missing Art. 14 notices. See Data sourcing law.

#How the leaders would respond

They would not need to. Instantly and Saleshandy can match any wholesale price because they buy from the same suppliers. If the catch-all lane caught on, it is a few weeks of work for Smartlead, whose campaign settings already include "isolated provider sending". The suppliers themselves (LeadMagic, Findymail, BetterContact) can sell to your customers directly at the same price.

#Scores, argued

Pain: 3. Bounces and credit confusion are real complaints, but buyers already have cheap options. The pain is mild irritation, not urgency.

Gap: 2. The waterfall, pay-on-valid billing and bundled verification are all shipped by incumbents and specialists. Only the catch-all send policy and send-time re-verification are open.

Size: 3. The data wallet is large (Clay alone is estimated at $150M ARR by May 2026, a Sacra figure), but selling at cost means you keep little of it.

Moat: 1. You resell other people's data at a thin markup to buyers who can go to the same suppliers. There is nothing to hold.

Speed: 5. Supplier APIs exist and a working waterfall is days of integration work.

Safety: 3. Reselling personal data makes you a GDPR controller with notice and access duties. SMTP probing conflicts with Microsoft policy. Licensed, notified data from EU-native suppliers lowers both risks.

#What would kill it

#What this means for an entrant

• Ship it as a feature in month two, not as the pitch. Pay-on-valid, verification free on every plan, catch-all lane on by default (Built-in email verification).
• Stack it with Mailbox health engine. The catch-all lane only works if the engine can give a mailbox pool its own bounce budget.
• Stack it with EU-native compliant outbound. Wire in EU-native sources such as Dropcontact, which is pay-on-success with no contact database and EU servers, plus Leadfeeder and Cognism, and log provenance per record. "Data whose source you can name" is the EU version of this opening, and the only one with a moat (Non-English markets).
• Publish bounce rates by supplier. You will see real bounces across all customers. A quarterly report is the neutral benchmark this layer lacks.
• Do not build a database. KASPR deleted its LinkedIn-derived database in 2026. Proxycurl shut down after LinkedIn sued. See Built-in lead database.