Jess pulled 847 records, counted 288 matched. 34% match rate. The SDRs said the tool was broken. She spent three evenings going through 559 failures one by one. The tool was not the problem.
The SDR team had been complaining for three weeks before Jess pulled the actual numbers. They said the email finder tool was broken. They said the match rates were terrible. She pulled the last four weeks of enrichment output — 847 company records — and counted what came back.
Verified email addresses: 288. That was a 34% match rate.
She felt two things simultaneously: the SDRs were right that something was wrong, and the explanation was not "the tool is broken." A tool does not run at 34% on a list of 847 companies without those 288 being genuinely correct. Something was producing failures on the other 559. She needed to know what, because the answer to that question was the answer to whether she was switching tools or fixing the input.
She had been using the term "match rate" for two years without thinking carefully about what it was actually counting. She thought it meant: how often the tool finds a valid email for a given input. That was approximately right but imprecise in a way that was making the diagnosis harder.
A match rate in an email finder context is the ratio of verified, deliverable email addresses returned to total records submitted — a single aggregate metric that collapses multiple structurally distinct failure types into one number, each of which has a different root cause, a different fix, and a different downstream impact if left unaddressed.
The aggregation was the problem. When the SDRs said the tool was returning a 34% match rate, they were describing a combined outcome of several different problems as if it were one. Jess opened a new spreadsheet and started categorising the 559 failed records manually — not to find the tool's failure, but to find her own.
Nine were acquisitions — the acquired company's domain passed to the email finder, employees migrated to the parent company's infrastructure months earlier. Eleven were rebrands — old domain sometimes still live but no active employee mailboxes configured. Eleven were records where the data provider had returned the wrong domain entirely — a subsidiary, a holding company, a different entity with a similar name.
This was not a match rate problem. This was a domain quality problem manifesting as a match rate problem. The email finder's output was accurate. The input had been wrong.
A catch-all domain accepts delivery to any address at the domain regardless of whether a specific mailbox has been provisioned. An email sent to zzz.randomstring@company.com receives the same acceptance response as an email sent to a real employee. The tool connects to the mail server, asks whether the address is valid, and gets "yes" for every probe — because the server accepts every address. The tool flags the domain as catch-all and returns nothing. That is the honest response. There is no verified result to return.
Industry data puts catch-all configuration at roughly 22% of active B2B domains — which made 124 of 559 failures exactly what the distribution predicted.
The first one: the name field read Chen, Michael. Last name, comma, first name — the default export format from at least three CRM database systems. The tool's name parser had received it and returned nothing rather than guess at the intended order. Six more in the same format. Three records with diacritics or multi-word surnames the tool hadn't normalised. Two records with no name at all — just a job title. Three records with full legal names contacts didn't use professionally.
Fourteen failures that were not the email finder's fault and not the domain's fault. The problem had originated in a CRM export format that nobody had reviewed before passing it to the tool.
Email finder tools construct addresses through pattern inference: they identify confirmed addresses for existing employees, establish the format, and apply that pattern to the contact being searched. When no confirmed addresses exist in the tool's database for a given company, there is no pattern to infer.
Forty-eight records from companies with fewer than 25 employees. No press coverage. No blog posts with author bylines. No PDFs with footer contact details. Nothing that would have seeded an email pattern database. A 2025 benchmark across 20,000 B2B contacts confirmed meaningful coverage gaps correlating directly with companies below 25 employees, regardless of which email finder was used.
The email finder had returned a verified address at the company's domain. SMTP verification had confirmed it as deliverable — because the old mailbox had not been deprovisioned when the contact left eight months ago. The mail server still accepted connections to it. The address was technically live. The person was not there.
Jess checked a random sample of 60 of the 288 matched records against LinkedIn. Seven of the 60 — 11.7% — had contacts no longer at the company. Extrapolated: approximately 34 of her 288 matched records were technically verified but operationally wrong. The true operational match count was 270, not 288.
She manually searched 10 at random. She found emails for six — not from any database tool, but from PDFs, press releases, event speaker bios, and LinkedIn posts. The email finder had returned nothing not because the emails did not exist, but because they did not exist in its index.
No email finder tool has complete coverage of the B2B universe. Every tool has gaps — companies that are real and active but not represented in its data sources because they have never appeared in a source the tool indexed. This is not a tool failure. It is a data coverage reality that no single tool resolves completely.
| Root Cause | Records | % of Failures | Recoverable |
|---|---|---|---|
| Wrong domain in input | 31 | 5.5% | All 31 — upstream domain fix |
| Catch-all domain — SMTP verification impossible | 124 | 22.2% | ~56 — LinkedIn-first workflow |
| Name format problems in input | 14 | 2.5% | All 14 — normalisation script |
| Small company — no detectable pattern | 48 | 8.6% | ~17 — size-tiered routing |
| False positives within matched output | 18 | 3.2% | Removed via quality gate |
| Not in tool's database | 342 | 61.2% | ~68 — supplementary tool |
| Total failures | 559 | 100% | ~186 additional contacts recovered |
Every company name resolved to a current operational domain, confirmed by MX record check, before the list reaches the email finder. The 31 wrong-domain failures addressed at the source.
Records flagged as catch-all moved to a LinkedIn-first workflow with engagement-based address confirmation on first contact. Recovered approximately 56 of 124 catch-all records.
Comma-separated names inverted, diacritics normalised, no-name records flagged for manual contact identification before enrichment. All 14 name-format failures recoverable.
Below 25: LinkedIn-first. 25–50: email finder with manual review. Above 50: standard automated. Recovered approximately 17 of 48 small-company records.
20% random sample of matched results checked against LinkedIn before sequence enrollment. Wrong-company matches trigger full cohort review. Removed the false positive layer contaminating the 288 matched records.
All records returning no match from the primary tool passed to a second tool with different data source composition. Recovered approximately 68 of 342 database-gap records.
A 34% match rate is not a statement about the tool's capability. It is a statement about what the tool received.
The confused reply on that Wednesday afternoon — from a person who had left the company eight months earlier and was genuinely puzzled about being in a sequence — had cost them one sequence slot and started the entire diagnosis.
The thread that unravelled 559 failures had begun with someone who no longer worked there, politely asking to be removed.
The SDRs had been right that something was wrong. They had been wrong about what it was. "The tool is broken" requires no further work. "Here is specifically what broke and why" is correct.
FindCompanyDomain resolves company names to verified, MX-confirmed operational domains before they reach your email finder — so the tool starts with the right input on every record.
