The Call She Made at 7pm on a Wednesday
She called a friend who ran data operations at another B2B company — someone she trusted to tell her something uncomfortable if that was what the situation called for. She explained the metrics, the gap, the rewrites that had changed nothing. The friend listened and asked one question. Not about the copy.
Nadia paused. She knew the list had been enriched. She said something approximate — "it was clean when we got it" — and her friend said: "That's not the same thing."
A data mismatch in B2B outreach is a condition in which a contact record is structurally complete and technically valid — passing every verification check, containing no missing fields, generating no errors in enrichment or delivery systems — but contains a divergence between one or more data points and the current operational reality of the contact or company it represents.
The record is not broken. It is out of sync. And because it passes all standard quality checks, the mismatch produces no visible signal except the downstream symptom Nadia had been staring at for three weeks: a reply rate that defies everything else the metrics say.
The friend told her that data mismatches are not one problem. They are a category of problems — distinct types, distinct causes, distinct detection methods, distinct fixes. Most teams treat them as a single list quality issue and apply a single intervention that addresses none of them fully. Nadia went back to her list that night and started auditing for each type.
What She Found — Four Distinct Mismatch Types
Domain-Contact Misalignment
The first check Nadia ran compared the email domain in each contact record against the current operational domain of the company that contact was supposed to be at. The result: around 40 records where those two things did not match.
She opened the first flagged record. A contact at a mid-size SaaS vendor she recognised. The email domain in the record was the company's old name. They had rebranded eight months earlier. The old domain was still live — it resolved, it accepted SMTP connections, it passed verification — but no employee had an inbox configured there anymore. Her outreach emails had been delivering to a catch-all server and disappearing.
The pattern held across the flagged records: acquisitions where email had migrated to parent infrastructure, rebrands where the old identity was still partially live, contacts who had changed companies entirely. Every single one had passed enrichment verification. Every single one had delivered. Not one had been received.
Delivery is not the same as receipt. She had just not thought about the difference before.
According to LinkedIn's 2024 State of Sales report, 41% of B2B contact data is inaccurate or outdated at any given time. Nadia had always read that statistic as an abstract warning. She now understood what it looked like inside a single 300-person list.
Role Decay
The second audit explained something that had been bothering Nadia since week one. If emails were not reaching anyone, open rates should be low. But 32% was a healthy rate. Somebody was opening the emails. They just were not replying.
She compared the title field of every contact against their current LinkedIn profile, looking for contacts whose current role had moved materially from the role her sequence was designed for. Around 25 contacts had changed roles — not companies, roles within the same company.
One had been a Director of Sales Development when the record was created. He was now a CRO. Her message was offering to help him build his SDR team's outreach process — a problem he had promoted out of. He had opened the email. Of course he had. The subject line was relevant to a role he had held eight months ago. He had read two sentences, understood it was not for him, and closed it. No reply. No unsubscribe. Just a click in the open-rate column and silence.
That was the 32% explained. Good subject lines, wrong person. B2B contact records decay at 22–30% annually, driven primarily by role changes — a figure consistently reported across Salesforce State of Sales and HubSpot research. In a list built six months before a campaign launches, around 25 role-decay contacts is exactly what the decay rate predicts. Not bad luck. Arithmetic.
Company-Stage Mismatch
The third audit moved from contacts to the accounts themselves. Nadia's ICP was specific: B2B SaaS companies between 50 and 200 employees, Series A to Series B stage, with a dedicated sales team but no formal revenue operations function yet. The list had been built against that profile. But six months earlier.
She checked current funding stage and headcount for each account. Around 20 had changed status materially.
One had closed a Series B six months ago. Headcount was 340. They had a VP of Revenue Operations. The problem her product solved had been solved. By someone else. Her sequence had been running into a company that had already bought. The stage field still showed Series A. The record said one thing. The company was something else.
Others had raised new rounds past her ICP ceiling, been acquired, contracted significantly, or effectively ceased operations. Three touches per account, around 20 accounts: roughly 60 emails to companies where the commercial premise had already changed.
Intent Signal Decay
Part of the list had been built from intent signals — indicators that specific companies had been actively researching topics in her product category. When those signals were captured, they were real.
The problem was the timeline. She mapped signal capture dates against the sequence launch date. Around 40 contacts had intent signals more than 90 days old by the time her first email reached them.
In B2B SaaS, a buying cycle from initial research to vendor decision typically runs 30 to 90 days. A contact actively researching in March who receives outreach in June has, with high probability, already resolved the question that drove their research. Their titles were correct. Their email addresses were valid. Their companies were in the right stage. The only thing that had changed was the behavioural signal that had justified prioritising them — and there was no field in the enrichment output that showed signal age.
What the Numbers Said When She Added It Up
Some contacts appeared in more than one category. When she deduplicated, roughly 96 contacts were affected by at least one material mismatch.
She pulled Q3 engagement data and split it by mismatch status.
| Contact Segment | Count | Reply Rate |
|---|---|---|
| No material mismatch | ~204 | ~0.6% |
| At least one mismatch | ~96 | ~0% |
| Total (weighted) | 300 | 0.4% |
The problem had never been the message. The problem had been measuring message performance against a denominator that included 96 people the message could not have reached effectively regardless of how well it was written.
The Four-Check Framework She Built Before Q4
She did not rebuild the Q3 list. What she built was a pre-sequence audit process for Q4 — four checks, one per mismatch type, run before any contact entered any sequence.
-
Domain-Contact Alignment CheckFor every contact, resolve the company name to its current operational domain and compare it against the email domain in the record. Any mismatch goes to manual review before enrollment. On the Q4 list of 280 accounts, this flagged around 30 contacts in the first pass — contacts that would have become another cohort of confirmed-delivered, zero-received touchpoints.
-
Role Currency CheckFor every contact in a function-specific sequence, verify current title against LinkedIn before enrollment. Any contact whose current role has moved out of the targeted function gets re-evaluated or removed. This added 45 minutes to the pre-sequence workflow, removed around 18 contacts from enrollment, and rerouted 7 to a different sequence matched to their current role.
-
Company Stage ConfirmationFor every account, verify current funding stage, headcount range, and ownership status against a live source before any contact enters a sequence. Nadia used Crunchbase and LinkedIn company pages. It flagged around 12 accounts: some had grown past her ICP ceiling, others had changed ownership.
-
Intent Signal Age GateFor every intent-sourced contact, log the signal capture date and flag any signal older than 60 days at the point of enrollment. Signals between 60 and 90 days old receive a modified sequence that acknowledges the research context may have evolved. Signals older than 90 days move to a re-engagement track, not a primary cold sequence.
Four hours added to Q4 list preparation. Around 61 contacts removed from the primary sequence. The sequence launched with 219 contacts instead of 280.
What Q4 Came Back With
The sequences were not dramatically different from Q3. The copy was better — she had learned things about the audience — but not so substantially better that it explained a 7x improvement.
The improvement came from two places. Some of it was genuine craft improvement — better targeting, tighter messaging, accounting for roughly half a percentage point of the gain. The rest came from the four checks. From not enrolling 61 contacts who should not have been in the primary sequence. From measuring reply rate performance against a denominator of 219 people who were actually reachable — rather than 300 people, a third of whom were structurally unable to respond regardless of what the email said.
She wrote a one-page note for her SDRs and the data team. The last line:
"A data mismatch is not a list quality problem. It is a structural misalignment between what a record says and what is currently true. There are four types. Each has a different cause, a different check, and a different fix. Treating all four as one problem is why we spent three weeks rewriting copy that was never the issue."
The data ops friend got a message the same week. "You were right. It wasn't the message."