The CRM you don't trust is the CRM you report from

Thirty-seven percent of staff regularly fabricate data for leadership, according to Validity's 2025 survey of 602 organizations in the US, UK and Australia (Validity, 2025). Validity is a data-quality vendor, so its research comes with a point of view. The number still describes something recognizable: a report is due, the data does not join, and someone fills the gap.

For field marketers, the gap tends to open in the same place. The event happened. The guest list exists. The pipeline lives in the CRM. And the two do not agree.

How bad is the record

Validity's headline finding is that 76% of organizations say less than half of their CRM data is accurate and complete. Respondents said they lose an average of 16 deals a quarter to poor data, and that CRM users spend 13 hours a week hunting for basic information (Validity, 2025).

A year earlier, Validity's survey of CRM administrators ranked the problems: incomplete data (68%), missing data (65%), incorrect data (61%), duplicate records (53%) and expired data (49%). Forty-eight percent said customer data was decaying faster than the year before (Validity, 2024).

Other vendor research points the same way. Data and analytics leaders surveyed by Salesforce estimate that 26% of their organization's data is untrustworthy, and 49% say their companies occasionally or frequently draw incorrect conclusions from data (Salesforce, 2025).

What bad data does to an event report

Picture a single dinner guest making their way into your reporting. They register with a personal email address. Their company appears as "Acme" on the form and "Acme Holdings Inc." in the CRM. They changed jobs four months ago, and the contact record still shows the old title. The account executive never added them as a contact on the open opportunity.

Each of those is ordinary. Together they mean the guest never shows up where it counts. The attendance does not reach the account, so the account's open deal shows no event touch. The seniority mix in your recap is off. And the question leadership actually asks, whether the buying group was in the room, cannot be answered.

Multiply that across a season of events and you get a reporting gap that looks like an attribution problem but is really a matching problem. It sits inside a broader integration gap. Only 1 in 5 enterprises has integrated its primary event platform with the rest of its martech stack (Forrester, 2024), and 71% of organizations in a Forrester Consulting study commissioned by Cvent said disconnected event tools limit their visibility (Forrester Consulting for Cvent, 2026).

The temptation to fix it with AI

The obvious move in 2026 is to point an AI tool at the mess. Some of that is sensible. Models are good at proposing likely matches between "Acme" and "Acme Holdings", normalizing messy company names and spotting probable duplicates.

The risks are specific. A language model asked to fill a missing field will often produce a plausible value rather than a blank, and a plausible wrong title is harder to catch than an empty one. No model knows that someone changed jobs last week. Fuzzy matching can fold a subsidiary into its parent, or merge two people who share a name. And a one-time AI cleanup does nothing about the process that created the mess in the first place.

Bain, writing about sales AI projects, observed that sometimes as much as 80% of existing data has to be removed because it is old, inaccurate or confusing (Bain, 2025). Validity framed its 2025 report around the disconnect between data quality and AI implementation. And half of martech leaders told Gartner their technical or data stack is not ready for AI agents (Gartner, 2025).

Who owns this

The least technical finding may be the most useful. Thirty-five percent of organizations are unsure who owns CRM data accuracy, and 55% have no full-time employee dedicated to it (Validity, 2024). In the 2025 survey, 57% said they had moved to manual data cleaning while cutting dedicated data staff (Validity, 2025).

Event data sits in the worst spot in that ownership gap. Marketing collects it, sales relies on it, operations maintains the system, and nobody is assigned to reconcile it. So it falls to whoever builds the report, usually the night before it is due.

Practical moves

1. Measure your match rate. Take last quarter's attendee lists and count how many attendees match a CRM contact attached to the right account. That percentage is the honest ceiling on every event number you report. Put it at the top of the recap.

2. Fix capture before cleanup. Ask for a work email at registration and make company a required field. Check guests in against the list at the door, so the attended flag reflects who actually came, walk-ins and substitutes included.

3. Let AI propose and a person approve. Use a model to suggest matches, merges and name fixes, each with a note on why. Never let it write a job title or fill an empty field without a source. A reviewer signs off on each change before it touches the CRM.

4. Name an owner for attendee records. One person, with a deadline after each event (within five business days is a reasonable default) for every attendee to be matched, created or flagged. Put the name in the event brief so it cannot drift.

5. Report the gap instead of papering over it. Picture a recap line that reads: "Of 42 attendees, 36 matched to CRM accounts; the figures below cover those 36." That sentence costs nothing with a skeptical CFO and protects every other number on the page. The alternative, as 37% of Validity's respondents apparently know, is making something up.