The scenario below is a composite drawn from patterns we've seen repeatedly across healthcare data operations — not a real client, not a specific engagement, but a fair representation of how these projects tend to go.
The Situation
A mid-size medical device manufacturer decides its CRM is a mess. Sales reps are working off contact records that are two or three years stale — practice affiliations that no longer exist, physicians who've moved to different health systems, titles that were accurate when someone typed them in but haven't been touched since. Marketing ops leadership pitches a data enrichment project: pull in a vendor, match the existing database against a cleaner source, fill in the gaps, standardize the fields.
Everyone agrees it's overdue. There's budget. There's a kickoff call. For about three weeks, it looks like a success story in the making. Match rates come back respectable. Fields that were blank now have values. The enrichment vendor delivers exactly what was scoped — a refreshed file, mapped to the CRM schema, ready to upload.
Then the project quietly stops mattering.
What Didn't Work
The first problem wasn't the enrichment itself. It was that the project had been scoped as an event rather than a process. Leadership approved a one-time cleanup, budgeted like a software purchase rather than an operational commitment. Once the enriched file landed, there was no owner assigned to keep it that way. The vendor relationship was transactional — deliver the file, close the ticket.
Within a few weeks, the usual entropy started reasserting itself. Physicians changed affiliations. A regional health system absorbed two independent practices, and none of those org-chart changes were reflected anywhere. Reps started noticing discrepancies — a contact enriched with a title that didn't match what showed up on a call — and instead of flagging it, they just stopped trusting the new fields altogether. That's the real damage: it wasn't that the data went stale, it's that the reps went back to their own spreadsheets and personal notes, because at least those felt current.
Second problem: nobody had defined what "enriched" actually needed to mean for this specific sales motion. The vendor filled in standard firmographic and provider fields — specialty, NPI, facility affiliation — but the device sales team's real bottleneck was knowing which facilities had budget authority centralized versus which ones let individual department heads sign off. That distinction never made it into scope because nobody on the buying side had articulated it clearly at the outset. So the enrichment was technically correct and practically underwhelming.
Third, and maybe most common: the enrichment sat on top of a CRM with inconsistent record-creation habits. Duplicate accounts, contacts logged under old facility names, some records manually entered by reps with abbreviations only they understood. Enrichment can't fix a structural mess underneath it — it just gets absorbed into it. A clean data file merged into a messy database produces a slightly less messy database, which nobody experiences as improvement.
By week five, the project had a Slack channel nobody posted in and a dashboard nobody opened.
What Changed
The turnaround didn't come from a better data source. It came from treating enrichment as an ongoing operational discipline instead of a deliverable, and from getting specific about what the sales team actually needed to act on.
The team pulled in a small group of frontline reps — not just marketing ops — to define which fields actually changed behavior. It turned out reps didn't care that much about a generic specialty tag; they cared about knowing whether a target account had recently changed its procurement process, and whether the physician they'd been calling had moved practices. Those are harder signals to enrich for, but they're the ones that actually get used.
They also assigned an internal owner — not a vendor, not a committee, one person — responsible for triaging discrepancies reps flagged. When a rep noticed a contact's facility was wrong, there was now a two-day turnaround to fix it, instead of the correction disappearing into a ticket queue. That single change rebuilt trust faster than any accuracy improvement did. Reps started reporting bad data instead of quietly ignoring it, which meant the feedback loop actually closed.
On the technical side, they stopped treating enrichment as a monthly or quarterly batch job and moved to a cadence tied to specific triggers — new account creation, deal stage changes, a set number of days since last verification — rather than a blanket refresh. That made the ongoing cost more predictable and made it easier to justify as a recurring line item instead of a one-off project that needed re-approval every time.
They also brought in an external partner, NPLUS Global among others considered, specifically for the harder-to-source layer — affiliation changes and organizational structure shifts in healthcare systems, which move faster and are harder to track internally than most CRMs can keep up with on their own. That wasn't the whole fix, but it filled a gap that internal data entry was never going to close on its own.
The Outcome
None of this produced a dramatic before-and-after chart. What changed was quieter: reps stopped bypassing the CRM fields and started updating them, because corrections actually got made. Marketing stopped running campaigns off segments that quietly excluded half the accounts due to mismatched facility names. The sales cycle didn't get shorter overnight, but the number of "wait, this contact doesn't work here anymore" conversations dropped noticeably, and that alone bought back real selling time.
The bigger shift was cultural rather than technical. The company stopped asking "is the data clean" as a yes-or-no question and started asking "who's responsible for keeping it usable this quarter." That reframing is unglamorous, but it's the difference between a project that stalls in month two and one that's still delivering value a year later.
Most enrichment failures aren't data problems. They're ownership problems wearing a data costume.
Ready to see what we can build for your ICP?
Send us your ICP — sample in 2–3 hours, full delivery in 48–72 hours.
Request a free sample →