There's a conversation I have in almost every engagement. It usually starts with a marketing director or an executive saying something like: "We're getting leads. They're just not converting."
So we go to the data.
And what we find, more often than not, isn't a conversion problem. It's a tracking problem. Leads are coming in from five different channels. Some are tagged correctly in the CRM. Some aren't tagged at all. A few have lead sources like "Other" or "Unknown" because someone skipped the field. And nobody, not marketing, not admissions, not the executive, can tell you with confidence which channel is actually producing admits.
That isn't really a marketing failure. It's an attribution failure, and until you can see the problem clearly there's no way to fix it.
Attribution Is Not a Marketing Problem. It's an Operational One.
Attribution gets treated like a marketing function, and it is, partly. But in behavioral health, attribution is really an operational question: how did this person find us, and what happened to them once they did?
If your CRM doesn't capture lead source at intake, marketing has no feedback loop. They're spending money, on Google, on community outreach, on referral development, and getting no signal back on what's working. So they guess. Or they optimize for volume instead of quality, because volume is the only thing they can measure.
Meanwhile, admissions is receiving leads with no context. They don't know if this person came from a warm referral or a cold web search. They treat every inquiry the same way, regardless of what actually predicts conversion.
Both teams are doing their jobs, they're just working from different information, while leadership tries to make budget decisions from reports that don't reflect what's really happening.
This is what I mean when I say attribution is an operational problem. It doesn't really live in the marketing department. It lives in the system, or more often in the gaps between systems.
What "Dirty Data" Actually Costs You
Dirty data gets written off as an administrative nuisance, but the cost is operational, and it shows up in the quality of your decisions long before it shows up in a spreadsheet.
You can't pull accurate conversion rates. If lead source fields are inconsistent, if the same referral hospital is entered seven different ways by seven different people, your lead-to-admit rate by source is meaningless. You might be pulling 40% of your admits from one relationship and have no idea, because it's scattered across a dozen variations in the CRM.
You can't identify your best referral sources. BD teams without clean activity logs and source tagging can't tell leadership which relationships drive volume and which have gone dormant, so they end up managing relationships by feel. The failure belongs to the system, not to the BD team.
You can't make a defensible budget case. When a marketing director can't show which channel produced which admits, they lose the conversation with the CFO. The result is budget cuts based on gut feel, not data, which usually means cutting the thing that was actually working.
You can't spot the leak. Every pipeline has a point where conversion breaks down. Maybe it's inquiry-to-first-contact. Maybe it's assessment-to-admit. Without clean data at each stage, you can't isolate where leads are falling out. You just know they are.
The Three Places Attribution Breaks Down
In behavioral health specifically, I see attribution fail in predictable places.
1. At intake. The lead source field gets skipped. Or it gets filled in with whatever's easiest, "referral," "online", without the specificity that makes it useful. One person entering "St. Clair Hospital" and another entering "physician referral" for the same source destroys your ability to segment later.
The fix is structural: required fields, defined picklist values, no free-text lead source entry. If someone can type anything into the field, they will. Lock it down.
2. At the marketing-to-admissions handoff. Marketing passes a lead. Admissions receives it, but the source data doesn't travel with it, or it gets overwritten, or it never gets attached to the eventual admit record. By the time you try to run a source-to-admit report three months later, the connection is gone.
The fix is process: define exactly what information must transfer at the handoff, confirm that it attaches to the record, and spot-check regularly.
3. In BD tracking. Business development is often the least systematized function in the revenue operation. Referral sources get tracked in spreadsheets, or not tracked at all. There's no consistent log of which conversations led to which referrals. When you try to calculate referral conversion by source, who referred the most, who referred the highest acuity, who hasn't sent anyone in 60 days, you're working from memory instead of data.
The fix is a BD activity framework: defined logging standards, required fields for every referral contact, and a weekly review cadence that keeps the data current.
What Good Attribution Actually Looks Like
When attribution works, the day-to-day experience is mostly one of clarity rather than complexity.
Marketing knows which channels produce admits, not just leads. They can make a case for budget based on actual ROI, not impressions or click volume.
Admissions knows where each lead came from and can route accordingly. A warm physician referral gets a different response than a cold web inquiry. Speed and tone adjust to context.
BD has a tiered view of referral sources, which are active, which are lapsing, which are producing high-conversion referrals versus high-volume referrals that don't convert. They know where to spend time.
And leadership can look at a dashboard that connects marketing spend to pipeline activity to admit outcomes, without wondering whether the data is accurate.
None of that comes from the technology itself. It comes from process, because the CRM only holds what the team puts into it.
If You Don't Have Clean Data, That's the Finding
One thing I tell every client at the start of a diagnostic is that a lack of clean data isn't a problem to work around. It is itself the finding.
The absence of accurate attribution data is itself a signal, that the system hasn't been built with reporting in mind, that ownership of data entry has never been clearly assigned, or that nobody has made the case for why it matters.
Making that case is usually the first job.
Once leadership understands what they can't see, and which decisions they're making blind, cleaning up the data becomes an easy case to make. It stops looking like administrative housekeeping and starts looking like what it is, the foundation under every revenue decision the organization makes.
Orbital Behavioral Health Partners works with behavioral health operators to build the revenue systems, data infrastructure, and operational clarity that drive consistent admissions performance.