Andrew Bramlett
[personal email]  ·  Memphis, TN
Admission-level attribution · paid-search control

I made paid search traceable to admissions. Then I made the result actionable in bidding.

Behavioral-health operators can usually see clicks, calls, and admissions, but not the line connecting them. I designed the attribution layer that joins first-touch call source to one CRM episode and a confirmed admission, then turns admission economics into bidding values.

4 of 6
eligible active ad accounts live
64%87%
verified admission-to-source coverage · 275 admits, 90 days
-25%/+15%
early CPA / paid-search admission signal in the first live account
Four-month directional result; presented as an early signal, not causal proof.
The control loop
Ad interaction
click and session captured
Tracked call
first attributable source
CRM episode
calls and EMR status unified
Admission
outcome confirmed, value set
MQL, SQL, and admission values return to bidding.
Attribution establishes what happened. Valuation and bid control follow from it.
Only the ad platform’s click ID and a value are returned; patient and treatment data never leave. Clearing that bar inside HIPAA and 42 CFR Part 2 is why few operators have this loop.
What leadership can now answer
Is paid search creating value?

Read acquisition cost against expected admission value at a sustainable spend level.

What does marketing really cost?

Compare all-in marketing cost with expected value and, when reconciled, realized revenue.

Where should the next dollar go?

Increase, hold, or reduce campaign budgets using admission-linked performance.

Attribution is the foundation. Valuation and bid control are its outputs.
Expected admission value is set by facility from average length of stay by level of care. MQL and SQL values reflect their observed probability of becoming an admission, and the episode is normalized so upstream signals are not mistaken for realized revenue.
Admission-level attribution
1 of 2
Andrew Bramlett
Admission-level attribution · diligence appendix
Diligence appendix

Three passes raised verified admission-to-source coverage from 64% to 87%.

A 90-day case study at one residential facility, selected because I suspected gaps. The EMR confirms admission; the CRM joins that outcome to the call history in one episode.

64%87%
Direct-linked coverage vs the verified three-pass union.
275 total admissions
90 pinned days
1 representative facility
Every counted linkage was corroborated against a real tracked call.

87% = admissions with a verified tracked call and non-outbound source / all admissions. The unlinked 13% lacked a defensible earlier source; no match in this figure is modeled.

Three-pass identity resolution
01Direct-linked

The call ID attached to the CRM episode, verified against the call log before counting.

02Backlinked

The CRM opportunity ID carried back onto call records, with deduplication inside the episode window.

03Phone-stitched

Patient and family phone fields matched to inbound callers, with shared-number collisions screened.

Attribution rule: primary credit goes to the first call source in a 30-day episode. Direct or outbound triggers a 180-day lookback for a prior attributable source; later touchpoints remain attached without replacing primary attribution.

How the episode is valued
Expected admission value

Facility-specific expected revenue, based on average length of stay by level of care.

MQL · qualified call

A call longer than 180 seconds, valued at its observed probability of admission, currently about 20%.

SQL · opportunity with VOB

Valued at its observed probability of admission, generally 40-50%.

Admission and normalization

The admission receives the residual episode value; a reporting multiplier corrects for upstream events that never convert.

Production and privacy boundary
Operating today

Daily in three accounts; weekly in one. Two use Maximize Conversion Value and two use Target ROAS in an active test.

Internal matching only

Only GCLID, action, time, and value return. No patient or treatment data leaves. Client counsel approved one implementation.

Walkthroughs of the methodology and the dashboard behind these figures are available on request: [personal email] · [portfolio URL].

System architecture, validation, dashboard: Andrew · CTM-CRM matcher: Jordan Marlow
2 of 2