TL;DR
Record the referring relationship, the patient's reported discovery and the observed contact route in parallel. An endodontic patient can receive a dentist's recommendation and still use Google to find the office. Preserve that combined journey rather than awarding all credit to the final click or deleting search from referral-led cases.
Ask what each source field means
An endodontic owner asking whether referrals or search are growing needs a definition of each. A referring dentist identifies a professional relationship or referral provenance. A person's answer about discovery reflects what they report. A website session or telephone destination describes an observed route under the practice's setup. These can coexist for the same request.
Do not compare a count of referring offices with a count of website sessions and call the larger series the winning channel. Offices, transmissions, inquiries and appointments are different units. Begin with the business question: are more offices sending appropriate inquiries, are people finding accurate information, or is the practice receiving additional patient-initiated requests?
Use the measurement dictionary for the general event definitions. This report adds the referral relationship as a separate dimension so an agency's acquisition report does not erase a dentist's role in the patient's path.
Build a parallel-source report
Keep detailed reconciliation in the practice's approved system. A marketing report should use only data and summaries approved for its recipient and purpose. Do not export referral letters, diagnoses, tooth identifiers or personal search histories to fill a source field.
| Reporting dimension | Permitted finding when supported | Keep separate from |
|---|---|---|
| Referral provenance | Known referring office, no recorded referrer or unknown | The public route used to contact the practice |
| Patient-reported discovery | Dentist, search, another influence, mixed or unsure | Technical attribution inferred by software |
| Observed entry route | Measurable website route, telephone route or unknown | A complete record of earlier influences |
| Request identity | Practice-reconciled distinct request and repeat-contact status | Raw documents or event totals |
| Administrative outcome | Contact, consultation request, confirmed visit or unresolved | Clinical acceptance or treatment outcome |
| Observation coverage | Period, cutoff, missing fields and setup changes | An assumption that missing equals absent |
For management, create grouped categories such as referral known with search reported, referral known without search reported, no referral recorded with search reported and unresolved. These are reporting recommendations, not standard clinical classifications. Absence of a recorded referrer does not establish that no dentist influenced the decision.
Preserve disagreement instead of repairing it by guessing
If the person says a dentist recommended the practice but analytics shows search, retain both observations. The search may have helped them verify the address or understand the contact process. The evidence does not establish that the ad created a new patient independently of the referral.
If staff record a referring office while the person reports another influence, check definitions before treating it as an error. The office field may identify where documents originated, while the person is describing why they chose to inquire. Correct demonstrable entry mistakes through the approved process; do not force agreement simply to produce a tidy chart.
For Google Analytics, a session represents a period of interaction, and a person can start later sessions. Session totals therefore should not be renamed prospective patients. Google's session documentation.
A report showing sessions and referral requests should state their coverage separately. Do not divide unrelated totals to create a conversion rate. If the available systems cannot connect an approved inquiry outcome to a visit, report the gap plainly.
Keep the source question neutral
If the practice already asks about discovery, have operations and privacy owners approve any revision. A useful optional question allows uncertainty and several influences. It should not ask the person to confirm that the agency's AI work brought them in.
Avoid asking for screenshots of private assistant conversations or detailed health-related search prompts. The owner can examine broad discovery patterns without collecting those materials. The neutral source-question guide supplies the general approach.
For referral provenance, define who may update the field and what evidence supports it. Do not let a campaign parameter overwrite a known referring relationship. Similarly, do not automatically relabel a patient-initiated request as referral-generated merely because a dentist later sends records.
Use the report to decide the next improvement
If known referring offices send requests but people repeatedly use search to clarify basic logistics, examine the public information and referral instructions together. That is a service-information task, not proof that search deserves exclusive acquisition credit.
If patient-initiated inquiries ask for services outside the practice's verified scope, have qualified staff define the actual mismatch and review the promoted message. If referring-office activity changes, examine the actual office relationships and any process changes before attributing the movement to content publishing.
A recurring report should show actual categories, changes in missingness and the relevant outcome stage. Preserve the earlier definition if the method changes, or mark the comparison break. Rewriting historical fields to make the new method appear continuous conceals what the practice knew at the time.
End with a decision record: source dimension under review ___; evidence ___; unresolved connection ___; proposed operational or marketing change ___; owner ___; next observation ___. A parallel report is valuable when it changes a decision, even if it cannot assign exclusive revenue credit.
Frequently asked questions
Can a referral and search both receive credit?
They can both be recorded as influences when supported. Keep the reconciled total of distinct requests separate from channel participation so the combined report does not count a request repeatedly.
Should branded search be removed from the agency report?
Keep it visible with its actual meaning and limits. A referred person using a branded search is useful journey evidence, but it does not prove the agency independently created the referral.
What should we do with unknown source records?
Report them as unknown and inspect whether the approved collection process is working. Do not allocate them proportionally to channels to make the report look complete.