Which channels produced completed care?

A plan of care is a course of visits. A patient who attends once and drops out is worth a fraction of one who completes, and the difference between channels on that measure is large.

Ranking by billed episode rather than by new patient shows which channels produce patients who finish. That is usually the difference between a marketing budget that grows the clinic and one that just fills tomorrow's diary.

Most clinics treat referral as untrackable and advertising as measurable, which flatters advertising and hides the largest source of care.

Add the referral list to the lead file, each patient tagged with the referring practice, and it attributes like any other channel — including which referring practices actually produce patients who complete care, which is a conversation worth having with them.

A course of therapy is not one appointment, it is a plan of care running eight to twenty visits over six to twelve weeks, and the payer frequently decides how many of those visits will be funded. That ceiling is invisible to any advertising measurement, and it is not evenly spread across sources. Plans that authorize six visits and plans that authorize twenty arrive through different adverts, because the adverts that mention coverage attract people for whom coverage is the deciding factor.

This makes visit counts a misleading unit even inside the clinic. Two sources producing the same number of new patients can differ by a factor of two in funded visits per patient, and therefore in revenue, without either of them doing anything wrong. A clinic ranking channels on new evaluations is ranking them on the one part of the episode that every payer covers, and ignoring the part where the money actually accumulates. The gap compounds over an episode, because the visits a payer will not fund are the later ones, which is where a course of care either finishes properly or quietly stops. Clinics feel this as a retention problem long before they recognize it as a measurement problem.

Exporting completed, paid visits with a patient contact detail on each row solves it without any modeling. All the visits from one person join to that person and therefore to the inquiry that produced them, so the report ranks sources on episode value rather than on evaluations booked. No service type, diagnosis or body region belongs in the file: a contact detail, an amount and a date are all the match reads.

A growing number of practices run a self-pay line alongside the insured caseload: performance work, post-rehabilitation training, dry needling packages, sessions sold at a rate the clinic sets rather than one it is handed. The economics are completely different. A cash session collects in full on the day at a price the clinic chose, and an insured visit collects an allowed amount weeks later after a claim, an adjustment and sometimes an appeal. One of those two lines can be advertised at almost any price and still make sense, and the other cannot, which is the whole reason it matters that they are told apart in the report.

Advertising treats them as one funnel because they share a phone number and a website. That is the trap: a campaign producing cash-pay clients can be worth several times a campaign producing insured referrals at the same cost per inquiry, and no report inside an advertising account will ever say so, because the platform knows a form was submitted and nothing about what happened next. Clinics frequently underfund the profitable half for years on the strength of an even-looking lead count.

Seeing it does not need a column, and a payer column is not one to add: whether somebody's care was insured is health information, and CloseRev signs no business associate agreements. If the cash-pay line is the question, export only the cash-pay revenue, so the file carries nothing but a contact detail, an amount and a date, and the report ranks channels on that line alone. Cash work tends to produce fewer, higher-value visits, which shows up as value per sale — precisely the shape a volume-based dashboard is worst at seeing. It turns a vague preference for cash work into a number that can sit next to a budget line, and it can make the case for a dedicated landing page, which is usually the cheapest improvement available.

A surgeon tells a patient to go to a named clinic. The patient goes home, types that clinic's name into a search engine, clicks the advert sitting above the practice's own listing, and books. The advertising account records a conversion and claims the entire episode. Nothing in that record is false, and the conclusion drawn from it — that branded search is the clinic's most efficient channel — is completely wrong, because the referral did the work and the advert intercepted a patient who was already arriving.

This is the most expensive measurement error in outpatient therapy, and it is self-reinforcing: branded campaigns look efficient, so they get more budget, so they intercept more referred patients, so they look more efficient. Clinics have spent years bidding on their own name on the strength of it. The only way out is to know the referral source independently of the click, which means capturing it at intake and putting it in the file rather than inferring it afterwards. It is also the reason a first report in this segment often shows a smaller paid share than the clinic expected, and that smaller number is the useful one.

Where the intake form records a referring physician, add those patients to the lead file with the physician as their source, and that source is ranked beside paid media on collected revenue. A patient who appears both as a referral and as an ad inquiry is credited to whichever lead came first; the report does not show that overlap as a separate figure, but every match links back to the rows behind it. Only exact matches count automatically, which matters here because the wrong answer is one a clinic will act on for years.