Patient value and retention

Every other report in the catalogue measures a window: leads created this month, revenue won this quarter. These two measure a person — what they turn out to be worth, and whether they come back — which means they read forward past the window you selected. That one difference explains almost everything that looks odd about them at first.

Who counts as a patient

The clock starts when someone becomes a patient: whichever came first, their first won treatment or their first attended appointment.

• Enquiries that never converted are not in these reports. Conversion funnel and Lead sources already measure those.
• Patients imported from your practice software are included even with no enquiry record, because their attended appointments are enough.
• Someone who was merged into another record is counted once, under the record that survived.

Reading the cohort grid

Each row is the month a group of patients joined. Each column is how many months later. So the cell at row Mar 25, column m6 is that March intake, six months on.

The grid is a triangle, and the ragged edge is the point:

• Older rows are longer. A cohort from two years ago has lived through every column; last month's has lived through one.
• Empty cells mean "hasn't happened yet", not "nothing happened". A cell only appears once that month has fully finished, so a half-finished month is never compared against an older cohort's complete one.
• A visible 0 is a real zero — that month happened and nothing came of it.
• Values accumulate along a row. On Patient value the cell is everything that cohort has been worth per patient up to that point, so a row never goes down.

Columns are calendar months, not rolling ones. For a patient who joined on 1 January, m1 is February — which can be anywhere from one to sixty days after they joined, depending on the day of the month they came in.

Only the most recent 36 cohorts are drawn. If your window covers more, the card says how many older rows are hidden.

Why the tiles mostly have no comparison arrow

A cohort keeps earning after you measure it. Last month's patients have had a month to spend; the month before's have had two. Comparing them on value-to-date would flatter the older group every single time, and that is how most lifetime-value dashboards end up lying.

So the tiles split into two kinds:

• Value per patient (to date) — everything they have been worth so far. No comparison, because there isn't an honest one.
• Value per patient at 90 days — counts only what each patient spent in their first 90 days, and only includes patients who have had 90 days. This one compares, because both periods are measured at the same age.

Repeat treatment rate at 12 months works the same way: only patients who joined at least a year ago are in it.

What "value" means here

Two different measures sit side by side, and they will never match.

Recorded treatment value is the headline everywhere: the value of the journeys your team marked won. Where someone has filled in the confirmed amount, that is used; otherwise it falls back to the treatment's list price. The Confirmed-value coverage tile tells you how much of the total is a real confirmed figure rather than a price-list estimate — if that number is low, treat the rest as an estimate, and get the team filling in confirmed amounts.

Cash collected only appears if you have somewhere real to take it from — payments synced from your practice software, or Stripe deposits taken through SmileLine. It is never added to recorded value, because that would count the same treatment twice. If your practice software is connected, that ledger is used and deposits are ignored; if two connections both report payments, or a payment arrives in a currency that isn't yours, the tiles are hidden rather than showing a number that can't be trusted.

Why this doesn't match Revenue summary

It shouldn't, and the two answer different questions.

Revenue summary asks how much did we win in this window — each won journey counted on the date it closed.

Patient value asks what are the patients we gained in this window worth — every treatment those patients ever win, including ones that close long after the window ends.

They also differ on filtering: Revenue summary filters each journey on its own attributes, while Patient value filters the journey that brought the patient in and then counts everything they go on to do. A patient who came through Google and later had treatment on a different board still counts in full under Google.

Retention & recall

Returning means an attended appointment or a further won treatment, counted from the end of the course of treatment they originally came for — not from the day they joined.

That distinction matters more than it sounds. Someone who wins a treatment on the 5th and attends the appointment for it on the 20th has not come back; they are still on their first visit. Measuring from the day they joined would count that appointment as a return and show every cohort as almost perfectly retained, which is worse than useless.

The tiles:

• Returned by month 6 / 12 / 24 read the curve at that column, over the cohorts old enough to have reached it.
• Median time to return measures from the end of that first course of treatment to the next thing they did.
• Active patients are those seen — or booked in — within the last 18 months. Someone who joined last week is active.
• Lapsed patients are the exact opposite: no visit, no won treatment and nothing in the diary for 18 months. Archived and deceased records are left out.
• Value from lapsed patients is what those people were worth historically. It is not money at risk — you have already earned it. It is the size of the prize if you can bring them back.

Scheduling these by email

Both reports can be scheduled like any other. The digest carries the KPI tiles and the breakdown table; the cohort grid doesn't travel by email, so open the report itself to see it.