Which channels produced paying clients?

A client on twenty hours a week for eight months is worth tens of thousands of dollars. Judged on the first invoice, the channel that produced them appears unaffordable and gets cut.

Because the match is on the client, every week of billing credits the original channel — which is the only ranking that reflects the actual economics of the service.

Hospital discharge, case managers and physician offices produce a large share of clients, and agencies almost never know which of those relationships is actually working.

Referred clients added to the lead file, tagged with the referrer, rank referrers on billed care, which tells the community liaison team where to spend their week.

Almost every agency runs two recruitment campaigns out of the same marketing account: one aimed at families looking for care, and one aimed at carers looking for work. They share a budget, an agency, often a landing page template, and always a monthly report. They are not comparable in any useful way. A client inquiry produces revenue; a caregiver application produces the capacity to accept revenue, and in most markets it is the caregiver side that determines how much business the agency can actually take on.

The confusion is expensive because the metrics look alike. Both funnels report leads, cost per lead and conversion rate, so a single dashboard averages a caregiver application and a family inquiry into one number that describes nothing. An agency reading it concludes that leads are getting cheaper or dearer overall, when what has happened is that one of the two funnels moved and the other did not. Meanwhile the decision that matters — whether to spend the next thousand on clients or on carers — is made on instinct.

Keeping the client side on its own footing is what makes the comparison honest. Export the invoiced care against the inquiries that produced it, and the client funnel is ranked on revenue rather than on lead volume, which is the only basis on which it can sensibly be weighed against a recruitment campaign that produces no revenue at all. Separating the two may show the correct answer changes by season, while both have been funded at a fixed ratio for years.

The hourly rate an agency bills varies substantially by who is paying. Private pay families are invoiced at the agency's own rate and settle quickly. Long-term care insurance policies reimburse against a policy's terms, often after a claim process that takes weeks and sometimes requires the family to pay first and be repaid. Programs funded through veterans' benefits or a state waiver run on their own rates and their own paperwork. Same carer, same visit, three quite different amounts and three quite different collection timelines.

Sources are not evenly distributed across those categories, which is what makes it a channel question. Adverts written around cost and affordability attract families looking for a funded route; adverts written around quality and continuity attract private payers. An agency ranking channels on hours delivered or on clients started is ranking them on a unit that conceals a spread of thirty or forty per cent in what those hours are worth, and the direction of that spread is stable enough to plan around once anybody measures it.

The fix is in what gets exported, not in a column. A payer-type column would be health information, and CloseRev signs no business associate agreements, so it stays out of the file. Where advertising competes only for private-pay families, export only private-pay invoices, so the file carries a contact detail, an amount and a date and nothing else, and the ranking is computed on money collected from the book the campaigns are actually fighting for. The channel ranking on collected money may already show that the cheapest source of inquiries is the most expensive source of revenue, and that the campaign everybody wanted to cut was bringing in the families who pay the full rate. The rate is a fact the agency already holds in its billing system, and it has simply never been asked to sit next to a channel name.

A daughter three states away searches at eleven at night, rings in the morning, and arranges care for a parent who does not use the internet and may not know the conversation is happening. The inquiry record therefore carries the daughter's mobile number and email. The billing record carries the client's name and address, and often a landline nobody answers. Those are two different people with two different contact details, and a naive join between them finds nothing at all.

Guessing would be worse than finding nothing. A system willing to link records on a shared surname, an overlapping address or a nearby date would manufacture attribution out of coincidence, and in a segment where one client represents tens of thousands of dollars of invoiced care, a single wrong link moves a channel's ranking visibly. So only exact email and phone matches count automatically here; the one weaker link the engine produces, a national number under a different country code, waits for somebody in the office to confirm, and records that share no contact detail are not linked at all.

The countermeasure is at intake rather than in the software, and many agencies are already halfway there. The inquiry form usually asks who the care is for; what it rarely does is put a contact detail the billing record will also carry on the inquiry row. Recording the client's own phone or email there, or making sure the responsible party's details appear on the invoice as well, is the single largest improvement available to the match rate in this segment, and it costs one column in a report somebody already runs. The missing field was never a technical limitation.