What does it actually cost to acquire a telehealth patient?

Nobody publishes a credible telehealth customer acquisition cost benchmark, and most of the ones you will find were invented by someone selling you something. What is public is this. Hims and Hers spent $798.5 million on customer acquisition in 2025 and finished the year with 282,000 more subscribers than it started with, which is roughly $2,832 for each net subscriber added, against $81 a month in revenue per subscriber. That is not their customer acquisition cost, because they do not disclose churn and nobody outside the company can separate a new patient from a re-acquired one. It is the cost of net growth. In 2023 the same calculation gave $764.

The number quadrupled in two years. The rest of this page is where it came from, what it does and does not mean, and what it should change about how you spend.

Why is there no telehealth CAC benchmark?

This is the question most people arrive with, and it does not have an honest answer.

A benchmark is an average across companies. Those companies treat different conditions, at different price points, with different margins, through different channels, and they do not agree on what a customer is. One counts a booked visit. One counts a completed visit. One counts a subscription that survived a first refill. Averaging those produces a number with no referent. When an agency quotes you a telehealth CAC benchmark, it is quoting its own client mix, which is a sample of whoever happened to hire it, or it is making the number up.

Off-Label does not have a benchmark dataset, and this page is not going to pretend otherwise.

What does exist is disclosure. Public companies file, they get audited, and one of them is the largest advertiser in this category by a wide margin. Their filings will not tell you what a patient costs you. They will tell you what a patient costs the best-funded buyer in the auction you are bidding in, which turns out to be the more useful fact.

What do Hims's own filings say it costs them to add a patient?

The method here is not ours. Preston Alexander worked it out in The Healthcare Breakdown in April 2024: take the customer acquisition spend Hims discloses, divide it by the change in subscriber count, and you have the cost of adding a subscriber. He ran it on the 2023 filing and got about $785. What follows is his method with two more years of filings, and one correction to the inputs.

The correction, because two numbers are circulating and they are not the same number

Hims discloses acquisition spending in two different places, and they do not match.

In the management discussion, the company reports customer acquisition costs: $379.7 million in 2023, $594.5 million in 2024, and $798.5 million in 2025.

In the advertising footnote, it reports advertising costs for customer acquisition and content production, a wider figure that bundles in the cost of making the ads: $390.3 million in 2023, $604.6 million in 2024, $814.9 million in 2025.

The widely repeated $785 figure divides the footnote number for 2023 by that year's net adds. Most people quoting a recent Hims figure use the management discussion number. Crossing the two inflates the base year and understates everything that happened since. This page uses the management discussion series, the narrower one, throughout.

The three-year series

Cost per net subscriber added, from the Hims and Hers 10-K filings for 2023, 2024, and 2025:

The shape is the story, and the shape is not a slope. It is a cliff.

Through 2024 the machine worked the way a healthy paid engine is supposed to work. Spend 57 percent more, add 39 percent more subscribers, watch the cost per add drift up 12 percent. That is a company buying growth at a stable price, and it is what everyone in this category assumed they were watching.

Then 2025. Acquisition spend rose 34 percent. Net adds fell 59 percent. The cost of adding one net subscriber went to $2,832.

Before anyone draws a conclusion from that, here is what the number is not

It is not Hims's customer acquisition cost, and it should not be called one.

Hims does not disclose gross additions, and it does not disclose churn. Nobody outside the company can separate the money spent acquiring a patient who had never been there from the money spent winning back one who left. The $798.5 million bought both, in a proportion the filings do not reveal. Divide it by net adds and you get the cost of net growth, which is real and knowable, and narrower than CAC.

This matters enough to say twice: if you see this number quoted anywhere as Hims's CAC, including on a page that took it from here, it is being quoted wrong.

So why publish it at all?

Because the trend survives the ambiguity. Whatever share of that spend is churn backfill rather than new acquisition, that share did not change by a factor of 3.7 in two years. The level is uncertain. The direction is not.

And because both readings cost you money. If most of that money bought genuinely new patients, then the entry fee to this category is close to three years of revenue per patient, and you need a bankroll to pay it. If most of it bought back patients who had already gone, then the category is a treadmill, and you would be renting the same people forever at a rising rent. There is no third reading in which a new entrant whose only lever is paid media comes out of this fine.

The other half of the arithmetic

Hims got better at monetizing patients over these two years, and it was not close to enough.

Monthly online revenue per average subscriber went from $54 in 2023 to $81 in 2025, a 50 percent improvement, which most companies would take. Over the same period the cost of a net subscriber rose 270 percent. At $54 a month, a $764 subscriber pays back what it cost to get them in about 14 months of revenue. At $81 a month, a $2,832 subscriber takes about 35 months. That is revenue, not profit, and gross margin went the wrong way too: 82 percent in 2023, 74 percent in 2025.

A patient became half again as valuable and nearly four times as expensive.

One footnote on that series, in the interest of not being surprised later: Hims stated in the 2025 filing that it will stop reporting monthly online revenue per average subscriber after the first quarter of 2026. 2025 is the last full year of this number.

What actually makes a patient cheap?

If the largest and best-capitalized buyer in the category watched its cost of net growth quadruple, then "get better at paid" is not the lesson available here.

Paid media is the one lever every competitor also has, which is exactly why it gets bid to its clearing price. An auction does not reward the company that wants the patient most. It rewards the company that can pay the most, and that company is almost never the one reading this page.

Two things make a patient cheap, and neither is purchasable in an auction.

Differentiation

A patient who wants specifically what you do, and cannot get it from the next result down, does not have to be outbid for. The auction only prices attention that is up for grabs, and a patient who came looking for you was never up for grabs.

This is why undifferentiated telehealth is so expensive to buy patients for, and it has nothing to do with how many competitors exist. It is that when your promise is the same as everyone else's promise, the only remaining variable is price, and the only way to move price is to bid. How you build a position a competitor cannot copy in a quarter is its own problem, and it is where most of the return in this business hides.

Distribution

Distribution is any route to a patient that does not run through an auction. An audience that already exists. A referral base. An employer contract. A health system relationship. A partner who is already in the room when the need comes up.

It is unglamorous and it compounds. Every patient who arrives through it arrives at a cost that does not reprice against you every year, because there is no second bidder. The reason incumbents are hard to dislodge is rarely that their ads are better. It is that they built a way for patients to reach them that does not involve buying the patient twice.

And why both together is a moat

Either lever alone makes patients cheap. Both together is why the companies that have both are hard to compete with, and why a company with neither is not really competing with them at all. It is subsidizing them, bidding up the price of attention they also want, and losing the auction anyway.

What don't the filings tell us?

We do not know why 2025 broke. The filings show what happened, not why. It is tempting to reach for the obvious candidates, and this page is not going to. The FDA sent thirty warning letters to telehealth companies over compounded GLP-1 marketing. The manufacturers reset entry pricing and started selling direct. The compounding window closed. Any of those could have moved the auction, or the mix, or the churn, and a company's own operational choices could explain the whole thing without any of them. The filings do not adjudicate between these, so neither will we.

We do not know Hims's gross additions, so we do not know their CAC, so nobody does.

We do not know how much of the 2025 spend went to reacquiring lapsed subscribers, which is the single fact that would most change the interpretation.

If you read these filings differently, they are linked above, and the arithmetic is three lines long. Show us where it breaks and this page will change.

Frequently asked questions

What is a good customer acquisition cost for a telehealth company?

There is no credible published benchmark. An average across companies with different conditions, margins, channels, and definitions of a customer produces a number with no referent. Compute your own ceiling instead: what you can afford to pay follows from your patient capacity and your margin, not from an industry average.

How much does Hims spend to acquire a customer?

Hims and Hers disclosed $798.5 million in customer acquisition costs for 2025, and grew from 2,229,000 to 2,511,000 subscribers. That is roughly $2,832 per net subscriber added. It is not their CAC. Hims discloses neither gross additions nor churn, so the figure bundles new patients with re-acquired ones, and no one outside the company can separate them.

Why did Hims's cost per net subscriber jump in 2025?

The filings do not say. Acquisition spend rose 34 percent while net subscriber additions fell 59 percent, and the disclosures show what changed without explaining why. FDA enforcement over compounded GLP-1 marketing, manufacturer price resets, and the close of the compounding window all fall in the same period. None of them can be established as the cause from public filings, and this page does not claim otherwise.

Can you compete with Hims on paid alone?

An auction rewards whoever can pay the most, not whoever wants the patient most. Hims spent $798.5 million on acquisition in 2025 and still watched its cost per net subscriber added quadruple in two years. A new entrant matching them on the one lever they share is bidding up the price of attention it will lose anyway.

What matters more, differentiation or distribution?

Either one makes patients cheap to acquire, because both create demand that does not have to be won in an auction. Having both is what makes an incumbent hard to dislodge. If you have neither, that is the project, and it matters more than any campaign decision you will make this quarter.

What is a B2B partnership in telehealth go-to-market?

A route to patients through another organization's existing audience, rather than through an ad auction. Ro is a partner in Planet Fitness's PF Perks program, which Planet Fitness describes in its 10-K as brands paying it fees and commissions for offers made to its 20.8 million members. Affiliate deals work on the same principle. The cost scales with results instead of with what a competitor is willing to bid.