Key takeaways
- The sequence matters more than the checklist. Pharmacy and LegitScript have external lead times, so they start first, not last.
- Corporate structure comes before everything. An MSO structure is what lets a non-clinician own the business in states with corporate practice of medicine rules.
- LegitScript certification is what unlocks paid acquisition. Without it, Google, Meta and TikTok will not run prescription ads.
- The single most expensive decision is who holds the merchant account, because it determines what leaving costs later.
- Build time is not the constraint. Fulfillment contracting and certification are, which is why platform choice compresses a five-week launch into five days.
Decide what you are actually selling
Before structure, before software, decide the category. Weight management, testosterone, menopause, sexual health, hair loss, peptides and longevity are not variations on one business. They differ on whether the initial visit can be asynchronous, whether labs gate the first fill, how often a patient reorders, and how tightly the category is scrutinised.
That single decision cascades into everything downstream. A testosterone brand needs a synchronous visit and lab ordering built in from day one because testosterone is a DEA Schedule III controlled substance. A hair loss brand needs neither, and would be wasting money on both.
Pick one category and get it working before adding a second. Operators who launch three categories at once usually discover they have built three half-configured funnels rather than one that converts.
Get the corporate structure right first
Many states apply corporate practice of medicine doctrine, which restricts non-clinicians from owning entities that practise medicine. The standard resolution is a Management Services Organization structure: a professional entity owned by licensed clinicians provides the care, and a management company owned by you provides technology, marketing and administration under a services agreement.
This has to be genuine rather than cosmetic. Agreements that let the management company direct clinical judgment defeat the purpose entirely and recreate the exposure they were meant to remove.
Do this before you build anything. Retrofitting a corporate structure after you have patients is considerably more expensive than getting it right while the entity is empty.
Line up providers licensed where your patients are
Telehealth is regulated by the patient's location, not the clinician's. A provider licensed in Florida cannot treat a patient sitting in Ohio without Ohio licensure, regardless of where your company is incorporated.
For a national brand that means a credentialed network with coverage in every state you intend to advertise in, plus the operational discipline to assign each case by patient state. Credentialing is a standing operation, not a one-time task: licences expire, renew and lapse, and coverage gaps show up as unexplained decline rates in specific states.
- Verify state licensure, DEA registration where applicable, NPI, board certification and malpractice coverage before a provider sees a patient.
- Assign cases by patient state automatically. Manual routing fails at exactly the volume where it matters.
- Track which states you have thin coverage in, because that is where your funnel quietly leaks.
Solve fulfillment before you think it is urgent
This is where most launches actually stall, and almost nobody expects it. Contracting a compounding pharmacy, qualifying a cold-chain vendor for temperature-sensitive therapies, and building the integration takes weeks. Then you discover your compound cannot ship into eleven of the states you were about to advertise in.
A single pharmacy relationship also means you inherit that pharmacy's outage, that pharmacy's state map and that pharmacy's price. A routed network across many pharmacies makes a backorder a routing event rather than a stoppage.
Start this first. It has the longest external lead time of anything on the list except certification.
File LegitScript on day one, not at the end
Google, Meta and TikTok all require LegitScript certification before they will run ads for prescription products. Without it, paid acquisition is effectively closed to a direct-to-consumer telehealth brand.
Filed unaided the process commonly runs three to six months, and most of that is rebuttal round trips: LegitScript comes back with questions, and each unanswered cycle costs weeks. Filed as a managed process alongside your build, it runs in 7 to 14 days.
The mistake is treating it as a launch-day task. It has an external clock you do not control, so it should start when the build starts, running in parallel rather than in sequence.
Decide who holds the merchant account
This is the decision with the longest tail and the least attention paid to it at signing. If the platform is merchant of record, the patient's card is charged into the platform's account, and the platform remits your share on its own schedule against its own count of active patients. You are paid a fee out of revenue you never received.
If your brand is merchant of record, the money lands in an account in your name and you pay the platform a fee out of it, which is the ordinary direction for a vendor relationship.
It matters on exactly three days: the day you want to leave, the day you are acquired, and the day you want to renegotiate your own processing economics. On each of those it decides everything, because whoever holds the stored card tokens holds the ability to keep billing your patients.
Only then, buy traffic
Traffic is last for a reason. Every dollar spent before certification lands, before fulfillment routes cleanly, and before intake screens properly is a dollar spent converting patients you cannot serve, cannot ship to, or will have to refund.
Screening before checkout is the specific discipline that separates a clean funnel from an expensive one. An ineligible patient who is declined before a card is captured costs you nothing. The same patient charged first becomes a refund, a processing fee and a dispute risk.
Five days, because the slow parts are already done
The sequence above takes most operators four to six weeks, and the delay is almost never software. It is pharmacy contracting and certification, both of which have external clocks.
PharmaBro opens a full branded clinic in 5 days because those clocks are already run: 30+ compounding pharmacies pre-integrated and routing, providers credentialed across all 50 states, and LegitScript filed in parallel with your build at $0 rather than sold as an add-on. Payments settle into a Stripe account your brand owns from day one.
Conclusion
Starting a telehealth business is not difficult in the sense of being technically hard. It is difficult in the sense of having six dependencies with different lead times, where getting the order wrong costs you a month.
Do structure first because it is cheapest when the entity is empty. Do pharmacy and certification early because they have external clocks. Decide merchant of record deliberately because it is the hardest thing to change later. Buy traffic last, because everything upstream of it determines whether that spend converts or refunds.
Frequently asked questions
How much does it cost to start a telehealth business?
Platform costs vary enormously by model. PharmaBro publishes $15,000 to $50,000 one-time setup and $1,500 to $5,000 a month with a 3% to 1.5% transaction fee. Many platforms take a percentage of billings instead, from the high teens to roughly half of patient revenue, and many do not publish pricing at all. Beyond the platform, budget for entity formation and healthcare counsel, and for acquisition.
Do I need to be a doctor to start a telehealth company?
Generally no. An MSO structure lets a non-clinician own the management company while a clinician-owned professional entity provides the care under a services agreement. This is what satisfies corporate practice of medicine rules in the states that apply them.
How long does it take to launch?
On PharmaBro, 5 days for a full branded clinic, with LegitScript landing in 7 to 14 days in parallel. Assembling the same stack independently commonly takes four to six weeks, and platform implementations elsewhere in the category commonly run 30 to 60 days.
What is the most common reason a launch slips?
Pharmacy fulfillment, followed by LegitScript. Both have external lead times you do not control, and both are routinely scheduled last because they feel like paperwork rather than product. Start them first.
Anmol SethiFounder
Ran direct-to-consumer telehealth brands and paid a revenue share on every dollar they earned. Built PharmaBro as the infrastructure he wanted to buy.