What the build actually costs
Start with people, since that's where most of the money goes. A coaching platform that real clients depend on needs, at an absolute floor, three developers and one QA person.
That's not a study, it's arithmetic on public wage data. Do it with your own numbers and the shape won't change.
Then there's delivery risk. McKinsey and the University of Oxford's BT Centre for Major Programme Management studied more than 5,400 IT projects and found the large ones ran 45% over budget and 7% over schedule, while delivering 56% less value than predicted. Seventeen percent overran by more than 200%.
That research is from 2012 and covers projects above $15 million, so it isn't a direct read on your roadmap.
You don't just get the platform late. You get less of it than you signed up for.
The line item nobody budgets
Coaching data is health data. The moment you hold it, you own an access-controlled, encrypted, audited system that has to stay that way permanently, and that has to survive your clients' security reviews.
Nobody scoping a build in a conference room budgets for that. They budget for scheduling, messaging, and a dashboard. Compliance is not a feature you ship once. It's a standing obligation with an annual cost and a tail risk that can end a small coaching company outright.
Buying general software is not the same as buying a coaching platform
Most failed "buy" decisions aren't really buy decisions. They're a purchase of general-purpose software followed by a build project nobody planned.
If the platform doesn't arrive with the program already in it, you've bought a database. Five things have to be included, or the build you avoided has simply moved:
The touchpoint item is not a nice-to-have.
Almost the entire gap opened at sessions three and four, where coaching alone got two people to session three and none to session four, while coaching plus technology got 17 and 12. That's the capability you're deciding whether to build. It isn't scheduling.
The five questions that settle it
| Question | Build | Buy |
|---|---|---|
| Is the software what your customers actually pay for? | Yes | No, they pay for outcomes |
| Can you fund engineering, QA, and security permanently, not just through launch? | Yes | No |
| Is your coaching workflow genuinely unlike anyone else's, and is that difference why clients choose you? | Yes | No |
| Do you already own a behavior change content library? | Yes | No |
| Can you wait 12 to 24 months before a client sees anything? | Yes | No |
Most coaching companies answer "buy" to four of five and build anyway, usually because one internal champion is confident it's a six-month project. It is the 56% value shortfall that gets them, not the 45% cost overrun.
What buying looks like when it works
An insurance services company selling full-service wellness programs to its own clients moved coaching onto the Avidon platform in early 2021. It went from in-house coaching staff to one full-time coach delivering 2,849 sessions, up from 1,700, at roughly 30% lower cost.
Worth noting who that buyer is. They don't run a wellness program, they sell one to their clients, which is the same position most coaching companies are in. Read the full case study.
For scale context, Avidon's platform has been deployed more than 500 times across 50,000+ participants. Enterprise implementation runs two to four weeks, and pricing starts at $35 per month. Compare that against $700,000 a year in salary and a 12-month runway before a client sees a screen.
When building is the right call
It sometimes is, and pretending otherwise would be sales talk. Build when all three of these are true:
- The software is the product. You're selling seats, not sessions.
- You can carry engineering, QA, and security as permanent functions, through a bad quarter as well as a good one.
- Your workflow is genuinely unlike anyone else's, and that difference is the reason clients pick you.
The hard question isn't whether you could build it. Most competent teams could. It's whether the version you'd have in eighteen months would beat what you can license this quarter, and whether the eighteen months were the best use of the only engineering capacity you have.
