Introducing: the Miles & Mountains Coaching App!

Josh Fields
4 days ago
4 min read

Updated: 20 hours ago

A screenshot of the app, showing the week of training, color-coded, a comment notification, and subjective feedback. — The Training Log page of the app.

Why I built my athletes an app

A collaborative, customizable and flexible training log has always been at the heart of how I coach. It provides space for all the details and context that matter most to the athlete—and good coaching— while allowing us to trim out everything we don't need. Most importantly it prioritizes the collaborative nature of the coach-athlete relationship.

But for years it had one real weakness: it was big, and didn't work well at all screen sizes, particularly mobile devices. If you've ever tried to open a Google Sheet on your phone, you know the problem. Pinching, scrolling sideways, hunting for the right cell with cold hands at a trailhead... and sometimes it just doesn't load at all. And if you're an athlete with 1500 days in your log, all of them with conditional formatting? Forget about the official Sheets app loading that.

An app has been a someday-project for a long time, not only to gain mobile utility, but for general usability. It can show only what you need most of the time without presenting a barrier to things you still need occasional access to.

Why not shift to a platform?

The obvious move for most coaches would have been to switch to one of the big platforms. I didn't, and that route was never really in the cards for three reasons.

First, those platforms are built around data. So much of it, and so insistently, that eventually you're coaching the stats instead of the runner. Scroll social media for a few minutes and you'll see tons of runners stressing over what their watch is telling them, when in reality it's erroneous, irrelevant, or both.

Real coaching lives in the context of the training we do, not the statistics.

Pre-made platforms also flatten the personality out of both the coach and the athlete. They decide what you're allowed to track and what a week should look like. In my experience that costs you the two things that matter most: honest feedback, and training that actually fits the person doing it.

Plus, the coach and athlete have very little control over what happens in the future. If the app gets bought by venture capital, switches to AI-assisted coaching or worse — the only option is to leave and lose all the data in their plan and log.

What's in the M&M Coaching app?

But the training plan and log isn't the whole story—there's a lot more to my coaching. A strength and mobility exercise library with 50+ exercises, most with video examples. Trackable exercise routines. Athlete-exclusive articles. A glossary. Charts to track progress and places to keep track of races and goals. A handbook for crews and pacers.

Bit-by-bit, I'd built a lot of material that exists only for the people I coach, and it was scattered across links, tabs and my website. It wanted a home, and particularly someplace that made it easy, not just accessible.

I didn't want an app instead of my system. I wanted an app for it.

So I built one. Athletes add it to their home screen and it works like any other app. They can check today's workout before they head out, log their run when they finish, tick off their strength work or read an article on the couch. Almost everything I've made for my athletes is in there, with no login required. It just works.

It took awhile, but it's now been rolled out to all the athletes I work with.

Main training plan/log page
The mobility exercise Library
Log entry fields
Menu

Here's an example of how a few different sections look.

On the AI question.

While I designed the app entirely myself, I used AI to help write and refine the actual code. I'm a running coach, not a software developer. I've had some experience with a variety of aspects of HTML and Apps Script over the years, and came in with a solid starting point, but needed help to fully code the app in a robust way.

There wasn't another route to this. A build this size runs somewhere in the tens of thousands of dollars — and that only buys the first version. I'm still adding to it, and want to continue to do so. A solo coaching business can't keep a developer on call, which means the honest alternative wasn't a hand-built app. It was no app.

I don't reach for generative AI when a human does the job better, which is exactly why my AI usage stops at the code: the design and the coaching are the parts a person does better, and I've kept it that way.

Custom, careful building and testing

The building process was slow and deliberate. I didn't just prompt a couple times and take whatever it spit out. I made over 200 numbered versions across the app and the backend system behind it, and an automated test suite with over 900 tests at last count. That suite ran on every version, and each version had about a week of real athletes using it before anyone else saw it.

Nothing about this is a generic running or coaching app. It couldn't be. It's built around my specific way of coaching.

One thing that's important to note: every athlete's training is written by me. Every word of feedback in their log is written by me. AI has never written a workout, read an athlete's notes, or made a single decision about how they train — and it never will. AI isn't integrated into the app itself in any way.

Behind this app, unchanged, is the same collaborative training log I've always used. Still built around each individual athlete, and still the place we actually get things done. This is just a better way to see it.