First results from the new StoryLearning App

Welcome to something new…

Olly in Maker Mode.

From here on, I’m going to focus this newsletter on documenting how I’m evolving my 13-year-old business, StoryLearning, by building a new AI-powered software product alongside it.

The first milestone goal: $1m ARR.

In my mind, if we can hit that in the next 12 months, it puts firmly on the table the possibility of pivoting the whole business to be the app.

If we can’t… well, at least I’ll have some war stories, battle scars, and a fine collection of empty whiskey bottles to show for it.

I have no idea if this is realistic, or even a good idea.

But I’ve got a decent track record with this kind of thing.

And either way, it’s going to make for good content.

Why this should matter to you

If you’ve run an online education business for any length of time, you’ve likely had a rocky few years.

Best case, you’ve felt the ground shift under you.

Worst case, you’ve seen a real drop in leads and sales, maybe even considered shutting up shop.

Either way, one thing’s clear to me:

Now is not the time for business-as-usual.

Now is the time to evolve.

I’ve been banging this drum for a couple of years now. Time to put my money where my mouth is.

Why software, why AI

For a business like mine, software has always been a theoretically smart move, because it adds infinitely scalable, high-margin recurring revenue.

(Plus, genuine equity value, if that’s on your radar.)

And anything with AI is a theoretically good idea, because swimming with the tide is a lot easier than swimming against it.

The hard part is figuring out how to add genuine, real-world value with this technology, rather than building something “powered by AI” just for the bloody sake of it…

which, let’s face it, is what most people are doing.

You’re looking to create a product that draws on your unique IP to create something genuinely unique.

And…

Touch wood…

I think that’s what we have here with the StoryLearning App.

The other thing is timing.

Why now?

Well, AI has made software development dramatically faster and cheaper, which removes the biggest historical risk of building software: the eye-watering cost of developers.

We’re at a strange moment where most consumes are still operating in a pre-AI world, but the cost for us to build in it has dropped by something like 80%.

That feels like a real window of opportunity that has made me decide to pounce now – in July 2026.

The backstory

About a year ago, I had an idea:

Build an AI-powered software version of my books.

These books are short story collections written in different languages and proficiency levels, designed to help people practise reading in a foreign language. They’ve sold over 2 million copies and are one of the things StoryLearning is best known for.

The success of these books comes down to the IP – the way I write the books. (i.e. they’re really good)

And my theory is that now, with the app, I can bring the exact same IP into the software product. (i.e. the app follows my process when creating stories)

Where software beats a book is that it’s malleable.

Story too long? We shorten it.

Too easy? We make it harder.

Want to drill a specific word in Spanish? We can bring that word back across your next five stories, so the more you read, the more personalised it gets.

Sure, a sophisticated learner with enough know-how could do a version of this themselves with ChatGPT.

But most people won’t.

Most people have neither the skill nor the inclination, and plenty would happily pay $15 a month to have it done properly.

Hence, why I’m making it.

Finding a partner and building the MVP

At the end of last year, I connected with my technical co-founder, an old contact who happened to be at a crossroads in his career.

When he heard the idea, he was straight in.

We did a deal and set a simple goal: launch an MVP within 6 months.

We could have gone faster, but there was no rush. He was working part-time, and we wanted to get the first version right.

For launch, we pulled together a tiny interest list from our most engaged customers and capped the first cohort at 100 people.

We launched about a month ago and those 100 spots filled in just over half an hour 🤯.

Needless to say, I had a fair amount of firepower at my disposal with my existing audience, so I wasn’t particularly shocked that I’d managed to fill a beta group, but the speed of it was still a great sign.

The MVP ran smoothly with very few issues, and it quickly became clear which features people were most excited about, and what was missing.

We spent the following month building those out in priority order, then reopened to the same interest list a couple more times.

That brought total subscribers to 315, out of an interest list of around 1,250, a roughly 25% conversion.

That’s the best list-to-paid conversion I’ve ever had for anything.

Pretty cool.

This brings me to the headline numbers as they stand today – one month in:

StoryLearning App — Key Numbers (as of 31 July 2026)

Active subscribers: 351 (280 yearly, 71 monthly)
MRR (prorated across all plans): ~$3,408/mo
Median daily reading time: ~17 min per active learner
Monthly churn: 19 cancellations (~5%)

Where we are now

Despite the early traction, we haven’t actually gone out to the main StoryLearning list yet.

Let alone to our YouTube channels, podcasts, affiliates, or anything else.

So there’s a lot more gas left in the can.

That’s next week:

A proper launch, hopefully bringing in thousands of people and taking the product well out of beta.

So far, it’s good news, but it’s also fairly predictable. After all, I’ve spent 13 years building an audience, and it’s at moments like this where that tends to pay off.

Selling to your own list is the easy part of any $15/month software product. Real revenue only comes with many thousands of subscribers, which means the product has to succeed in the open market, not just with people who already know and trust me.

(We’d need around 7,500 active subscribers to hit $1,000,000 ARR.)

There’s a long way to go.

What’s coming in this series

Going forward, I’ll be writing about the wider growth strategy.

This is new territory for me.

I’m experienced at marketing to my own list, far less experienced at growth marketing: going out into the open market and properly selling one product.

That’s never been my style, but it’s the job now if this is going to work.

And I want it to work, because while courses and coaching have proven more durable than I expected a couple of years ago, selling information is living on borrowed time.

I’d rather be swimming with the tide than against it.

So expect less about how I sold subscriptions to my own list (any tinpot marketer can write about that) and more about how I’m evolving an education business for the long term.

A few early lessons for you

Find a good technical co-founder. Vibe coding is fine for the most basic apps and dashboards, but for anythign else it’s a waste of time. Watching how our co-founder has architected the app for future growth has convinced me: get the tech right from day one, or you’re storing up technical debt that’ll be painful to undo later.
Talk to your customers from the start. Our biggest insights have come from 5-10 Zoom calls each week with paying customers, right from the beginning. You know less about your customer than you think until you do this, even if you hate the idea of getting on a call. It speeds up your ideation cycle enormously.
Make opinionated software. I borrowed this from 37signals’ Getting Real: boring and predictable kills software, so embrace the weird and wonderful. Anyone who knows the StoryLearning method will spot my fingerprints all over this app’s features. Take “strict mode”: toggle it on and it strips out every in-reading aid, like word translations, so all you can do is read. It runs against the industry-wide push toward gamification and bells and whistles, by taking all the options away…but the die-hard fans love it.
Launch before you’re ready. You’ve heard this one before, but we did this, and I’m glad we did, because it quickly showed us which features were duds and which were missing. And the answers came from paying customers who’d put their card in, not from opinionated bystanders or an overzealous founder.

What’s next

For the next few editions of this, I’ll probably report on how our internaal launches have gone, and where my focus is turning next.

I’ll share some of the metrics from our early users, including some genuinely striking in-app engagement and time-on-task numbers. That’s for another day.