How we got from chaos to a final product in 8 weeks

About two months ago, we got into Techstars. This story isn’t about that, though. It’s about doing more faster.

Published in

Making Mimo

8 min readAug 16, 2016

The first thing I noticed during our Techstars introduction is that all startups are at very different stages. While some of them already have dozens of employees, other startups are in the process of pivoting and starting from scrach.

At Mimo, we are somewhere in between. Our two proof-of-market apps have been on the App Store for over a year now. Swifty just passed the 1M download mark but, although it generates a decent amount of revenue through its in-app purchases, its customer lifetime value is too low to be a sustainable revenue stream.

Focus

In the long term, we want Mimo to become the leading source for mobile education, a place where you can dive into almost anything, in a learning style that is a lot closer to a game than old-school textbooks or video lectures.

Before Techstars, we were chasing a lot of different paths. We liked what we’ve created with Swifty and Javvy but we also felt that B2B was a thing. We wanted to release Mimo at the end of the summer but didn’t really know how to get there.

Here’s our initial plan:

Build a course creator that allows everybody to create courses in an interactive format and submit them to us
Build a tool that allows us to review the courses of the instructors in a timely manner
Create a marketplace and allow customers a way to rate and review the courses
Split the revenue with the instructors in a fair and secure way
Launch on iOS and Android devices (yes, that’s phones and tablets)

Then we had our initial deep dive with Rob, our Techstars Managing Director, and that was when things changed quite a bit. His question was:

What would the app look like if we were to release it four weeks before Demo Day?

Having Mimo out before Demo Day would give us the chance to show some actual data to investors. “Doing more faster” is just one side of the coin, we also had to figure out what we wanted to do faster. So we started taking the initial plan apart and looked for things we could remove. After two intense hours with Rob, the updated plans for Mimo looked like this:

Create all the content in-house and stop working on the course creator
Drop the revenue sharing system
Drop the rating system
Launch with ten courses on computer science and programming
Launch on iOS only and drop the Android version
Launch in seven weeks

Phew! To some of us, this sounded like a very drastic change, maybe even a pivot. Still, it also meant that we finally had a clear plan that we could execute on. 😃

Asking the people what they want

We decided on the number of ten courses for the initial release, but we still had to figure out which computer science-related topics we would go with. We had different ideas of figuring this out but, in the end, we decided to include a simple survey in Swifty.

The survey was obviously somewhat biased by the fact that Swifty’s users are iOS-only and mostly new to programming, which resulted in more advanced topics getting less votes than I would have thought (yes, we want to release more advanced courses in the future).

We also added an input field to the survey, which produced some odd responses

Because Mimo’s lesson format didn’t support images at the time, writing a high-quality course on Xcode/iPhone development wasn’t possible. So, sadly, we had to move the two topics that got the most interest back in the priority list.

Here’s the list of courses we decided to go for:

1. HTML
2. CSS
3. JavaScript
4. Swift
5. Java
6. Hacking
7. C++
8. C#
9. Ruby
10. Programming Basics
11. Python
12. SQL

Building a content pipeline

Great. We had a list of courses we wanted to release along with the app but, now that we didn’t look for external instructors to write the content, we had to find a way to create the entire curriculum in-house.

Swifty alone has more than 300 interactive lessons. A single lesson takes Henry, who’s in charge of content, around 15–25 minutes to create, creating ten additional courses within eight weeks seemed almost impossible (especially because the quality of the content was something we didn’t want to compromise on). That meant that we had to look for outside help as soon as possible.

We set out to hire a bunch of freelance developers with a desire to teach people. Not the easiest position to hire for in general but, with the added time constraint of finding someone in 1–2 weeks, it became a real headache. Thankfully, there are some advantages for Techstars startups and in the end, the Techstars network helped us find the right people.

In a matter of a week we hired four “curriculum engineers”, all from different countries and all worked remotely 🌎. Henry had to devote almost all of his time to managing the content team over Skype and Slack, a task that isn’t easy.

The course creator we built from scratch and used to write Mimo’s content with

Coming up with a business model

With the content team up and, well, writing, we also had to zero in on Mimo’s business model.

Here’s what the “business model” of Javvy and Swifty looked like:

Download the app for free
Get the first two chapters for free
Get the third chapter for free if you share a branded message on social media
Unlock all other chapters for $2.99

We had long discussions with Rob about the new business model and came to the conclusion that the old model worked fine for our stand-alone apps, but having to purchase every Mimo course on its own did not seem like a good way to go forward.

Our first intuition was to let the users decide between the one-off purchases to unlock a single course and a monthly subscription that would unlock the entire course library. But this approach has its flaws. Our courses differ in length, so we would have to price them all individually, which makes it more complicated for us to manage. The consumer on the other hand has to decide between the different purchasing options, which we deemed too confusing.

The other strategy that came to our minds was to offer a subscription that would unlock the whole content. We were hesitant to get rid of the “single-course purchases” because we were not convinced that users who wanted to learn, say, Swift would also want to learn other programming languages.

So we designed an experiment to test if users would be willing to go for a monthly subscription. We realized that, in order to get any real results, we had to actually charge the early-bird users.

Testing our assumptions

We built a quick landing page that displayed the different courses and features that we wanted to launch Mimo with. Then we linked to a Plasso “space” that allowed the users to choose between a monthly and yearly subscription. Since Plasso took care of the recurring payment handling and the Stripe charges, setting the experiment up didn’t take more than a couple of hours.

While all this was going on, we were also looking at a way to grow our mailing list. We started rewarding Swifty users that signed up as beta testers for Mimo with a dark coding theme, which worked really well. All of a sudden, we were collecting 2,500+ email addresses per week.

This mailing list allowed us to send the subscription experiment to a large number of potential users. We started by offering a batch of 3,000 people a subscription for a price of $2.50/month ($25/year) and limited the offer to 72 hours, so that people had an incentive not to wait around.

Once we looked at the results, we were surprised to see that 5.5% of the users who opened the email really subscribed to Mimo, with 41% of them going for the yearly plan. (Keep in mind that these are subscriptions for an app that was more than a month away from being launched.)

After the first experiment, we were pretty confident that users would go for a subscription if the price was right, but what is the right price?

To answer that question, we built another two landing pages and sent them to different batches of 3,000 Swifty users. This time, however, we priced the subscriptions at $7/month ($70/year) and $5/month ($50/year). Once again, the results were surprising.

At $7/month, the conversion rate dropped to 1% and we did not sell a single yearly subscription. We attributed this partly to the fact that $7 per month puts us in a price range where we are being compared to products like Netflix and Spotify. At $5/month, however, things looked a lot better.

After seeing these results, we started to feel a lot more confident about this change in the business model.

Here’s what we landed on (for now):

Download the app for free
The first chapter of every course is free, so you can get a glimpse of what’s to come
The second chapter of every course is free if you share Mimo on social media
Unlock all courses and chapters for $49.99/year

Sprint to launch

We figured out content creation and the business side of things, but we still had to build the app itself. Unfortunately, because of the different lesson format that we created to improve the learning experience, we could only reuse small parts of Swifty’s code base.

We hired two freelance iOS developers in order to launch on time, one of them is still working with us today (hey Frank 👋). Together, we spent many late nights in the office, trying to get the interactive lessons just right.

Joe and Henry at the Techstars office, well after 2 am

Although we didn’t quite manage to launch on time (Aug. 9th was the goal), we are still incredible content with the things we achieved over the last two months. There are still a few bugs but, from a user experience point of view and the amount of content, Mimo is far beyond anything Swifty ever was.