Livetweetapp

Why we’ve failed at the Lean Principles building a Live Tweet App so far.

10/05/2013 by Alexis Serneels

Are you launching a startup ? Like me, you think you’ve read enough about the Lean principles to start applying them ? Be careful, if you are not very much disciplined, you will eventually forget them and to put the cart before the horse.

We’ve experienced that. Here is our story.

The launch

All happened in two days, six months ago.

We had bought the domain name and I had finished the rapid development of our last project : a Twitter Wall web-application displaying moderated tweets based on a hashtag. The project was pretty much composed of a homepage, a registration form and a functional dashboard for search, moderation and display. The product was live, we could start welcome users.

Today, we have 304 registered users, 11 paying customers and generated $1,272 of revenue.
This is not a real success. That’s for sure.

It’s time I share with you how we’ve ended up here. And why we’ve decided to start all over again afterwards.

Do what I say, don’t do what I do

I’ve read them all. I do agree with all their principles.
From « The four steps of Epiphany » by Steve Blank to « The Lean Startup » by Eric Ries or even the first web-based version of « Running Lean » by Ash Maurya to printing the « Business Model Generation «  on A3 pages and attending workshops, I’ve spent hours understanding Lean Startup principles. I’ve been a good student.

But with our product, LiveTweetApp, every day I spent in the last six months, I haven’t applied any of the lean methodologies in the correct order. And believe me, I’ve not done any of this on purpose.

From Zero to a Minimum Viable Product

The decision to launch a Twitter Wall application actually came from one of our B2B clients (as we are firstly a web development company) who asked us if we could provide a Twitter Wall for their coming event. We’ve searched for existing solutions and mainly found expensive ones. So, we’ve decided that it would be cheaper (for the client) and better (for us) to quickly build this Twitter Wall application by ourselves.

And we did. We built it in-house, sold it to our client and it was a success.

Taking the wrong assumptions

From there, ignoring all Lean principles, we’ve taken this one-time positive feedback as a sign that if one customer loved our in-house Twitter Wall application, thousand of other people would have to love it too and were ready to pay for it.

And here we are, back to those two crazy days releasing our public version of LiveTweetApp.

From a Lean Startup point of view, we :

  1. Skipped the process of actually understanding the Problem. Our only problem hypothesis was « a client wants to display tweets on a screen during an event ». Our direct solution was a Twitter Wall, without further researches.
  2. We didn’t take the time to check if the problem was worth solving (by starting our new product anyway, we could have simply suggested our client another already existing solution)
  3. Didn’t build any demo neither done any interviews and went straight to release based on unverified assumptions.

Of course, we did some stuff right eventually.
But I’ve to admit they were implemented in a various order, without taking the time to learn anything from them.

  1. Users were able to upgrade (meaning : pay) their account from day one.
  2. We are measuring main data such as Acquisition, Activation, Retention. But numbers and figures are laying there, in Mixpanel, and we haven’t used any of them to rationalize decisions.
  3. Sent various one-to-one emails and collect feedbacks from registered users. This has been done though follow up emails and surveys as Qualaroo.

Where are we at this stage ?

Here is a good picture of our present situation. I’ve segmented it under the main Running Lean steps by Ash Maurya.

Problem / Solution Fit

  1. We haven’t clearly defined the demographics of an early adopter.
  2. We haven’t clearly validated if the real problem (emphasized strongly) was « not being able to display moderated Tweets on a screen during an event ». Is it a must-have ?
  3. We haven’t clearly defined the core features needed to solve the problem as we are still wondering internally if we should open our « Twitter Wall » to other social networks by example.
  4. We haven’t selected a price the customers are willing to pay. We still offer both monthly or one-time payments because we don’t really know which one our customers want. (We are back to the first question : who are our early adopters)

Conclusion : This is not absolutely surprising. After all, it’s clear we’ve skipped those foundation learnings months ago.
We have one good point though : registrations are coming daily (about 5 to 10 per day), early adopters are there and we can get in touch with them, analyze their background and start to write some hypothesis to test.

Problem / Launch Fit

This step is quite better of course, because we’ve jumped directly to it.
Here are the facts we learned eventually :

  1. Early adopters make through registration (acquisition) and displaying tweets (activation) workflows.
  2. Early adopters accept our pricing model. We’ve sold to 3% of them right ? Ok, this may maybe need some adjustments.
  3. We get around 70 unique visitors to our landing page and getting about 5 to 10 registrations per day.

Let’s launch ! Wait ! We already did that. What then ?

That’s our main problem and main point of this article. We should launch, learn, optimize and pivot. But at this stage we have no idea where to. Launch what ? What do we optimize first ? We’ve hit a wall.

This is only due to our lack of initial learning of our customers. Who are they really (for now they are just usernames and emails with an ID). What are they really looking for when registering for our product ?

Conclusion

Building a startup is exciting, appealing and you will be driven by the wish to scale faster and attract the most of users you can. You may end up skipping some very important steps to cut to the chase where in fact, you are not doing yourself a favor.

You should take the time to validate and learn, a lot.

And that’s what we are about to do from now. We’re going to start our building process from the beginning, taking advantage of the users and data we already have.

We will need your advices and suggestions. This post is the first of many. We want to log our learning and share them with our customers, with everyone. Communication is the key here.

Stay with us for the next article about Building a LiveTweetApp.
Don’t hesitate to comment below. I’ll read you from Brussels.