lean-mobile-startup-slides

Hello

Hands-on Entrepreneur and Tech Executive.

Now: Building fun, engaging apps at Levitum.

Past: Multiple startups. One Valley IPO. One exit.

Polyglot. Android + iOS + Python.

Program Chair, Droidcon India (2011-13).

Built apps with > 2 million downloads, 4+ ratings.

Are You a Lean Mobile Startup?

Arvi Krishnaswamy

@twitortat

arvi@alumni.iastate.edu

Not all Flappy Endings

The average app expends 9-12 man months of effort. #fact

The average app gets < 1000 downloads. #fact

~22% of downloaded apps are opened just once. #fact

An Untold Story

Vijay has this great idea for an app. He spends the next 6 months over nights and weekends building the app. He then publishes it to the Play Store and gets friends & family to download and check it out. All of them say the app is nice. Vijay celebrates, and starts to plan on ways of planning a roadmap, monetizing the app, and growing the user base. Downloads quickly climb up to 500. But then, they slowly taper off. They flatten at around 1000. Frustrated, he tries to make the app focused on a niche, and explores freemium and in-app ads. As more complexity is added, the existing active users leave. Vijay keeps adding new features his friends suggest, keeps trying to polish the user experience. But he is unable to find users for his app.

What can we learn from Vijay's story?

Vijay built an app no one wants

Vijay built an app no one wants to solve a problem no one had

Stop wasting time & start really learning

Key Lean Principles

Start by asking the right questions

What are the right questions?

• Who is your target user?
• What is your core value proposition?
  • If your apps main differentiator is experience, what is the core unique experience.
• What are the major leap of faith assumptions you are making.
• How can you validate these assumptions?

A Prototype is an Experiment that helps Validate Assumptions

Prototype for Early Validation

• Start with mockups or an explainer video. Or, even just paper & pen sketches which you share with users.
• Separate experience related feedback from the core value proposition
  • HTML5 prototypes with a degraded experience can help validate the core proposition.
  • Early adopters are typically willing to work around teething issues.
• Adopt a concierge approach.

The POP app.

• Use pen and paper to quickly sketch mockups
• Use the POP app to take snaps of the sketches
• Stitch the snaps together with a navigational heirarchy.
• Share your demo with users for feedback.
• For early experiments, focus heavily on interaction design, but deemphasize visual design.

How would you build an Ola Cabs app MVP?

Do Things

That

Don't Scale

Yes, you read that right. Counter intuitive?

Accrue technical debt, and assess the importance as a hypothesis requiring validation.

Measure Early

• Focus on the core use cases, nail them first.
• Focus on first time experience initially, but measure engagement early.
• A/B test right from the start.
• Choose a platform carefully
  • iOS has more early adopters, and a paying community.
  • Android has no review process, smoother alpha/beta distribution mechanics.
• Set up end to end cohort analysis
  • Instead of Google Analytics, consider Localytics or MixPanel.

Get Out Of The Building!

• Meet users, observe them using your app. Ask open ended questions (Why? How?)
• Blind test users using your app. Get their permission to screen record their interactions. Review it with them, and probe to understand their thought process.
• If you have competitors, have the users blind test your app and theirs, and ask for feedback.

Challenges in Applying Lean Principles to Mobile

• Limited flexibility with Alpha/Beta channel distribution
• Big Bang App Launches = Higher Rankings = Better Distribution
• Ability to push out updates quickly (iOS)
• Lack of mature technical frameworks to support CI, cohort analysis, A/B testing.

Anti-Patterns

• Buy cheap, targeted users that help validate your hypothesis. F&F are cheaper, but not useful.

Anti-Patterns

• Consider launching your experiments under different code names on the app store.
• Once the experiment is over, direct those users to your new MVP with a new code name.
• Users of the new MVP will not see the old app and its ratings.
• This greatly derisks your 'big bang' app launch.
• Another possible idea - launch your app in a representative country (Australia, for instance) before publishing it in the U.S.

Key Takeaways

• Ask the right questions.
• Talk to your users.
• Establish a rapid Build - Measure - Learn loop.
• Be honest with yourself.
• Avoid wastage.

Ask Me Anything

@twitortat

arvi@alumni.iastate.edu

www.arg0s.in

Bonus - How would you build an Instagram app MVP?

Bonus - How would you build a DrawSomething app MVP?