24
Launching Mixer: What worked, and what didn’t Wednesday, 9 January 13

Launching Mixer: What Worked, What didn't

Embed Size (px)

Citation preview

Launching Mixer: What worked, and what didn’t

Wednesday, 9 January 13

We wanted to build a service that would help us better connect with the people around us.

Introduction

Wednesday, 9 January 13

Wednesday, 9 January 13

Theoretical Foundations

• Recognizing the important concept of shared identity in communities

• Providing opportunities for individuals to interact with one another around a common theme: location

• Understanding communities require a dynamic where each person is able to communicate with all others in the group (George Homans; 1950)

Wednesday, 9 January 13

Launching

Wednesday, 9 January 13

Approaching the press• Focus on an influencer

• Meet them whilst out networking and have a conversation

• Pre-brief them by giving a heads up

• Negotiate

• We’ll launch when you’re ready

• Get to the point

• Have a story

• Send images in web ready format

• Details about competitors

Wednesday, 9 January 13

Marketing is really important

• People often wrongly assume apps will market themselves.

• Other apps we’ve built have done well with no marketing, we assumed Mixer would be the same.

• TechCrunch got us 500 installs on the first day but not core users who didn’t stick around.

• Marketing strategy is as important as the app itself!

Wednesday, 9 January 13

Post-launch

Wednesday, 9 January 13

Monitor everything

• Growth and user acquisition most important.

• Identify why some users stick around and why others don’t.

• App Store analytics and rankings.

• Listen to feedback.

• Be prepared to iterate.

Wednesday, 9 January 13

App Store challenges

• Know what matters (App name, Developer Name, Keywords, Title of in-app purchases)

• Get in top 100 = organic downloads

• Easier to rank in some categories over others.

• Pick the right keywords.

• Localize app store listing.

Wednesday, 9 January 13

What did we learn?

• Focus on the things your app does differently to others

• Narrative is important

• Needed a new way to distribute

• Profiles were weak

• Need for niche focus

• Needed to encourage more engagement

Wednesday, 9 January 13

You can always build on an MVP

• Focus on one key feature

• Well built

• Well designed

• Don’t build ‘nice to have’ features until users demand them

• Must be compelling

Wednesday, 9 January 13

Technical

Wednesday, 9 January 13

What did Mixer need?

?

Wednesday, 9 January 13

What did Mixer need?

?

Wednesday, 9 January 13

What did Mixer need?

?

Wednesday, 9 January 13

[PFFacebookUtils logInWithPermissions:permissions block:^(PFUser *user, NSError *error) {    if (!user) {        NSLog(@"Uh oh. The user cancelled the Facebook login.");    } else if (user.isNew) {        NSLog(@"User just signed up through Facebook!");    } else {        NSLog(@"User logged in through Facebook!");    }}];

Wednesday, 9 January 13

Use the rights tools

• We used Parse (Parse.com) as our backend to increase simplicity and decrease development time. (it’s practically free!)

Wednesday, 9 January 13

Use the rights tools• Problem: we needed to identify mutual

Facebook friends and interests (Likes).

https://graph.facebook.com/me/mutualfriends/?user=USER_ID

SELECT page_id, name, pic_big FROM page WHERE page_id IN (SELECT page_id FROM page_fan WHERE uid = USER_ID)

Mutual Friends (Graph):

Mutual Interests (FQL):

Wednesday, 9 January 13

RTFM (or “read the docs”)

• Problem: we needed users to spread the word about Mixer; share out your Mixer posts to Facebook and Twitter.

• We already had a easy way for Facebook access (and fetching Twitter credentials is easy enough since iOS 5).

• For Facebook, Open Graph was perfect, but read carefully to really see what Facebook’s API can offer.

Wednesday, 9 January 13

RTFM (or “read the docs”)

Wednesday, 9 January 13

Keep up to date

• Mixer development started before iOS 6 and before Facebook iOS SDK 3.

• Over 40 Parse SDK updates since 1st July 2012.

• Keeping up to date is key - permissions and sessions changed dramatically in v3 (and then again in v3.1) which we could have planned for.

Wednesday, 9 January 13

Consider your edge cases early on

• We found large milestones easy to accomplish.

• But we did not account for our edge cases.

• What if users invalidate sessions?

• What if users change their profile picture?

• Consider solving these problems early on.

Wednesday, 9 January 13

Mixergetmixer.com

@connellchris chrisconnell.me

@rossbeale

Wednesday, 9 January 13