Top Banner
35

Measure or Die

Sep 06, 2014

Download

Technology

Neil McCarthy

How Yammer uses data to make product decisions
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Measure or Die
Page 2: Measure or Die

Measure or Die

Neil McCarthyKen Pascual

Page 3: Measure or Die

Justifying featuresExecution decisionsDetermine success

Page 4: Measure or Die

HOW DATA HELPS US PRIORITIZEStep 1: Justifying Ideas

Page 5: Measure or Die
Page 6: Measure or Die

Previous signup flow experiment

Page 7: Measure or Die
Page 8: Measure or Die
Page 9: Measure or Die

Clues from the past

Page 10: Measure or Die

Brainstorming

Page 11: Measure or Die

Idea!

Page 12: Measure or Die

Does this align with our goals?

• Vision: Yammer is a primary communication tool that users use to get work done.

• Goal: Grow the number of users that use the product.

• Problem: We lose too many potential users in our signup flow.

Page 13: Measure or Die

How important is this?

Page 14: Measure or Die
Page 15: Measure or Die

big drop!

Page 16: Measure or Die

And HALF of our invites come from the signup flow!

Page 17: Measure or Die

Conclusion: This seems important.

Page 18: Measure or Die

Why justify?

Page 19: Measure or Die

HOW DATA HELPS US DETERMINE THE SUCCESS OF THE PROJECT

Step 2: Determining Success

Page 20: Measure or Die

What’s your hypothesis?

Page 21: Measure or Die

What we came up with…

If we move the email confirmation step to the end of the signup flow,then more users will make it through the signup flow and more users will be retained.

Page 22: Measure or Die

Secondary objective: Kill the Desktop app

Page 23: Measure or Die

We need another hypothesis

Page 24: Measure or Die

“If we remove the prompt to install the desktop app, then we will lose an acceptable amount of engagement”.

Page 25: Measure or Die

A/B Test Results: Killing the Desktop app prompt

Page 26: Measure or Die

A/B Test Results: Killing the Desktop app prompt

Page 27: Measure or Die

Interpretations

• Hypothesis disproved (in a good way).• Users find a way to post.

Page 28: Measure or Die

Results!

Page 29: Measure or Die

A/B Test Results: Moving the email activation step

Page 30: Measure or Die

A/B Test Results: Moving the email activation step

Page 31: Measure or Die

Interpretations

• Users saw the steps in a different order, and dropped off in similar percentages, BUT– Less users saw the prompt to “check email”

because it was at the end.– Thus, less users activated via email.– Thus, less users had the opportunity to engage.

Page 32: Measure or Die

Why we chose this example

Page 33: Measure or Die

How do you do this?

• Gotta log data!• Choose your own metrics• Data driven vs. data informed• Build your processes around data• How much PMs should know about statistics.

Page 34: Measure or Die

Q & A

Page 35: Measure or Die

Thanks for coming!

• Neil McCarthy, @the_neil• Ken Pascual, @vorped• Adam Zadikoff (original PM for this project),

@zadikafka

• Test results from the first signup flow test: https://medium.com/yammer-product/13eeb6435a73

• These slides on Slideshare coming soon!