how buffer uses looker to create user targeted emails

14
How uses to create user targeted emails

Upload: sunil-sadasivan

Post on 08-Aug-2015

63 views

Category:

Data & Analytics


3 download

TRANSCRIPT

How uses to create user targeted emails

Numbers• Founded Late 2010

• 48k DAU / 190K MAU (~3k daily signups)

• $470k Monthly Revenue Run-Rate from ~33k customers

• ~600k social media posts scheduled per day

• 31 Team members

• 9 Engineering, 3 Marketing, 7 Support, 2 data

Sign-up/Onboarding lifecycle emails (2012-2015)

Welcome Install Extensions Buffer Tips

Why didn’t we iterate on emails?

• Hard coded templates in our code base

• Required developer time to think and implement email triggers

• Small developer team, emails not a dev priority over product

• Targeting users part of our larger data problem

Led to not-so-great marketing habits

• Mass emailing our 2 million users

• Lower open and click rates

• Higher spam rates and bounces

• Rough experience for users with irrelevant messages

October 2014

Dec 2014Awesome! Moving fast with data

analysis… But iteration can’t keep up.

Solution: Remove developers out of the equation

• Empower any team member to create/style their own emails

• Give them the power to target individual users (without needing dev help)

• ie. Build our own version of customer.io, leveraging Looker to target users

Example: New Targeted Welcome emails

Example: Welcome email to users who joined 3 hours ago and who have set up their schedule

Buffer’s Trigger Email CMS

How it works

• Simple cron job running every 10 minutes

• Runs Looker generated SQL queries in redshift to grab user ids

• Sends users notifications/emails

• Some regex magic to look up currently used PDT

How the team is using it

• Started with marketing and lifecycle emails

• Outbound support. Predicting areas for confusion and errors

• Customer Development, targeted surveys

Where we want to go

• In app notifications (mobile and web)

• Dynamic data through templating (eg. %column2%)

• Request to Looker (adjusting column order reflected in SQL statement)