IDENTIFY THE BEST CONVERTING SOURCE AND ASK YOURSELF “IS THIS DIG IN FURTHER INTO INDIVIDUAL CAMPAIGNS AND ASK YOURSELF “WHICH ONES DO I KEEP OR DUMP?” Cohort Analysis: What Is It? Why Does It Matter? FIGURE OUT HOW TO GET REPEAT BEHAVIOR AND FORM HABIT LOOPS Cohort Analysis Who sticks around from one time period to another? Analyzing cohorts increases your chances of having people upgrade over time or buy again. If you don’t have good retention, nothing else matters. Brian Balfour, VP Growth, Hubspot • What % of the user base is still active? • What differentiates groups of people? • What actions can you take to make people stay? Example Cohort Report This row shows how many people signed up in February 2015. This row shows how many people signed up in July 2015. People signed up in February 2015. People signed up in July 2015. These columns show how many months have elapsed since the sign up month. These cells represent the % of people that have come back within the first month since signing up. Darker cells represent hot spots of high retention rate. Lighter cells represent low retention rates. The most important thing is the curve of this line. This is a bad retention curve because this line reaches 0. Retention starts at 40% at the first month, which means you’ve already lost 60% of your original users. By month two, you’ve dropped to 15% of your original users. By month 3 and onwards, barely any original users are present. Over the course of 3 months, you’ve essentially lost all of your users. This is a good retention curve because this line NEVER reaches 0. Retention starts at 100% in the first month, meaning everyone has stuck around. From month 2 onwards, you drop to 40%, but maintain there. With improvements to your marketing, product, and efforts, you hope to increase retention every month. 1. Click on the cell 2. Click on the “View the 102 people” Use the words and phrases from customer responses in the marketing copy to increase conversions Message Experimentation Thomas H. Davenport, Professor, Babson College Experimentation Figure out how one channel works for you. Experiment on other channels to figure out how to get them to work. • What do you do now that could be improved? • What inputs do you control? • Do you have a culture of using data to make decisions? The real payoff will happen when the organization as a whole shifts to a test-and-learn mind-set This is where we lifted product adoption by 12%. Notification • 17.2% conversion rate • over 2 months • 538 conversions Lightbox • 30.84% conversion rate • over 3 days • 278 conversions 1 Perception is everything. Tap into values, feelings and storytelling within your marketing. 2 Segment your audience to identify the best performing customer groups - then optimize. 4 Experimentation grants huge lifts if you have the culture and process.