Use qualitative & quantitative data to analyse the stickiness of your product.
Companies spend pots of money on several marketing channels, reaching out to customers to sign up to their service. But does it make sense to throw 10$,15$ or 20$ ads at a new customer, when they only stick around for a few days or months? The answer is probably no. So before spending money on acquisition, it’s essential to fix your churn (“leaking bucket”) and check your retention rate.
The retention rate allows you to monitor your performance on attracting and retaining customers. Counterpart of the retention rate is churn. It is also very hard to give an absolute benchmark on a good retention rate performance, as it strongly depends on your specific business. In the following you can find a simple guide on how to explore the stickiness of your product and if users are coming back to use your service after signing up.
1. Qualitative Data
The first step is not the subscription to an expensive tool and pump the shit out of probably most unnecessary data. No! Getting first insights is much simpler: Talk to your customers!
It is one of the most powerful tools every entrepreneur can choose from. This may be surprising and basic, but knowing your customer’s pain points, having the gut feeling on how important the issue is and tackling them together with the product team is key to achieve a high retention rate. Qualitative data generated from interviews or support requests, give you a first guess on where to measure and what to look for. Of course a gut feeling does not give back any retention rate yet. Therefore step to — use quantitative data and combine both.
2. Quantitative Data (Cohort Analysis)
After talking to your customers and creating first hypothesis on where problems are and how to improve them, every company calling itself “data-driven”, should definitely use hard data complementary to the qualitative approach.
One of the highly recommended and quick ways to gain deeper insights on your retention rate is a cohort analysis which is explained in the following.
Cohort Analysis on retention
Generally speaking a cohort analysis focuses on the behaviour of a predefined group (=cohort) over a certain period of time. Although the analysis can be conducted in almost every sector with a wide range of various characteristics, in this case we look at business analytics and especially retention. To set the frame of the cohort analysis you mainly need four inputs as described below:
Cohort Objective: First it is to clarify which question you want to answer with your cohort analysis on retention. Be creative when it comes to splitting it down. A few examples for different questions on retention:
- How does retention differ between sign ups by social logins compared to e-Mail login?
- Do I see an increased retention rate after releasing certain new product features in a particular month?
- Are mobile users more likely to login again than desktop users?
There are endless combinations of variables leading to deeper insights and help understanding your product. The example of signups and log in again after months is a very simple one and serves well for illustration. But once understood the concept there are much more insightful options.
Cohort Metric: This is the metric that is being measured for each cohort. It could be user retention, revenue, session duration, sales or any other metric you want to deep-dive into.
Cohort Size: How large should the cohorts be? Do you want to divide your cohorts into days, weeks or months? You can also further classify the cohorts not only in a timely manner, but for example on device (mobile/desktop), age or referral. Again, it mainly depends on your business.
Time frame: This is simply the time period you want to conduct your analysis on. This can either be several days, weeks or months. In our example we have a timeframe of 6 months.
How do I read the graph above?
The graph above is probably one of the most common and also most insightful ones. You can find it on tools like Mixpanel, Google Analytics, Amplitude and the like. The visualization of the cohort analysis above allows you then to compare the metrics and user behaviour of different cohorts over a certian period of time.
You can see the month of user sign up on the y-Axis and a timely measurement (here months) on the x-Axis. The percentage inbetween tells you how many people logged in again several months after signing up. Further subdividing allows you to find the cohorts with best / worst performance, and what characteristics these cohorts have in common.
Now you determine which different user characteristics / activation campaigns / changes in product or whatever adaptions you have made over time had an impact on your retention rate. The result is an improved understanding of your user’s behaviour and actionable analytics to improve it positively.
Conclusion:
Hard data comes in the form of numbers, text or graphs and is used increasingly to make better decisions and gain insights. Nevertheless in many cases the opposite data source, qualitative data, is overlooked.
Don’t forget that regular user feedback and gut feeling can be the kickoff to start an analysis (retention rate in this case) more detailed. It gives you the hints you need for a targeted and successful analysis with hard data. Only a combination of both can deliver the results you need in order to really understand why users stick to your product (high retention rate) or not.
