Tame The Retention Beast With Dunzo’s Insights Into Retention Metrics

Efforts go down the drain when loopholes get bigger. It is similar to putting water in a bucket full of holes. The water will flow out much before the bucket is filled up to the brim. Somewhat similar is the story of those companies that do not focus on (customer) retention but tighten all other screws. 

Brijesh Bharadwaj, director of product management, user products and growth at the hyperlocal delivery service provider, Dunzo, helped us understand the importance of targeting retention and identify the right retention metrics when he shared his knowledge at The Makers Summit 2021, the virtual product conference hosted by Inc42. So, let us fix the leaky bucket before it is too late with some key insights from his masterclass. 

Why Retention Is Of The Essence

You can bring in new users to the platform, but what will happen if they do not stay? It is crucial to make the product sticky for users as stickiness impacts the growth of the product and the company. Speaking about a basic cycle, Bharadwaj said if more users were clinging to an app, they would soon spread the feedback among friends and families, thus growing user acquisition as never before due to word-of-mouth publicity.

For example, with more users retained, there will be more files shared at Dropbox and more links will be exchanged. Similarly, with more retained users at Dunzo, there will be a higher number of referral codes shared, thus enticing many new users to become a part of the Dunzo family. Therefore, different techniques are used by different businesses to raise the number of retained users and other top-funnel numbers. 

STREAM THE SESSION HERE

Retention Is The Most Important Growth Metric

Although faster user acquisition is a good metric, higher retention leads to speedier recovery of marketing spend and better monetisation. Over his/her lifetime, a retained user may help a business earn, say, $50, which is then multiplied by the number of retained users. This specifies the amount that a business has made by focusing on the retention power of the product, said Bharadwaj. This way, you have a larger fund set aside for paid acquisitions that can be used advantageously to outprice competitors in the ad market.

Awareness Is Not Enough; Define Your Retention Metric

To define the retention metric of your business, you need to understand the frequency of problems or the needs that are solved by the product and the value event. Giving the example of Dunzo, Bharadwaj said that the problem could be customers running out of groceries and its frequency could be once or twice a week. In the case of Kindle, the need could be to read a book and this could occur once or twice a month. The value event, in each case, is getting the groceries and reading the book, respectively. 

But then, how will you find your ‘value event’? This crucial question can be answered by examining your business’ data. Check the retention of users at different events throughout the process to spot your ‘value event’. For Kindle, there can be three important events: Buying a book, finishing a book, and the reader giving the book a four-star rating out of five after completing it. If you are confused about your own ‘value event’, find the largest cohort of users at each event to build your retention metric. 

Analyse How Much Stickiness Is Enough

Bharadwaj talked about analysing how much retention is good, with the help of graphs. According to him, a flattening retention curve is a positive indication that users are staying for a longer period and not moving out. On the other hand, a downwards sloping curve indicates that troublesome times are already here. It is imperative to examine all data regarding the number of sign-ups and the number of users retained by the product to understand what is working and what is not. You can make an excel sheet, collate weekly data and add not just the figures but the weekly percentage of growth or fall.

There will be weeks when sign-ups are high, but retention is negative. Or there will be times when a newly launched product chain may not work as per expectations. So, performing a proper analysis is a smart way to reapply the strategy that saw the numbers grow in the past, while you push the strategies away that did not work out

Ways To Improve Retention Strategy

When you have completed defining the retention metric of your business and also analysed what is working wonders for the company, it is time to go a step further by improving the retention strategy. Bharadwaj thinks bucketing the users based on demographics like age, gender, income group, whether they have kids or not and so on is crucial to dive deeper into users’ requirements. A growth leader must choose which of these factors can meaningfully impact a user’s experience. 

For instance, the use case of Ola/Uber can be influenced by a user’s profession. Discounts on airport rides will not positively impact a 20-year-old college student but will benefit a high-income business executive. 

Whether a user values convenience or money can also shape the value proposition of a product. Bharadwaj said that users with pets or children are more likely to order food from Dunzo than a person without them, thus impacting the frequency of product usage. For Ola and Uber, it is important to know how many alternatives are there in place — do users own cars or have regular access to carpools? Thus, one needs to bucket user attributes favourable to a product’s growth. 

STREAM THE SESSION HERE

Apart from user demographics, acquisition channels need to be prioritised as well. For a few companies, organic installs that occur without any advertising may bear fruit, but Facebook ads may not. For others, it can be a referral programme that brings in inflated retention numbers instead of Google ads. Every business needs to peep into its channels and trash the options that burden it financially rather than elevating it. 

By simply looking at product categories and extracting the data around whether pizza lovers or healthy eaters have higher retention for a food delivery app one can actually improve retention. This way, users can be shown ads based on their preferences and interactions can be increased, thus enhancing user experience and raising retention. 

Moreover, it can be a particular OS that is yielding better results and delivering a finer experience or a specific product feature that is being used frequently. Once you have understood what is causing better retention, you can run experiments and lean more towards those user attributes. 

Run Fast With Your Experiments

However, running long-term A/B tests on retention is difficult as one cannot wait for 12 months to view the differences in retention levels. The trick is to run experiments on events that occur early on in the product life cycle and also correlate with retention, said Bharadwaj.

If one can identify an event X that when completed within a certain period Y, it becomes highly likely that the desired event which is the value event will be accomplished. The X and Y can be spotted via varied hypotheses. For instance, a food delivery app can create different hypotheses like ‘two orders within 30 days’ or ‘four orders within 45 days’ and increase or decrease the time and order intensity. For each hypothesis, user retention can be calculated, and small jumps will be visible depending on the hypothesis.