Super Bass-O-Matic Subscriptions: Lessons from Rovco on Customer Retention

acksup

As you know, Rovco has traditionally sold delicious bass on a transactional basis to its “Super Bass-o-Matic” customers.

Spokesman: How many times has this happened to you? You have a bass, and you’re trying to find an exciting new way to prepare it for the dinner table. You could scale the bass, remove the bass’ tail, head and bones, and serve the fish as you would any other fish dinner. But why bother, now that you can use Rovco’s amazing new kitchen tool, the Super Bass-o-Matic ’76. Yes, fish-eaters, the days of troublesome scaling, cutting and gutting are over, because Super Bass-o-Matic is the tool that lets you use the bass with no fish waste, and without scaling, cutting or gutting.  Here’s how it works: Catch a bass, remove the hook, and drop the bass – that’s the whole bass – into the Super Bass-o-Matic. [drops the bass into the blender ] Now, adjust the control dial so that the bass is blended just the way you like it. [turns blender on and grinds it to a pulp] Yes, it’s that simple!

[Cut to Bass-Drinker who drinks a glass of bass] Bass-Drinker: Wow, that’s terrific bass!

Rovco’s non-contractual relationship with its customers made the timing and number of bass purchases hard to model. As most of you know, Rovco has created a service which enables its customers to have fresh bass delivered to them twice a week on a subscription basis. As you can imagine, these bass subscriptions are very popular, but there are still the usual customer acquisition cost (CAC), cost of goods sold (COGS), average revenue per user (ARPU) and customer churn (retention) issues to deal with. Some customers still prefer to buy their bass when they feel like it on an a la carte basis.

I have written about the opportunities and challenges of a subscription business model here and here. I tried to keep these posts simple to make the ideas more understandable. In this post, I will provide a more nuanced view of customer retention. Among the challenges Rovco and other subscription businesses inevitably encounter is that different subscription customers have different retention rates – some customers will go through just about anything to stay with you while others will run for the exits at the first available opportunity.

Churn is almost always reported as an average but actual churn rates can vary based on factors like sales channel, demographics, license type, license term, and so on. I know what you might be thinking – c’mon, the overall churn rate is a helpful summary measure, so what’s the worst that can happen? In many situations, assuming that customers from all these different channels, demographics, and licenses have exactly the same desire to remain with your company over time can be dangerous – most of the time, it will underestimate the overall future activity and thus the value of your customer base. It can also lead to a very mistaken view of the how well a business is doing and what strategies should be pursued to make it better.

These differences across customers, which marketing professors Daniel McCarthy and Peter Fader refer to as “customer heterogeneity,” should be embraced and leveraged, not ignored. A business with “good heterogeneity” (i.e., a company which has pockets of customers who are much better than the average customer in the portfolio) will be dramatically more valuable than a business with the same overall average churn rate but whose customers all tend to be the same. In other words, it is important to differentiate between businesses that have low retention heterogeneity from businesses that have high retention heterogeneity. Without that, you will not have an accurate understanding of the business – how much your business is worth and what levers can be pulled to manage the value of your customers over time.

Let’s examine this idea through an example. The following spreadsheet screen shots of the Bass-o-Matic business illustrate the point. Imagine that all Bass-o-Matic customers had the same 18% chance of churning out of the subscription each month. This is how their customer lifetime value (CLV) might look:

MOd c

Now imagine instead that while Rovco’s average monthly churn rate was still 18%, it actually had two customer segments – a “die-hard loyal” segment that absolutely loved getting those bass twice a week and would do so for a long time to come, and a mediocre segment that would churn out much more quickly than average. Let’s assume the die-hard loyal segment is small (30% of the customers) with a 1.7% monthly churn rate, while the mediocre segment is large (70% of the customers) with a 25% monthly churn rate. This is how the CLV figures shake out:

MOdD

The average CLV moves up from -$257 to +$83 when we move from assuming one customer segment to two. The way the math works out, for a given average churn rate, assuming all customers are the same results in the most pessimistic forecast for the future value of the cohort. The more “spread out” or “heterogeneous” the customer retention rates are, the more a single retention rate will undershoot CLV, because the good customers are “thrown out with the bath water” by lumping them in with everyone else. This “heterogeneity factor” is an important but often overlooked additional dimension to customer value. This is the essence of what McCarthy and Fader do for public companies in their article on valuing subscription businesses using publicly disclosed customer data. In this work, they show not only that you can “back out” customer heterogeneity from public data, but also that it predicts far better than models which assume there is none. They have applied this same methodology to other public companies like Blue Apron and HelloFresh, and to private companies through a company they started, Zodiac.  The methodology is certainly more involved, but they pretty convincingly show that the alternative, while simple, is dangerous and performs poorly in practice.

Taking it a step further, imagine that you as a manager knew that there were two customer segments that had such dramatically different loyalty profiles. The strategies you could pursue next are clear. You could find out all the features and characteristics that are associated with that better customer segment to try and acquire more like them. You might show them a bit more love than customers in the other segment, because they are worth a lot more to you. And so on. These would certainly not be the most natural strategies to pursue if all your customers were indeed the same.

Embracing differences across customers is the lifeblood of what McCarthy and Peter Fader call  customer centricity. It is impossible to be customer centric if all your customers have the same loyalty or retention rate, and yet this is far and away the most common assumption made when the popular media discusses a company. All the worse that this assumption results in the most pessimistic view of your customers’ value, for a given average churn rate.

As we venture outside of the news media and high-level analyses of the companies in them, the opposite objection often comes up: “we do that segmentation stuff already!” Unfortunately, most people do segmentation in a very suboptimal way. While it is tempting to think that you can segment your customers purely on the basis of observable features like demographics and psychographics, is this the best approach? Demographics and personas have a notoriously weak ability to sort out your good customers from your bad ones. Measures of customer satisfaction like the highly popular Net Promoter Score are significantly challenged. It is much more effective to turn to your best data source on customer loyalty – the actual renewal behavior of your customers, which you get for free from your transaction log. You may not know why Bob loves to eat Bass on subscription so much, but if his renewal data says it’s true, then it’s true, demographics be damned. Incorporating customer history into the segmentation procedure is easy – you can do it in Excel, and there are notes which document in detail how to go about it. The payoff for making this small leap is the ability to truly pursue customer centric business strategies. At a time when the product centric mindset is starting to show some cracks, this payoff looms large relative to its cost.

Summing up, while customer retention is more complex than the popular media would have you believe, it ultimately boils down to a very simple concept – your customers are different from one another. It’s high time that we stop ignoring it and start embracing it.

Please note that these are not the actual spreadsheets I use in my work which are proprietary. They are simplified for teaching purposes. You will need to create your own spreadsheet perhaps using the one provided by David Skok here http://dskok.wpengine.netdna-cdn.com/wp-content/uploads/2017/02/What-is-your-TRUE-LTV-02-2017-1.xlsx  as a starting point.