Math Secrets Behind Subscription-Based Distribution Models

October 31, 2017 BY Charles Fallon

subscription based distribution

One important evolution in e-commerce has been the proliferation of subscription-based business models. From meal kits to dive-bar t-shirts, companies connect directly to consumers and curate a weekly or monthly selection of products based on the individual profiles of those consumers. If these companies live up to their brand promise, then they hit a gold mine of recurring revenue from loyal consumers.

The subscription concept is not new, as anyone who fell into the trap of signing up for a Columbia Records membership in ancient, pre-internet times knows. However, technology enables consumers to tailor and adjust their deliveries in ways that they could not with Columbia Records. (Does anyone remember that deflating feeling of getting Michael Bolton’s Soul Provider in the mail?)

Behind this limited customization is a complex distribution requirement that most subscription-based businesses do not understand nor provide sufficient infrastructure to achieve productively. Here are a few basic things that any subscription-based distributor needs to understand before figuring out how to build an operation to support the business.

Customization is a Combinations Game:

The default preferences of your subscribers already create a certain variation in the kinds of packages you will send to them. Let’s pretend that you send out snack boxes to 100,000 subscribers.

  • If every subscriber gets the same package – let’s say 5 snacks per week – then life is easy. Set up an efficient assembly line to pack 100,000 identical boxes.
  • Let’s say that half the subscribers want 3 snacks per week and half want 5 snacks. Beyond that, you permit no variation. Again, like is still pretty easy with 2 runs on one line or 2 lines.
  • Let’s say that each week, you have 6 snacks available and while you always set up a default selection for each subscriber, you also allow them to customize which snacks you want. 

Now, life gets hard because:

o For the 5-snack per week subscribers, there could be as many as 6 combinations that make up a shipping box¹ 

o For the 3-snack per week subscribers, there could be as many as 20 combinations that make up a shipping box²

Suddenly, small customizations can take a simple, straight-forward production run and turn it into something much more complex. We have one meal kit company that begins with 11 meals on its weekly menu and generates, on average, about 150 different meal combinations to pack and ship out.

Pareto’s Law Still Rules:

While the subscription business is a combinations game, Pareto’s Law still rules. In the above example where the 3-snack per week subscribers could create up to 20 combinations of shipping box, 20% of the combinations will generate 80% of the shipping boxes.

Part of what drives this Pareto effect is our tendency towards laziness: many subscribers will accept the default selections. The other part is that no matter how cherished every product will be to your merchandising team, there will always be dogs and losers that leave your customers cold.

How this affects the operation can vary based on the actual numbers. What matters is that there will be box combinations that are hugely popular and processed in large runs while there will be box combinations that are extremely unpopular and would fail if processed like a the large run boxes. Moreover, within the combinations – both popular and unpopular – there will be fast-moving and slow-moving individual selections.

Pareto’s Law is to distribution what the mafia was to Michael Corleone: no matter what you do, you can never get away from it. However, in subscription-based distribution, the combinations game that comes from customization adds a scale of complexity that can create all sorts of unnecessary costs and service level penalties if not understood when designing and implementing your operation.

¹ Mathematically, selecting 5 from 6 is calculated as 6! ÷ (5!(6-5)!) = 6
² Similarly, selecting 3 from 6 is calculated as 6! ÷ (3!(6-3)!) = 20