Exponential Growth, with Marshmallows

In the early 1970s a trio of psychology professors at Stanford University began to explore the willingness of people to postpone their immediate gratification for the sake of longer-term gains.

In a series of tests published in 1972, which became known as “The Stanford Marshmallow Experiment” the group examined how long children could resist the urge to eat a marshmallow that was placed in front of them if they were promised more marshmallows later.

“If you wait a short while, you will get another marshmallow later”

The initial tests were focussed on whether distracting the children from the marshmallows was an effective way of increasing their wait times. In a result that will shock absolutely no one who has spent time with small children, distracting them worked rather well.

More interesting than the initial study, was the litany of follow-up studies done between the 1980s and the 2010s which linked higher delay behaviours to better outcomes later in life, such as higher SAT scores and perceived competence.

How does this help me grow my business?

In the business world, there are two types of growth that you are likely to encounter: Linear and Exponential.

The linear approach to marketing

Most businesses are focussed on a linear model as there is a clear, direct relationship between marketing work going out and sales coming in.

For instance: 

A typical day for a marketer working in a linear environment would be to prepare and send a batch email blast to their customer base with an enticing offer: 25% off Marshmallows. 

The email gets sent, receives the industry-standard open rate of 27%, a click-through rate of 1-2%, and the business sees a corresponding spike in the number of marshmallows sold. Let’s say they sell 100 bags of marshmallows instead of 20. The following day, the marketer starts over with a new offer for a new product…

The below graph is a typical example of what the linear model might look like for the sales numbers of marshmallows.

When the business wants to see a larger increase in the total number of products sold, they hire additional marketers and begin outputting two campaigns a day (resulting in 160 additional bags of marshmallows being sold).

Over time, this leads to incremental revenue growth, and the business will rely on economies of scale to grow their profit margins.

Moving to exponential approach to marketing

By contrast let’s consider an exponential model of growth.

You don’t need a degree in data science to notice two very important things about the exponential curve on the graph above:

1. The right-hand side of the curve is steep and extremely favourable.
2. The left-hand side of the curve is flat and not very favourable.

For most small businesses, cashflow is an ever-present concern. Quite naturally their owners adopt the linear approach to get some revenue flowing in, and to keep the lights on. The problem that owners often encounter is that once the revenue does start to flow, they find themselves committed to this model and the type of thinking that goes with it. As such, the businesses either don’t revisit the issue, or can’t ascertain how to move from one model to the other (they want to have their marshmallows and eat them too).

The key to migrating from a linear model to an exponential one is to shift the business’s focus from output to capability. In maths parlance this would mean looking at the rate of change (the slope) of the curve instead of the numbers.

The tried and tested pathway to increasing capability is by leveraging automation. From water mills, to steam power, to the computing revolution, it has worked staggeringly well for businesses over the last few hundred years. In marketing, because we’re a creative bunch, we refer to our branch as “marketing automation”.

Marketing automation is about setting up rules so that a piece of software (or several pieces of software) can handle situations in the marketer’s absence. This frees up the marketer to build the next automated set of processes. These processes can then be repeatedly stacked up to increase the baseline number of sales.

In contrast to our linear approach above, a typical day for someone in a marketing automation role would be to log into a piece of software and define a set of rules for identifying a target group of prospects. For instance, it might look for anyone who viewed more than two different flavours of marshmallows on your website, or anyone who read a recipe on your blog which included marshmallows in the ingredients list.

The marketer would then instruct the software package to let customers know that there was a 25% sale on marshmallows. If the prospect was an existing customer, the software could send them an email if they leave the site without completing a purchase. If the prospect was visiting your site for the first time, the software could deploy a popup with a coupon code, if they looked like they’re about to leave the site. 

This wouldn’t immediately result in a spike in the number of marshmallow bags sold (like the email blast), but it should lead to a small consistent improvement (perhaps 25 bags a day instead of the usual 20).

The following day, while the marketer is doing the same thing for “candy bars” the marshmallows are still selling at 25 bags a day, and they will continue to do so every day hence. By the end of the month, the automated approach will have outsold the blast approach from the start of our story. Meanwhile, the marketer has been free to create automated campaigns for the remainder of the products in the inventory.

As you can probably tell from the above graph, this approach on its own doesn’t get us exponential growth. To get exponential growth we have to continually look for efficiencies and optimisations in the interconnected processes that we have set up. Which we can now do, as we have freed ourselves up from sending daily blast emails, and established a solid baseline of marketing activities.

Optimisations are the secret to building organisational capability

Optimisations allow us to “cheat” the system in a couple of ways:

1. Three 25% improvements are usually easier to implement than a single 75% improvement. So the incremental effort required does not scale with the benefits attained.
2. By looking for improvements in connected areas of a business, we can leverage a compounding effect. As the below picture shows, two 25% improvements add up to 56%, while three adds up to 95%. This gives us a bonus result of 6% and 20% respectively.

(Horrah for maths!)

What is the key takeaway for my business?

If your cashflow situation is tight, send the email blast and survive to fight another day. But if things aren’t that dire, and you can afford to take a bit of time to automate and optimise, then prioritise those activities instead.

If you can wait a short while, you will get a lot more marshmallows in the end.