When not to use this mental tactic

There are some situations in which making decisions on the margin is not sensible. Consider the fashion retailer Nordstrom, renowned for its laissez-faire return policy. Nordstrom’s customer satisfaction guarantee includes full reimbursement of returned products even without receipts, often years after the purchase. At the margin, Nordstrom shouldn’t accept these refunds: the marginal costs of paying refunds for used clothes that might be unsaleable results in a marginal net loss. But taking into account the reputational benefit of the liberal return policy, Nordstrom’s approach might lead to an increase in overall profits and greater customer satisfaction.

Marginal units

Instead of considering the total or average amount in every decision, the marginal thinker is only concerned with the additional unit, such as the next hour of driving, the next piece of pizza, or the next investment. One of the fascinating characteristics of marginal units is that they are what economists describe as typically decreasing in whatever input factor is required.

What does that mean? For example, take eating pizza. The first slice will be delicious, and so will the second, third and fourth. Soon you’ll be getting full, and the next (marginal) slice of pizza won’t be as satisfying as the first one. In the chart below, the vertical distance, measuring satisfaction (or, as an economist would term it ‘utility’) declines with each extra piece of pizza – eventually even turning negative.

Or consider an energy company active in oil exploration. The first oil fields are easy to tap into as their location is known and their reservoirs are large and accessible (assume, to simplify, that all oil fields have oil reservoirs of equal sizes). Maximising their profits, the next (marginal) oil field project will be the ones that are slightly harder to access, and so on. At the end, the only fields that remain are those that are hardest to reach.

In the following chart are the expected exploitation costs and revenues for four oil fields. The lengths of the arrows indicate exploitation costs per site. The executive committee, acting on behalf of shareholders, would start with those projects promising highest profit first (oil field 1), and work their way downwards. The expected revenues of oil field 4 are below the exploitation costs, at least at the given moment. This may change as technology advances and then your marginal utility would change too.

Marginal thinkers ask themselves: “What’s the most value-adding activity I could spend my next five minutes on?”, “How can I put this $100 to best use?” or “Given the current customer base, which customer should we take care of next?”

Sunk costs

Sunk costs are costs that have been incurred as a result of past decisions. They are unrecoverable. Because they happened in the past, they are different from the other costs a company faces, such as R&D or materials. They are independent from any event in the future. Sunk costs have been incurred no matter what you choose to do next. They cannot be influenced.

Even though we talk about costs here, what is meant is any kind of outlay (time or attention) expended to bring about a better result in the future. A cost could be learning a skill, for example. Say you are told that an attractive business position opened up in Latin America. The deal is sealed, and in a year from now, you and your family plan to move to Mexico City. In anticipation of your new role, you invest many hours per week in learning Spanish. However, due to bad luck, the position in Mexico City falls flat. At this junction, it is natural for most of us to say to ourselves, “Well, I’ve already invested at least 200 hours in learning Spanish. I might as well make the most out of it and continue to practise it. You never know!”

But it is actually not the most rational thing to do, unless you really enjoy learning languages or anticipate that knowing Spanish will be advantageous to your future career or life. The fact that you have learned Spanish in the past should not in itself be enough reason to continue to do so. Assuming you don’t benefit in speaking Spanish in any other way, the costs of learning Spanish are sunk and unrecoverable. You simply can’t undo the past, or allocate the cost (your time spent practising) in any other way.

Humans tend to hold on too tightly to their past decisions and hence fall into the sunk cost fallacy. We do this because our mind subconsciously connects decisions that are related, even if they are in the past and don’t make a difference for future decisions. Marginal thinkers are only interested in the future outcomes of their decisions and future trade-offs. They don’t have any allegiance to whatever happened in the past.

How can you address this? Armed with The Decision Maker’s Playbook, you can take on a new role in your organisation: the sunk cost bias police. Whenever you hear someone saying something like, “We’ve already spent so much time on this” or “We spent a lot of money setting up the business in Angola,” you can be virtually certain that sunk cost bias is at work. We’ve found it helpful to respectfully but directly intervene at those moments by saying: “That’s actually not the point. That time and money is gone. The question is whether the next dollar or hour we are about to invest is worth it, knowing what you know right now.”

Decreasing marginal returns

This idea is key to marginal thinking. In fact, without decreasing marginal returns there would not be the need to think at the margin.

Decreasing marginal returns are ubiquitous in life. For example, fertiliser improves crop production, but only up to a point. The more fertiliser you add, the lower the impact produced per unit of fertiliser and overkill can even reduce the yield.

The same is also true of the satisfaction we get from activities. Riding a rollercoaster is fun for a few minutes, maybe even an hour. After that time, it quickly gets less exciting. Even the satisfaction of being on holiday, something most us look forward to more than anything else, doesn’t linearly increase with the duration of the holiday. It is fun to forget which week day it is during the first two weeks, sit at the pool, drink margaritas and live and let live, but after a while even this gets dull and one wishes to go back home (and potentially even back to work).

The same is true for the satisfaction we derive from income. We often imagine that more money is better, and that we cannot have too much of it. One of the best-known studies on the relationship between happiness and income levels was done by Betsey Stevenson and Justin Wolfers in 2013. On a life satisfaction scale from 0 to 10, doubling your income doesn’t translate into doubling your happiness. Rather, it buys you half a point on the scale. Stevenson and Wolfers illustrate this using an example: “GDP per capita in Burundi is about one-sixtieth that in the United States; hence a $100 rise in average income would have a twenty-fold larger impact on measured well-being in Burundi than in the United States.”⁵³

Marginal thinkers are very aware of the shifting trade-offs as time goes by (as they move up the line), and ready to make decisions on the margin – only taking into account their current position.

Opportunity costs

The concept of opportunity costs is closely related to marginal thinking. Let’s say you have one hour available this afternoon to be productive. You could either tidy up your inbox and reply to emails or improve your Spanish. Note that as a function of time invested, both of these activities have decreasing marginal returns just like the yield/fertiliser mentioned. Typing away for a whole hour will make you tired. Holding constant the quality of your replies, the twentieth email will take longer than the first. And for your vocabulary skills, the bottleneck will be your brain’s capacity to process and successfully memorise new terms.

So, for both tasks, email replying and vocabulary memorising, the effectiveness will decline over time. This is exactly why the curve on the previous page is concave. Let’s look at an example. Say you are at point B, which allows you to use the next hour to send 6 emails and learn 45 new words in this hour. And let’s say your holiday to Costa Rica is coming up and you’d like to maximise the number of new words you learn. How many email replies do you have to give up in order to learn 50 words per hour? The cost of learning five more words is six emails. The easiest way to remind yourself about opportunity costs is to ask, “What am I giving up to get this thing? Is that a rational trade?”

Note that the opportunity costs differ based on your starting position. On the margin, if you are at point C, your opportunity costs for one additional word are just 6/10 emails, – half the price than if you were on point A. Marginal thinkers keep opportunity costs in mind (the benefits you would have received had you taken another course of action) to make best use of their time.

Marginal thinking