Customer tips make up a significant source of income for people in most service industries, especially in the United States. For taxi drivers in New York City, passengers would typically add around 10% to the fare when they handed the driver a handful of money after their rides.

But a new dynamic emerged in 2007 when New York City cab companies installed credit-card readers. By January 2012, around 55% of gross taxicab revenue came from credit-card transactions, up from around one-third in 2009 when the readers became standard.¹ Total revenue also increased significantly. One reason for this success is that using a credit card reduces the pain of paying.² When someone knows they can pay with a simple swipe of a credit card rather than search for cash or withdraw some from an ATM, they are more inclined to pay and potentially also to spend more.

Another dynamic, however, had even more powerful implications for New York's taxicab system and, more broadly speaking, for how a company can influence buying decisions. Instead of leaving passengers on their own to calculate a tip, the credit card display offers three specific amounts. These amounts were expressed in absolute dollars for shorter rides and in percentage terms – usually 15%, 20%, and 25% – for longer rides.

So, let's take a quiz. Prior to the addition of card readers, New York City taxi drivers received an average tip of around 10% per ride. Once riders had prompts to help with their decisions of tips, what did the average tip increase to?

1. It didn't increase at all,
2. it increased to 15%,
3. it increased to 18%, or
4. it increased to more than 20%.

Before we get to the answer, we'll look at the underlying concept, a universal phenomenon which is generally referred to as the Power of 3. Humans have a hard time finding and using a scale to calibrate quality. When faced with multiple options, they tend to gravitate to the ‘golden’ or ‘magic’ middle. As Itamar Simonson and Amos Tversky pointed out in their seminal paper on the topic, these decisions have less to do with the actual merits of the middle offer, and more to do with that they call extremeness aversion. The decisions that consumers make are ‘often influenced by the context, defined by the set of alternatives under consideration’. They wrote, ‘[t]he attractiveness of an option is enhanced if it is an intermediate option in the choice set and is diminished if it is an extreme option.’³

There are several factors that determine how strong this effect is. The less frequently someone purchases a product – let's say a snow shovel – the more they may view the middle product and middle price as solutions that simplify their decision making and reduce their research costs. How the customer sees the information spatially also makes a difference. Building on the assumption that people prefer the middle option in a set, Taiwanese researchers Chung-Chau Chang and Hsin-Hsien Liu revealed that ‘an information format compatible with their characteristics makes the middle option more salient and attractive. In detail, the two researchers assert that the middle option becomes more attractive when presented in the middle position and more attractive when the buyer sees all the options together rather than separately.’⁴ A subsequent study, which includes prices with the options, confirmed that the selection of the middle option is strongly influenced by its spatial presentation and not extremeness aversion or its related concept, the compromise effect, which states that people gravitate to a middle option when it seems to be a well-proportioned balance between two extremes.⁵

The tips in the New York City taxicab offer a massive and ongoing test of whether the Power of 3 not only holds true, but also whether its effects endure. So what is the answer to our quiz? It is 18%, which is roughly a doubling of the tips and a very good approximation of the middle option presented to customers.^6,7

The practical implication of these studies is that if you have something you want a customer to choose, you do not give them your single best offer, nor do you focus on one alternative. Remember that the advanced version of the Invisible Game is about influencing decisions, not defining a choice. You offer the customer three alternatives in order to steer their eyes and their preference toward the option you have placed in the middle. To sum up the message in Figure 18.1, if you want people to buy more chicken and rice, you don't offer them better chicken and rice. You offer them sirloin steak.

The effectiveness of this approach is not limited to day-to-day consumers. It affects everyone, and it works best in complex sales situations in which the buyer can easily get overwhelmed or faces artificial pressures such as time. The presentation of three options simplifies their decision making.

The power of decoys

Imagine that you need an analyst with the skills to coordinate, crunch, and interpret a large number of complex datasets and spreadsheets. Your job advertisement attracts a lot of attention online, and your HR assistant helps you narrow down the candidates to two: Skyler, a recent MBA graduate from a prestigious business school in the eastern United States, and Taylor, a post-doctoral physicist from a highly respected Midwestern university. You have no clear choice based on their different backgrounds and personal pros and cons. Skyler had considerable work experience prior to business school, while Taylor wants to make the leap from academia to industry. You connect with each of them on a personal level, and you have full confidence that each would do excellent work and enrich your team.

The one remaining criterion that could separate them is salary expectations, but even that doesn't help as much as you expected. Skyler expects $130,000 and Taylor is asking for around $140,000 per year. That is a slight edge in Skyler's favour. Nonetheless, you still agonize over the decision and ask the two candidates to wait a little longer.

That's when your assistant brings you another candidate whose application came late in the process, after the first round of filtering. This new candidate, Jordan, just completed a PhD in physics at a decent university in midwestern United States. Jordan impresses you so much during the interview process that you now have three candidates who appear equally strong. As you conclude your own final interview with Jordan, you pose the question that will hopefully offer some guidance on whom to select.

‘What are your salary expectations, Jordan?’

‘Well, let me first say that I love atmosphere here, and I hope I'll get the opportunity to work here’, Jordan says. Then the hammer falls. Citing the fresh doctorate in physics and a wide range of proven quantitative skills, Jordan calmly says that ‘I can't settle for a job offer below $300,000.’

You let that number sink in, then smile and nod a few times.

‘Thanks, Jordan, we'll be in touch very shortly. My assistant will walk you out.’

Your assistant returns a few minutes later, leans in the doorway, and looks you straight in the eye: ‘So …?’

The question draws out the same smile you had after your interview with Jordan. Then you reveal the decision.

‘I'm making a phone call this afternoon,’ you say. ‘Taylor gets the job. Let's get the paperwork ready.’

It turns out that Jordan facilitated the decision without ever standing a chance of landing the job. The salary expectation shifted your perception of the value of a doctorate in physics, relative to another graduate degree such as an MBA. Jordan unwittingly became what scientists refer to as a decoy. When two options seem equal, the introduction of a third option can allow someone distinguish between the two otherwise equal options and make a decision.

Decoys can affect choices in several ways. The irony of decoys is that you can strengthen your portfolio – and boost your sales – by including a product nobody wants to buy. Let's look at an example from a complex portfolio. GrindKlemp, a manufacturer of industrial packaging machines, had included four options in its portfolio for years. The four core machines – the A, B, C, D series models – addressed different market segments and helped them carve out a leadership position in terms of both market share and technology.⁸ The machines had two main features: format and parts per minute. The format refers to whether the machine packs powder, gel, or both as sealed plastic capsules. Gel was the newer technology, but both formats were in demand in the market. Parts per minute refers to the number of individual items finished per minute. Table 18.1 below shows the standard prices and the specifications, ranging from the high-end D model down to the best-selling, mid-range C model to the basic A model.

When the venture capital arm of Chicago Entrepreneurs, Inc. (CEI) bought GrindKlemp, it looked for ways to reduce the size of the portfolio from five models to four or even three. The B model looked like the prime target. It had a feature profile that didn't fit its price point very well, and it accounted for only about 3% of the company's sales. That sounded like an open-and-shut case. It was hard to argue against the B model's downward sales trend and its cost burden, never mind the distraction of building marketing and training programmes around five models instead of four.

‘That could possibly be the worst move you could make’, someone said at the meeting where the new owners proposed killing off the B model. The person was a regional sales rep that Chicago Entrepreneurs had kept in her role after the acquisition.

‘Would you mind telling me why?’ said the CEI team leader.

The sales rep proceeded to explain the history of the B model. GrindKlemp had considered killing it off two years earlier, but decided instead to test the portfolio in a couple of markets to see what the effect would be.

‘One of those markets was mine’, the rep said. ‘When we took the B model out, we noticed that people gravitated toward the cheaper A model and away from the C, our best-selling unit. The B turned out to be a very effective decoy. We kept it in the portfolio because it makes the C look like a much better deal. We also did the maths and saw that the revenue gained from selling more C models more than offsets the cost of manufacturing and retaining the B.’

Joel Huber, John Payne, and Christopher Puto of Duke University first documented these kinds of decoy effects, which occur when an item in a set is ‘asymmetrically dominated by one item in the set but not by another’.⁹ In one of their studies, participants needed to choose between taking their date to a five-star restaurant which was a 25-minute drive away or to a three-star restaurant which was 5 minutes away. When the scientists introduced a third option – a five-star restaurant that was 35 minutes away – the study participants on average tended to favour the five-star restaurant located 25-minute drive away.

This is analogous to what GrindKlemp experienced. Practically speaking, the strategic role of the B model in the portfolio was not for customers to buy it. Instead, B's primary role is to make the C model stand out even more than it would objectively. As the researchers from Duke University wrote: ‘Adding such an alternative to a choice set can increase the probability of choosing the item that dominates it. This result points to the inadequacy of many current choice models and suggests product line strategies that might not otherwise be intuitively plausible.’¹⁰

The Power of 3 and the decoy are complementary ideas, not contradictory ones. If the relationship between the three options is balanced or symmetrical, people naturally gravitate to the ‘golden middle’. On the other hand, if the relationship between the three options is unbalanced or asymmetrical, people tend to prefer the option enhanced most by the asymmetry. It has been 40 years since the original fascinating study by the Duke researchers, but their findings remain an undiscovered gold mine whose effects have been underestimated or underappreciated, especially in B2B portfolios. Most key account executives and salespeople are unaware that many purchasers struggle with all the decisions they must make, and that decoys can make it easier for them to decide.

Very efficient levers for choice architectures with decoys are loss aversion, short-term benefits, and the value of zero. It remains up to you to decide precisely how you want you tap the Power of 3 and decoys to nudge your customers consistently and effectively to purchase the product you want them to buy. These techniques will not work every time, but in the long run, they skew the odds significantly in your favour.