Human Hardware: The Paradox Of Choice

In the last few years, there have been two best selling books published that seem diametrically and philosophically opposed: The Paradox of Choice by Barry Schwartz and The Long Tail by Chris Anderson. Both explore the explosion of choice that we have in our modern world and how we might be equipped to navigate through these choices. Today, I want to look at the value of choices in human hardware terms and how this impacts our interactions with search.

Choices, Choices Everywhere

First, let’s begin with Schwartz’s theory. Schwartz begins with a fundamental dogma of western industrialized nations: human welfare = freedom = choice. This has become accepted as an essential truth and has been baked into our consumer culture to the point where Schwartz’s local supermarket carries 285 varieties of cookies, 75 of iced tea, 230 soups, 175 salad dressings, and 40 toothpastes. The number of choices available on our store shelves has exploded, all based on our cultural conviction that more is good.

But Schwartz, building on the theories of Amos Tversky, Daniel Kahneman, and Herbert Simon, argues that the expanding number of choices places a cognitive burden on us. It’s stressful making choices, leading to increasing use of heuristic shortcuts in narrowing our choices, Simon’s classic model of satisficing. We arbitrarily narrow our choices, with each narrowing requiring cognitive effort. The more choices we have, the more narrowing and effort required. And if the decision is about something of minimal importance, such as the picking of the best salad dressing from the store shelf, the amount of effort required outweighs the potential benefit. We evaluate the amount of cognitive effort we put into decisions based on the benefit, or expected utility, we would get out of the decision. If we’re forced to put too much effort into a decision for too little return, we walk away.

Cognitive Jamming

Columbia University researcher Sheena Iyengar demonstrated this phenomenon, called analysis paralysis, in an oft cited study involving jam tasting in a California grocery store. She set up a tasting booth with a variety of exotic jams. Half the visitors to the booth had 6 jams to choose from and half the visitors had 24 jams to choose from. The objective of the study was to see if the number of choices presented had an impact on the amount of jam sold. It did. 30% of the visitors to the 6 choice booth bought jam, while only 3% of the visitors to the 24 choice booth ending up buying jam. Too many choices “jammed” our heuristic decision-making mechanisms, causing a complete shut down.

But what about Anderson’s Long Tail? In his book, Anderson also refers to Iyengar’s study (pages 170 – 172) but has a different take on it. He argues that it’s not so much the number of choices available to us, but the tools available to us to filter and order our alternatives. As he notes, his local grocery story has a lot more than 24 choices of jams, yet they still sell lots of jams. The difference is in how we make the decisions. Anderson quotes the conclusion of Iyengar and her colleagues in another study, “Knowing What You Like versus Discovering What You Want: The Influence of Choice Making Goals on Decision Satisfaction”:

“Despite the detriments associated with choice overload, consumers want choice and they want a lot of it. The benefits that stem from choice, however, come not from the options themselves, but rather from the process of choosing. By allowing choosers to perceive themselves as volitional agents having successfully constructed their preference and ultimate selection outcomes during the choosing task, the importance of choice is reinstated. Consider the request in Forbes’ recent ‘I’m Pro-Choice’ article: “Offer customers abundant choices, but also help them search.” We now know how.”

Combing the Long Tail

Anderson’s Long Tail shows that hundreds or thousands of choices can be a good thing, as long as we have ways to filter and order our options. We need ways to reduce the cognitive load, bringing it in line with our expected utility from the decision. If this wasn’t true, the iPod and iTunes, as well as Amazon, eBay, and the entire e-tail movement would all be abject failures.

Of course, the ultimate online filter is web search. Google is our satisficing tool of choice, universally applicable and ubiquitously available. And it’s in this usage that the paradox of choice again becomes apparent, but in a unique way.

Choice in Search Engine Results

Recently, we did an extensive study with Google in Europe. This was an extension of the Brand Lift of Search study we did with Google in North America. The goal was to see how the inclusion of well know brands in key positions in the Golden Triangle impacted consumer awareness and choice in buying decisions. As a quick recap of the original study, we found there was significant lift in both brand awareness and likelihood to consider for purchase when brands appeared prominently in search results sets.

In the European study, we wanted to explore which combination of positions yielded the best lift. So we tested different combinations of presentation of our test brands in four different market verticals: travel, financial services, fast moving consumer goods (CPG in North America), and consumer electronics. We set up three different presentations:

Option 1 – Test brand in both top sponsored and top organic position
Option 2 – Test brand in top organic and competitive brand in top sponsored
Option 3 – Test brand and competitive brand in top sponsored, test brand in top organic

We then tested which combination provided the best performance for our test brand, measured in terms of recall, purchase intent (both captured in a post-session survey), and clicks. In all three metrics, Option 3 performed better for the test brand, better even than option A, where the test brand had the same placement, but no competitor showed. In this case, choice proved to be a good thing for our test brand.

The human Need for Choice

This brings us to another human hardware issue. We need choice. We prefer options, even if we never choose them. In the book Predictably Irrational by MIT professor Dan Ariely, he calls it “keeping doors open.” A study involving a computer game in which participants clicked on one of three doors to earn money showed that even when they know which door would consistently yield the highest returns when clicked on, if participants knew that unclicked doors would disappear forever, they would keep clicking on them just to keep the doors as options. They never intended to use them, but they still wanted them there.

I believe this is the effect we saw in the Google study. When people were presented with a results set that included a couple of well-known brands in a prominent position, they felt more comfortable with the quality and relevance of the results set. And the closer those two brands were together in terms of physical location, the higher the perceived quality of the results.

This brings up another interesting aspect of our search decision model. For a search where we already have some experience with the product category, we pre-judge the results set based on the sites we expect to see appear. For these types of searches, we already have some brand relationships and we feel a sense of comfort when we see these brands appear near the top of the page. If we don’t see them, we begin to question the quality of the results.

Looking for the missing brand

In another study, we let participants search for laptop computers and we placed our test brand near the top of the results in various positions. But unlike the Google study, we didn’t include a well-known competitor. We let our test brand dominate the results set unchallenged. But when we analyzed our eye tracking results we noticed a non-typical tendency on the part of participants to hunt around the page, obviously looking for something. After a quick scan of the Golden Triangle (a few seconds), we saw an unusual relocation of scanning to the top of the right sponsored ads. In typical search interactions, scanning of the right rail is only done after a scan of the Golden Triangle, usually 8 to 10 seconds in duration. Even more typical, scanning of the right rail happens after a click or two in the Golden Triangle. But in this study, participant after participant was scanning the right rail almost immediately, obviously looking for something. Finally, after one of the sessions where we saw this happen, we asked the participant why they were doing this. The answer was they were looking for the competitor, whom they expected to see at the top of the page. They didn’t feel comfortable with the results set unless they had options to choose from.

What are the takeaways? Humans need choice, and if we can filter and order our results, more choice is generally better. But, as we start to consider our choices, there are finite numbers that we can cognitively consider at a time. One is too few, and dozens are too many. We use web search to help with the paradox of choice, and the Golden Triangle is the real estate where those decisions are made.

Gord Hotchkiss is CEO of Enquiro, a search marketing firm that produces search engine user eye tracking studies and other research. The Just Behave column appears Fridays at Search Engine Land.