UI Design Newsletter – October, 2003

Psychologists and Economists who study how people make decisions describe two types of decisions: compensatory and non-compensatory.

Compensatory decisions

Compensatory decisions are rational decisions. They involve:

• identifying the complete set of attributes that could impact the success of choosing any of the available options – including both positive and negative impacts,
• assigning a relative importance to each attribute,
• computing an overall value for each option based on the impact of attribute and relative weight, and
• selecting the option with the best value.

In compensatory decisions, when the final values for attributes are computed, negative attributes can be compensated for by equal or higher value positive attributes. For instance, a plane ticket that costs $50 more (negative attribute) may ultimately be the better choice because it is a direct flight (positive attribute).

Compensatory decisions are rational. However, people don't make compensatory decisions. Collecting and comparing all of the necessary data is simply too labor intensive.

Non-compensatory strategies are heuristics for decision making

In reality, we typically make non-compensatory decisions. This happens when people:

• don't collect all the relevant information systematically,
• fail to consider the relative importance of various attributes, or
• do not trade off the benefits of some attributes against the deficits of others.

Non-compensatory decision making essentially shortcuts the compensatory process to make the decision making process easier. For instance, in the airline example above, decision makers may simply not collect information on all the relevant attributes. In our airline example, they may consider price but not stopovers or overall flight time.

Alternately, people may find the task of comparing the tradeoffs too much. Instead of weighing price against layover, they may adopt strategies like "drop the choice with negative attributes." Continuing the example, under this strategy the higher priced ticket goes without consideration of the longer travel time. In reality, human decisions all include non-compensatory decision strategies.

"Satisficing," introduced by Herbert A. Simon in his Models of Man in 1957, is a term that combines "satisfy" and "suffice". Satisficing is a non-compensatory decision strategy in which we select the first adequate option, failing to explore or consider the entire set of options. Web site users do this all the time when they quickly scan a page until they get to the first hopeful sounding link name and click that. In a compensatory process they would read all the possible links, identify the most promising one, and then click.

Decisions by design

Could thoughtful information design help users make better consumer choices? Seems like it should. One benefit of the Web is that it simplifies the task of comparison shopping across merchants. The task that used to involve walking back and forth across the mall can now be accomplished in a fraction of the time – without moving from your desk chair. Merchants can organize and present all of the relevant information right at our finger tips to make the decision making process even easier. With relatively minimal work we can use the Web to make a well researched, informed, and even more rational decisions, right?

Unfortunately, shopping sites are not always created with the goal of comparative decision making in mind. How many times have you been online and thought I should be able to make a much better decision, but the site doesn't let me compare...?

There are exceptions. Some sites try to help users make better, or more compensatory decisions. But, does it help?

In a recent study, Jedetski, Adelman and Yeo (2002) hypothesized that Web sites designed to make comparing alternatives easier would lead users to better (or more compensatory) decisions than Web sites that do not. Before starting the study, they briefed their participants on the difference between compensatory and non-compensatory decision strategies so that they could reflect on the decision strategies that they used. After the briefing, participants used two target Web sites (CompareNet and Jango) to select items to purchase.

CompareNet

CompareNet allowed users to specify and display attributes side-by-side, and then compare among the alternatives. The site also facilitated comparison by providing the ability to sort on criteria such as price or reviews. Columns of attributes could easily be added or removed.

Jango

In contrast, Jango presented alternatives in a list for users to sort, but direct comparison required clicking deeper into the site, printing pages, or taking notes on the side. On Jango comparison was possible, but it certainly was not made easy.

Participants shopped independently for answering machines, baby monitors, golf clubs, and toasters. Data was collected using the "think-aloud" protocol. In this process participants think out loud while they shop so that the experimenters can note which attributes are noticed and considered. The "think aloud" procedure provides the experimenters a window to some of the mental processes of how participants made their decisions.

At the end of each "shopping" task, participants and researchers independently indicated whether the participant had used compensatory or non-compensatory strategies for that trial.

Both the researchers and participants agreed that participants were significantly more likely to use a compensatory strategy on CompareNet than on Jango.

Assuming that decision effort is relatively constant within individuals, these findings indicate that site design influences the quality of the participants' decisions. By presenting the information in a format that supported the selection and comparison of critical attributes, CompareNet clearly increased the use of compensatory decision strategies. In economic terms, participants made better decisions when they were using CompareNet.

Participants also reported greater satisfaction with CompareNet. While this is not surprising (we like to think we make good decisions), it is important. The perception that a site is helping or hindering customers to make better decisions, ultimately has impact on brand.

It's not rocket science...

Jedetski and colleague's work shows that the way information is presented on sites influences users decision-making processes. Users recognize when they are making good decisions (as opposed to satisficing) and they prefer sites that help them do that.

Other Web site options that support positive decision making include:

• sorting by price (which entails showing the price...something that sites STILL regularly fail to do!),
• prioritizing decision attributes so that critical options (e.g., price or travel time) appear early, and
• providing assistance or information to help shoppers unfamiliar with a product determine what the relevant parameters of consideration might be (e.g., in Jedetski's study, some participants were not familiar with what to consider when shopping for baby monitors.)

At a time in which many sites still bury prices, merchants and site designers who appreciate the usability of site interactions, and its effect on their brand, are starting to think in more sophisticated ways to support consumer decision making. Users feel better about themselves when they think they've made a solid and well considered decision. That halo may well extend to the merchant in the form of site loyalty.

References

Jedetski, J. Adelman, L., & Yeo, C. (2002). How web site decision technology affects consumers. IEEE Internet Computing, 6(2), 72-79.

Yes. People are systematically bad at making decisions. This can be an opportunity for profit (Look how the gaming industry succeeds because people generally overestimate the likelihood of very unlikely events – like winning the Lottery).

We actually don't design eCommerce sites for good and rational decisions by consumers. We design them for PROFIT. Therefore we can benefit from the tendency for non-compensatory decision making. Putting the high margin choices first is a simple but lucrative example. We might even make the comparison difficult to favor the higher margin choices. By drawing the user down into the site to get a price, we might get them to buy an item without potentially embarrassing comparisons. It might be simply too much work to transverse the site to view the alternative.

But would customers appreciate and return more often to a site designed to support more compensatory decision making? Sites like Amazon prove that customers value information that helps them feel confident in their decisions (even having occasional bad recommendations to give one a solid sense of impartiality). They also DO provide comparison functions (which are probably rarely used). But my suspicion is that customers value reasonable protection from bad decisions and an emotional FEEL of good decision support. They value feeling part of a buying community. This is probably more important than a solid presentation of facts.