Editor's note: Bryan Orme is a customer support consultant with Sawtooth Software, a Sequim, Wash., maker of marketing research software.

Market researchers face two challenges as they provide market intelligence for managers. First, they must meet managers' objectives with useful, valid results. Second they have to communicate those results effectively. Failure on either of these points is fatal.

Conjoint analysis provides useful results which are easy for managers to embrace and understand. It is no wonder that conjoint analysis is the most rapidly growing and one of the most widely used market research techniques today.

This article discusses the benefits of conjoint that managers are most likely to embrace and highlights a dangerous pitfall to avoid when presenting market simulators.

Realism begets better data

Many managers have limited experience with statistics and can be skeptical of or intimidated by advanced methods like conjoint. Unfortunately, simpler approaches can be unrealistic or even useless. Suppose we are conducting a study about laptop computers. Consider the following question:

Q10. When purchasing a laptop computer, how important is . . .

(Circle One Per Item)
Not Important Very Important
Brand 1 2 3 4 5 6 7 8 9
Battery life 1 2 3 4 5 6 7 8 9
Processor speed 1 2 3 4 5 6 7 8 9
Weight 1 2 3 4 5 6 7 8 9
Price 1 2 3 4 5 6 7 8 9

Respondents can answer this type of question rather quickly. The average respondent answers with high ratings, while the bottom half of the scale is largely ignored. This results in sub-par data for statistical analysis: skewed distributions, with typically little differentiation between attributes. Such "self-explicated" importances reveal little about how to build a better laptop. How much battery life will buyers trade off for a given increase in processor speed? Further, stated importances often don't reflect true values. It may be socially desirable to say price is unimportant - after all, respondents don't want to appear cheap. Yet, in real-world laptop purchases, price may become a critical factor.

Even though many managers won't understand (or care to) the statistical shortcomings of self-explicated data, they should agree that this exercise can't be very realistic. They'll concur that buyers can't always get the best of everything in the real world; buyers must make difficult trade-offs and concessions. When people are forced to make difficult tradeoffs, we learn their true values. Convince managers of this, and you are well on your way to helping them understand the value of conjoint.

Conjoint aims for greater realism, grounds attributes in concrete descriptions, and results in greater discrimination between attribute importances. Conjoint creates a more realistic context. Figure 1, from Sawtooth Software's Adaptive Conjoint Analysis (ACA) illustrates this.

Which of the following laptop computers would you rather purchase?

486 DX/2 (7-hour battery life) $2,250 or Pentium 75 (5-hour battery life) $2,750

Strongly Prefer Left Strongly Prefer Right
1 2 3 4 5 6 7 8 9

Of course, conjoint questions can also be asked one product profile at a time, as in traditional card sort. The rationale behind pairwise comparisons is this: people can make finer distinctions when they directly compare objects. For example, if I hand you a 4-pound rock, take it away, and then hand you a 5-pound rock, chances are you won't be able to tell me which is heavier. However, if you hold one in each hand, you'll have a much better chance of guessing which weighs more.

Another flavor of conjoint, choice-based conjoint, offers even greater realism. Figure 2 shows how Sawtooth Software's Choice-based Conjoint (CBC) approaches the same subject.

Choice-based conjoint questions closely mimic what buyers do in the real world. Including "none" as an option enhances the realism and allows those respondents who are not likely to purchase to express their disinterest. Choice-based data reflect choices, not just preferences. If we agree that the ultimate goal of market simulators is to predict choice, then it's only natural that we would value choice-based data.
Managers don't need to know about orthogonal designs, main-effects assumptions or how utilities are derived. You will probably bore (or even annoy) them if you tell them. Instead, managers need to grasp that realistic models result from realistic questioning methods and be comforted that conjoint is a reliable, time-proven method.

Brand equity

Conjoint provides useful results for product development, pricing research, competitive positioning and market segmentation. Conjoint can also measure brand equity, which is an especially critical issue for many managers.

Brand equity encompasses the intangible forces in the market which allow a product with a brand name to be worth more to buyers than one without. High equity brands command higher prices and are less price sensitive. Since brand equity goes directly to the bottom line, it's no surprise that managers are focused on brand equity.

Choice-based conjoint offers a reliable way to measure brand equity. It presents respondents with varying product configurations and asks which they would purchase or choose. Each brand is presented at different prices throughout the interview. The percent of times respondents choose each brand at different prices reveals preference and price sensitivity for brands. Compelling demand curves result when we plot these "wins" by price point and connect them with smooth lines, as shown below (Figure 3) for three hypothetical brands of pain relievers: Renew, Balmex and PainFree.

If the brand manager for Renew wants to quantify the price premium it commands over the other brands, choice-based conjoint reveals the answer. We use the same demand curves as above as a starting point
This time, as shown in Figure 4, we've drawn a horizontal line through points A, B and C to represent a level of equal relative demand or preference. If Renew is priced at $3.90 and Balmex at $3.50, respondents on average will be indifferent (have the same preference) between the two. This 40-cent difference (Point C minus Point B, or $3.90-$3.50) represents the premium, or brand equity that Renew commands over Balmex. Similarly, Renew commands a 60-cent premium over PainFree (Point C minus Point A).

Another approach to assessing brand equity results from comparing preference if all brands were offered at the same price. Imagine that we continue drawing the vertical line from $3.50 through point B until it intersects Renew's demand curve. That point is a relative preference of 32. At $3.50, Balmex and PainFree capture preference of 22 and 16, respectively. At this price, Renew is preferred by a ratio of 32/22, or it captures 45 percent higher preference than Balmex. Similarly, Renew is preferred by a ratio of 32/16, or 100 percent over PainFree.

Strategic pricing research

In an ideal world, researchers could accurately measure price sensitivity by manipulating prices in test markets and measuring changes in demand. While scanner technology has made this sort of analysis more feasible than ever before for many categories of consumer goods, these real world experiments face crippling hurdles. Market forces don't remain constant for the duration of the experiment: macro economic forces can alter demand; competitors change their prices and/or promotions; buyers stock up to take advantage of lower prices; new products are introduced. While conjoint pricing experiments are not as realistic as the real world event, conjoint experiments hold market forces constant. The relative preferences and sensitivities we observe in the controlled experiment should be borne out in the real world.

In the previous demand curve example, Renew holds the enviable position of being preferred to Balmex and PainFree at all price levels. Notice also that the demand curves are not parallel: Renew's preference declines at a slower rate than the other brands as price increases. Respondents are less price sensitive toward Renew than the other brands. The ability to measure unique price sensitivities by brand is an advantage choice-based conjoint enjoys over traditional main-effects-only conjoint.

Demand curves provide strategic information for pricing decisions. Suppose Renew's manager is considering initiating a price cut. Renew is the market leader, and her past experience suggests that the discount brands will react with similar price cuts. She could learn a great deal using conjoint data - enough to avoid a mistake. The slopes of the demand curves show that as prices are lowered, Renew will gain share at a slower rate than Balmex or PainFree. If she lowers price and the other brands follow, Renew's market share and profits should decrease.

Price sensitivity can be quantified for each brand by examining the ratio of preference at the highest price versus preference at the lowest price. Alternatively, the price elasticity of demand (defined as percentage change in demand divided by percentage change in price) can be easily calculated for each brand in a CBC study.

Some managers have been so pleased with this approach that they have funded wave after wave of conjoint tracking studies. They compare demand curves from each time period to quantify changes in brand equity, to gauge the results of previous pricing or other marketing mix changes, and to formulate future strategy.

Conjoint predicts preference, not market share

I was recently involved in a choice-based conjoint study for a manufacturer of personal computers. Our main contact was the pricing manager whose objectives were to measure market awareness, preference and price sensitivity for his sub-brands along with major competitors. We conducted the study disk-by-mail and were soon delivering top-line conjoint results.

Our client was skeptical when he saw that conjoint reported that one of their newly released brands, we'll call it "FastPC," was beating their well-established brands hands down. He insisted this couldn't be right and that we check the data. We did - somewhat nervously, I might add - but found no errors. In the meantime, he called his sales department for a sanity check. Sales reported that the FastPC was flying off the shelf. FastPC had exceeded all expectations.

While this happy-ending story warms us inside, it also illustrates a limitation of conjoint. Conjoint predicts preference, not market share. While the newly released FastPC was selling above expectations, its market share at that point still fell short of established brands. Given enough time, adequate promotion and distribution, we'd expect FastPC's market share to more closely align with conjoint results.

Conjoint models do not predict market share due to a variety of reasons, including:

1. Conjoint assumes perfect information. In the conjoint interview, respondents are educated about available brands and features. In the real world, obscure brands have less chance of being purchased. Conjoint cannot fully account for differences in awareness or preference developed through advertising and promotion.

2. Conjoint assumes all products are equally available. One brand is as conveniently selected as another in a conjoint interview.

3. Conjoint respondents might not accurately reflect potential buyers. Many won't have the interest, authority or ability to purchase at the current time.

4. Conjoint results reflect the potential market acceptance of products and services, given proper promotion, distribution and time.

Many researchers quantify factors conjoint cannot account for and build them back into the model using external effects. While this practice typically brings conjoint results more in line with actual market share, it draws us toward a troublesome paradox. As factors are accounted for to more accurately tune the conjoint model to market share, we become more likely to believe we actually have developed a valid market share predictor and more likely to misuse the model. Imagine the potential damage if costly resources are committed based on the assertion that, "It's worth it because the simulator predicts that market share will increase from 17 percent to 23 percent, which translates into an additional $8.27 million in revenue per year."

That said, conjoint models are excellent directional indicators. Conjoint can reveal product modifications that can increase market share but it will not reveal by how much market share will increase. Conjoint can tell us that the market is 20 percent more price-sensitive for Brand A than Brand B but we do not know the absolute price sensitivity of either one. Conjoint can identify which market segment will most likely purchase your client's product but not how many units it will purchase.

Summary

Conjoint analysis increases the return on research dollars by providing managers with useful, valid information. Conjoint's realism leads to more accurate results, and provides a strategic tool for quantifying brand equity and relative price sensitivity. To ensure success, researchers must carefully set management's expectations regarding what conjoint can and cannot do.

The market simulator is usually the most anticipated deliverable for managers. Don't let this enthusiasm get out of hand. Conjoint simulators are directional indicators which can provide a great deal of information about relative feature importances and preferences for product configurations. Conjoint simulators are not market share predictors. Many other factors such as awareness, distribution, advertising and product life cycles drive market share in the real world. While conjoint models can be fine-tuned to partially account for these elements, we mustn't let managers believe that adjusted conjoint models can accurately predict volumetric absolutes such as market share.