Editor’s note: Donna Wydra is vice president-consumer product practice for Socratic Technologies, Inc., a San Francisco research firm.  Bill MacElroy is president of Socratic Technologies, Inc.

How many of us have been involved in working on a new product concept that didn’t live up to expectations? Certainly, if you’ve been at it for a while, you have come across at least one example where the research predicted a big win but the launch didn’t generate the predicted enthusiasm with customers. Even as the tools for assessing the price/feature/brand trade-offs have become more sophisticated, we still hear many clients reporting (at conferences and in case studies) that their conjoint and demand modeling techniques didn’t accurately predict a new product’s acceptance. In one case presented recently, a research director shared the horror story of having a multi-round series of discrete choice testing predict one outcome only to have sales reach a mere 12 percent of the predicted results.

Is it that our research tools are just too crude for predicting the success of radically new products? Over time, we have looked deeply into the results and have found that the conclusions aren’t necessarily wrong, they’re just predicting a time period several years out. Of course, competent researchers always use the standard disclaimers that such analyses assume perfect knowledge among customers and fully developed distribution to emphasize the need for a “ramp-up” period. But even when these are factored in, the poorer-than-predicted initial results are still not fully explained.

So what is going on here? What other, hidden barriers or missing components are at work? The answer, we believe, lies in a concept borrowed from academic research on consumer couponing. People who have studied the factors that induce consumers to take advantage of coupons for new products found that to be successful in initiating a trial use, consumers must be “deal prone.” This concept of deal proneness identifies a number of factors that must pre-exist before a coupon activates a consumer response.

In studying the concept of deal proneness, we have discovered a widely applicable concept, that to one degree or another, acts to blunt the enthusiasm for new products entering an established market space. This phenomenon is called preference inertia. Preference inertia, in simple terms, is the degree to which the philosophy of “the devil you know is better than the devil you don’t know” prevents the target customer population from trying new things. We have found that this resistance to change is extremely potent and can blunt sales by up to 40 percent, regardless of the rational feedback given in product concept evaluations.

Preference inertia is produced by the combined effect of three factors: satisfaction with the brand/product currently being used; perceived risk in making a change; and the level of involvement that the customer has with purchasing products in the category. Each of these three sub-barriers will be different for various product categories.

Customer satisfaction with the current brand can be a potent barrier to trying a new alternative. If something is meeting their needs, customers feel less urgency to find and adopt new alternatives. We call this type of satisfaction barrier “micro-loyalty” because it doesn’t necessarily last long. Unless the current product continues to satisfy at sufficient rates, the customer will be tempted to try new alternatives and the preference inertia due to this factor may be overcome.

Perceived risk of making a change can also be a powerful factor in preventing trial of a new product. One example of a product category where risk is particularly high is motor oil. People tend to be loyal to the same primary brand for most of their lives, and it is often the brand their fathers used. Here the risk is embodied in the fear that using a cheap or even different type of oil can hurt the car engine which, in turn, could entail major repair costs. To further exemplify, it is easy to understand why new parents might not want to take a risk on switching the brand/type of baby food or diaper rash cream they use, while just several years later they may not perceive any risk in changing brands of pre-sweetened cereal or sliced cheese for their school-aged kids. Risk can be a potent barrier to trial, particularly if the product plays a role in some expensive, critical and/or urgent life situation.

Involvement with a category is another way of asking “how much does the customer really care?” If trying a new product involves time, education, practice and/or changes to the way the work- or life-environment is arranged, many people would rather not bother. Involvement is also related to the innate awareness of change within a category. If the customer isn’t actively scanning the market looking for a new solution, odds are that the initial message announcing a new alternative is falling on deaf ears. This barrier is most prominent in categories where the product or service is shopped infrequently or in those for which the purpose of the product is considered mundane, such as bleach or baking soda.

The practical application of the preference inertia effect is to deflate the estimates of product take rates over some initial period of time. As one might expect, preference inertia is not permanent and decays over time. The precise level of decay cannot be completely measured up front because it is affected by both changing customer needs and the level of spending to promote trial. But what we can do is estimate the “unadulterated” level of resistance in the first year or so of the product’s lifecycle.

In finance, the application of the preference inertia effect is to deflate the net present value of the cash flow in early years from the introduction of new products (by reducing revenue projections). This deflator (which we call the Phi deflator) is applied most heavily in the first year, and diminishes over time. A model for the level of Phi deflation that we have used in categories with high levels of preference inertia is shown in the graph and begins at a 40 percent level (anticipated volume in Year 1 is only 60 percent of the predicted long-run take rate) and reduces to 0 percent by Year 5. Where preference inertia is less potent, the level of effect and its duration can be much lower.

This concept of a demand deflator can also help marketing to rationalize the findings from choice-based and trade-off analytical techniques which, as previously mentioned, can overstate the initial take rates by a wide margin. If we assume that the levels of relative utility generated by our conjoint and configurator analyses are correct, but simply need to be deflated in the early periods due to varying levels of preference inertia, then the findings are more useful and realistic.

The level of preference inertia effect differs by category and the customers that make up the target audience for the products and services. We recommend taking measurements in conjunction with product volume measurements and calibrating the category results for use in future product introduction planning.