95% of new products introduced each year fail.
According to AcuPoll, a research agency out of Cincinnati, as many as 95% of new products introduced each year fail, resulting in massive losses. Considering the investment that goes into developing and rolling out a new product, failing to meet sales goals can be a devastating reality. While some industries fair better than others, the story looks the same across most.
McKinsey & Company conducted a study to better understand what sets those who are successful apart. The results seem simple: "Top performers were twice as likely as bottom performers to research what, exactly, customers wanted." Clearly the answer lies in a true understanding of customer preferences, yet in the same study McKinsey found that "more than 80 percent of the top performers said they periodically tested and validated customer preferences during the development process, compared with just 43 percent of bottom performers." While the statistic serves to underscore the author’s point, what is striking about this is the fact that almost half of the bottom performers are still leveraging research as well, begging the question: if so many companies are using research to understand customer needs and preferences, why is the failure rate so high?
Unfortunately, part of the blame must fall directly on us as market researchers. While of course there are factors which are well outside our sphere of influence, and this is not to say that anything malicious is happening, many market research practices are contributing to this high rate of failure for three key reasons: 1) Customer needs are not understood from the beginning, 2) the wrong methodology is being used, and 3) the data collected is not actionable.
In the case outlined above, at least most of the organizations are attempting to leverage data to drive the process, yet too often these decisions around product design are ultimately made based on "gut feel" by marketers. Until we understand preferences directly at the consumer level, we’re ultimately just taking a shot in the dark, jeopardizing product success and potentially contributing to product failure.
Both qualitative and quantitative data can prove to be extremely useful, but as marketers, we have to understand the advantages and weaknesses of both. The first step in collecting customer data is identifying the problem. This key step informs how the research study is designed, and designing a study without a problem to solve for is a futile exercise. Beginning with the end in mind allows marketers to design the best research process and identify the appropriate methodology. For example, if the goal of the study is to understand how user-friendly a prototype of a product is, a series of focus groups might be appropriate; however if the goal is to forecast demand for a potential new product, a focus group is absolutely the wrong methodology. Understanding methods for data collection is an important piece of expertise that markers bring to the table. How research is conducted can, in the end, contribute to either the success or failure of the product, and as marketers we have the responsibility to employ the best data collection method possible, given time and resources.
Finally, and most importantly, we design poor studies with flawed questions, resulting in data that is not actionable. The ability to collect meaningful data begins with the question design, and unfortunately poorly designed questions are influencing product decisions. Often customer surveys use scales that are abstract and unclear, for example, asking someone who just had their car washed if the service was “Superior, Excellent, Good, Mediocre, or Poor” would result in worthless data. What does “Superior” mean? How is that different than “Excellent” or “Good”? Without any context, or definition around these adjectives, the survey results do not provide any value. In addition to these ineffective scales, researchers tend to use a series of these scale-based questions to get at customer preference and behavior. Unfortunately using these questions as a stand-alone study often results in data that has little to no discrimination – whether or not the scales were designed well. For example, typically when faced with a series of questions asking “what is important?”, customers will indicate that everything is important. Some features may be rated “important” and some “very important”, but even if that’s the case, how can we determine where to focus when all of the data is the same? In the end, this type of data leads us down the road of attempting to be everything to everyone. We build a product with as many features as possible and in the end contribute to product failure rather than product success.
While there are many solutions to these three challenges, one of the easiest ways to approach them is to employ conjoint or trade-off analysis in order to more completely understand customer insights, use a flexible methodology that applies almost universally, and achieve results that are actionable. In a world where we cannot have everything, conjoint leverages the concept of trade-off questions in order to better and more comprehensively understand customer insights. Conjoint studies, supplemented by typical customer research questions, result in data that provides rich insights into both the rank order of customer preferences as well as the magnitude by which customers prefer one feature over another. This type of data gives us the guidance we need in order to pinpoint customer needs and design a product that excludes the features that are unimportant to customers.
Many marketers are using conjoint studies as a part of their research projects (certainly giving them a great advantage over those who are not), but there is one last pitfall that marketers encounter, which is using conjoint data in a vacuum. Conjoint trade-offs measure preferences, these insights are incredibly valuable, but used on their own, they will simply tell us that customers want the best products for the cheapest price. Not surprising; common sense tells us that when given the choice, customers want a $10,000 Ferrari. While we’d love to be able to deliver a sports car for next to nothing, realistically this is not possible. We have to introduce the constant of cost into the equation - that is the cost to developers in order to deliver the product to consumers. By introducing cost information for each of the potential features tested in a conjoint study, we can then optimize based on cost and preference. Optimization allows us to focus not only on the product features that are most popular with consumers, but in addition it allows us to identify the features that are most leverageable. We can determine where we get the highest return on preferences for each dollar invested, truly delivering the best product to consumers at the best price to the company, and ultimately driving product success.