STRATEGY & INNOVATION ARTICLES

If, as studies suggest, companies spend $8 billion per year on consumer research, why do 80 to 85 percent of new product offerings fail?

We think companies can do better. The approach we recommend focuses on understanding the “jobs-to-be-done” (or “jobs,” for short) customers are trying to get done in their lives. The value of this approach was described by Michael Raynor and Clayton Christensen in their bestselling book The Innovator’s Solution. The jobs approach works because it focuses the research on customer problems and the circumstances in which customers encounter them, rather than focusing on a particular solution a company might be hoping to offer. The jobs approach also allows for quite detailed insights to be surfaced that can guide specific product and business model design decisions.
 

Why your customers aren’t telling you what you really need to know

In our experience, most consumers participate enthusiastically in market research. People are typically eager to talk about themselves, and focus groups, interviews, and surveys encourage them — and in most cases, reward them — to do precisely that. However, traditional market research typically assumes that consumers are able to directly provide the needed answers, and, despite consumers’ best intentions, this is often not the case.

Many of the reasons for this disconnect between what your customers tell you and what you really need to know have been described by the fields of psychology and behavioral economics over the past few decades. For example, we now know that an insidious problem with applying traditional market research techniques to gain insight on new-to-the-world business concepts stems from the phenomenon of “cognitive biases,” subconscious mental filters that distort consumers’ answers to research questions in ways that reduce the value of those answers. These biases are always a pitfall to navigate in marketing research, but they can be exacerbated when the research is focused on gathering information about how consumers will act in a completely new set of experiences.

Market researchers have begun to use some of the insights of psychological and behavioral psychology to design research in a way that reduces results tainted by cognitive biases. For example, most market researchers are aware that the specific choices given to a survey respondent can influence their answers; you can increase the number of people saying they are willing to buy a $100 rice cooker simply by adding a $200 rice cooker to the choices, even if no one is interested in the $200 version. (Psychologists might refer to this phenomenon as the “compromise effect” — people like to choose the “middle” option out of a set).

However, while some common psychological pitfalls that lead to cognitive biases are known and can be controlled with careful design, others are more difficult to fix within the context of traditional market research techniques. The main problem, is that the “product-centric” emphasis of much market research does not reflect how customers themselves actually think about their problems, their lives, and their purchase decisions.

What is needed is a “customer-centric” approach to market research design. The jobs-to-be-done approach provides precisely that, enabling researchers to understand in detail consumer pain points and thus the broad landscape of possible solutions, rather than a narrow set of predefined product ideas. Specific solutions can still be tested, but in conjunction with jobs-based understanding, resulting in more reliable insights.

Note that the unique utility of the jobs approach arises from its emphasis on asking the right questions to uncover the right information about the market, and not from employing any specific market research techniques. In fact, the jobs approach can be implemented using any of the traditional market research techniques, including interviews, focus groups, surveys, and ethnography.
 

How jobs-to-be-done overcomes hidden customer biases

The jobs-to-be-done approach frames the problem of understanding the customer as a quest to identify and characterize the fundamental problems (the “jobs”) customers are trying to solve in their lives, for which they are seeking to “hire” products. This framing reflects the fact that customers don’t typically care about purchasing a product as an end in itself. Rather, as a problem arises in their lives that they need to solve, they look around for a product best suited to solve it.

The jobs customers are trying to solve might be satisfied to varying degrees by products and services currently available, or they might be completely unaddressed by any currently available product or service. Important yet unsatisfied jobs provide fertile grounds for innovation.

Contrast this method of framing customer insights with the product-centric or needs-based approaches used by many companies. Simply testing the appeal of a specific product can miss real opportunities if the product being tested does not have the right set of features and tradeoffs. Consumers might still react negatively to a particular product even if it is oriented toward solving highly important, unsatisfied jobs-to-be-done, if the test is framed in a product-centric way. This could result in missing a big market opportunity. However, framing the questions around jobs gets at the root causes of product purchases, and researchers are thereby able to decouple the root cause from any specific product.

The jobs approach also elicits more detailed and actionable information than typical needs-based approaches. It broadens the aperture of jobs explored beyond just functional ones to include emotional and social jobs, which are often as or more important than the purely functional jobs. It surfaces detailed information on the “hiring criteria” customers use when they evaluate different options for getting a job done, which provides important guidance on the specific features responsive offerings must possess.

The jobs approach also gives primary consideration to the circumstance in which a customer has a particular job, i.e. where, when, with whom, and while doing what a customer faces a particular problem. A customer might have the exact same job in two very different sets of circumstances (e.g., “at home” vs. “on-the-go”), and these can provide very different innovation opportunities. Finally, it helps companies to evaluate specific product designs to determine which best satisfies a consumer’s jobs.

We’ll illustrate the advantages of the jobs-to-be-done approach by showing how it overcomes three common examples of cognitive bias in market research: anchoring, emotional biases, and demand effects.
 

Anchoring

Research participants often “anchor” their responses to whatever informational influences are at hand. Just as a real anchor keeps a boat in place, cognitive anchors can tether consumers’ answers to recent information presented, and this can impact research results substantially. Most often, this anchoring influence will not be present in a real purchase situation — and therein lies the problem.

Anchoring manifests itself in many ways in traditional market research. Questions intended to probe product appeal often use the method of calling out and explaining particular product features, then asking respondents to rate these features or the product itself. For example, a respondent might be told about a shampoo’s new scent, and then asked to rate shampoo products containing it. Unfortunately, calling out the scent can cause this feature to be over-weighted in ratings and decisions made during the latter part of the survey.

Anchoring is particularly problematic when it comes to numbers. Since people often have a weak frame of reference for new products, any numbers appearing in a survey can strongly influence subsequent numeric responses — even seemingly benign, profiling data such as the number of items purchased in the last year.

For example, mentioning that a competitive product costs $5 strongly limits the amount a respondent might indicate they’d be willing to pay for a new product. These results both impede the predictive value of the survey and limit a company’s ability to envision new business models or price points. If Whole Foods, Apple’s iPod, or premium bottled waters had been shackled to industry price norms, they would have foregone significant value. And yet, traditional research techniques could have led them down this path due to the anchoring problem.

Framing a study around jobs-to-be-done minimizes the impact of anchoring. Asking about a consumer’s jobs-to-be-done before asking about specific new product ideas — the jobs that the new product is intended to satisfy — enables the researcher to first understand the nature of the unsatisfied job in a research context not clouded by price or product anchoring. Specific insights can be gleaned as to what that consumer feels the characteristics of a potential job-satisfying solution would be. Product concepts can then be developed out of these insights, and subsequent concept testing can include more specific, product-testing questions that are more prone to anchoring bias. This combined approach — and sequencing — enables a more accurate, comprehensive understanding.
 

Emotional biases

For decades, research on how people make decisions focused almost exclusively on the role of reason and thought. More recently, it has been widely acknowledged that emotions play a critical role in the decision-making process. Despite this, market research questions tend to focus heavily on exploring functional aspects of products and services. While this information is important, it is incomplete — the overall puzzle also contains emotional and social pieces.

Even when market research does probe emotional dimensions, the questions are often formulated poorly. For example, people are often asked to “associate” or “rate” a product or brand according to different emotional descriptors. This usually becomes a very academic exercise. After all, how “happy,” “optimistic,” or “proud” does your toilet paper make you feel? These associative techniques are also completely irrelevant for new product categories with which a consumer has no experience and with which he or she has not yet established an emotional connection.

On the other hand, social and emotional jobs are involved in the choice of nearly all products and services. Why does someone buy a particular computer? Functional characteristics like speed and memory are of course important, but, for example, people might also “hire” the Macintosh brand to fulfill an emotional job (e.g., the job of feeling fun and creative) and a social job (e.g. the job of conveying a “cool” image-to others). The jobs-to-be-done approach uncovers the actual social and emotional factors that lead to purchase decisions, rather than trying to force-fit a possibly nonexistent emotional connection to a particular solution.

A related market research challenge stems from the inability of people in the neutral, unengaged emotional state in which one ordinarily takes a survey to predict how they might feel and act in a more relevant emotional state, such as the “enthusiasm state” they are more likely to experience when contemplating buying a product, or the “frustration state” of having an unsatisfied job-to-be-done.

In such cases, respondents may be unable to respond accurately even if they wanted to, because they simply don’t know the “right” answer. The jobs methodology enables researchers to “emotionally prime” respondents by having them imagine the circumstances within which they would encounter a particular job-to-be-done, thus prompting the actual emotional state they would likely be in when making a purchase decision.
 

Demand effects

Survey respondents often modify their answers due to the influence of the survey-taking experience itself. As consumers go through a survey or interview process, they pick up information (or imagine that they pick up information) about the researcher’s agenda, objectives, and desired answers. The corresponding assumptions they make can impact how they respond.

This is somewhat analogous to manifestations of the Heisenberg Uncertainty Principle in physics, in which the very act of observing a particle’s state prevents the physicist from accurately characterizing it. For example, if a survey asks about cars with a particular focus on BMW models, the respondent may accurately deduce that this is a survey conducted on behalf of BMW, and their answers may be different than if they perceived the survey was by, say, Kia.

Demand effects come in many varieties. A common one is the respondent’s sense of what the research is about and what kinds of answers are “expected.” If the respondent has a clear framework for understanding the purpose of the study, they can reply “on auto-pilot,” without thinking too deeply about the true answer.

For instance, suppose you are asked “what’s your favorite restaurant?” You may have an automatic, standby answer, but it’s likely been a while since you really considered this deeply and re-evaluated your answer to see if your preferences have changed as you’ve aged or tried new places. Similarly, many market research studies, by focusing on the product, make it easy for a person to respond with an “auto-pilot” answer that is not necessarily the most accurate one.

Another common demand-effect trigger is the desire of respondents to be helpful, which may result in their trying (consciously or not) to provide the answers they perceive the researcher wants to hear. The opposite behavior (being purposely obstinate) is also possible, particularly when the respondent has experienced some perceived annoyance or insult while going through the survey.

Incentives built in to the particular survey platform can also be a problem: a respondent may answer in such a way to pass the screener and be able to take the survey, particularly if their reward for completion is high. Finally, respondents may answer hastily and without careful thought, simply to get the survey over with.

While demand effects occasionally lead to deliberate deception, it is more common for them to result in responses inadvertently “nudged” in undesired directions. Thus, simply demanding that the person be more “honest” is futile – and may even be counterproductive.

The jobs-to-be-done methodology reduces demand effects, again because of its overall framing through the lens of the consumer’s actual life and problems — decoupled from a particular product, industry, or company. In jobs-based lines of questioning, people are simply talking about the problems, hopes, and pains in their own life, subjects they know intimately well. Thus, the jobs approach is able to reduce implied norms and assumptions in behavior related to specific products. In this way, the survey is able to determine the consumer’s agenda, not reinforce the company’s agenda.
 

Conclusion

Successful market research elicits accurate answers to questions that are critical to a company’s business objectives. But just because you know what you need to know, it doesn’t mean that you can find it out by asking consumers directly. It’s not that people aren’t honest — rather, the way people actually make decisions involves a complex array of factors, including many unconscious cognitive biases, as the above examples illustrate.

The jobs-to-be-done methodology offers a powerful means to minimize and even overcome these biases, thereby getting more accurate — and more actionable — results.
 

Sidebar: Testing the Flying Toaster Concept

To illustrate how the jobs approach differs from traditional techniques, consider a fictitious example. Suppose you are considering launching a completely new-to-the-world product, inspired by computer screensaver graphics of yesteryear: a flying toaster. You have concluded that the product is technologically feasible, and believe that there is a market for it.

However, a number of fundamental questions remain: What is the best set of features for the product? How large is the true market size? And how much would people be willing to pay for it? To answer these questions, you diligently design and launch via the Internet a quantitative survey to potential consumers.

The table below illustrates the difference between how the questions would be framed in a traditional survey approach versus the jobs-based survey approach, and describes the advantage the jobs-based approach has in meeting the research objective.