Many innovative new products don’t succeed in the marketplace. One common reason: Companies don’t focus enough on understanding how customers evaluate products and make purchase decisions.

A lot of great new products fail — and companies often wonder why. Although the companies were careful to listen to their customers, the products still failed. This is not a rare occurrence. A recent study of almost 9,000 new products that achieved broad distribution at a national retailer revealed that just 40% of them were still sold three years later.1 (See “About the Research.”) Some of these products did not create value for customers and deserved to fail. However, many would have created value if customers had adopted them. But customers could not, or did not, recognize their value.

While most companies focus on customer needs, they do not think hard enough about how customers decide what to purchase. We now have ample insight into how customers evaluate new products. Yet companies generally focus primarily on creating value — without enough regard to whether customers will recognize this value.

To decide what to buy, customers need to know what products are available and how their features vary. Whether you are an airline choosing which aircraft to purchase, a college graduate choosing your first car, or a parent buying diapers for your infant, there are only two ways you can collect this information. You can search, or you can infer. The inference process uses the information you can search for to guess the information that you cannot easily search for. We will start by discussing the search process before turning to the inference process.

Customers’ Search Process

In July 2012, United Airlines Inc. announced a large commercial aircraft purchase, with an agreement to purchase 150 Boeing-737 aircraft for $14.7 billion. The deal took a year just to negotiate,2 and before that, the team from United engaged in extensive research trying to understand the capabilities of different aircraft and the costs of operating and servicing them. Imposing structure and discipline on this search process is one of the primary roles of a procurement department. The length and intensity of the search process is a function of its cost, the importance of the decision, and the customer’s expertise.

For a customer, the perceived benefit of searching for a better solution may not be the same as the actual benefit, particularly in markets with little recent innovation. This poses a challenge for significant innovations in such markets; customers may not find these innovations because they do not know to look. For example, when British shower manufacturer Aqualisa Products Ltd. developed an innovative new shower system for the U.K. residential market, customers did not initially adopt the product because they were accustomed to the U.K.’s low water pressure, which this product addressed. Customers did not realize that a better shower system was possible and therefore saw no reason to look for one.3

In other situations, the benefits of information are clear, but the cost of searching for it is simply too high. For example, when flying into New York City’s La Guardia Airport, many of us will use a yellow taxi to get to Manhattan. In 2014, there were more taxi accidents in New York City than there were taxi medallions.4 The implication: It is important to find a safe taxi driver. But what do we do when exiting the terminal carrying our luggage? Most of us jump into the first cab in the line. Although the benefits of finding a safe taxi driver are clear, the cost of searching for it is simply too high — particularly in settings where passengers are encouraged to just take the next available taxi. One suggestion: Walk around the taxi before you get in and see if there are any big dents, particularly around the passenger compartment.

The relationship between how much customers search and their prior expertise is surprising. In an influential study using data of new car purchases, researchers Sridhar Moorthy, Brian Ratchford, and Debabrata Talukdar showed that the relationship can be an inverted U.5 (See “The Relationship Between Expertise and Search.”) Customers with a lot of expertise (experts) may not search, because they think they already know what’s best. For example, a pharmaceutical company was recently surprised when doctors did not prescribe its new drug. The drug was a fabulous new product that created tremendous value for customers: It treated the disease more effectively and had fewer side effects than existing treatments. However, it was the first innovation in this therapeutic area for many years. Doctors believed that they already knew all there was to be known about how to treat the disease, and so their minds were not open to the possibility of new treatments.

At the other end of this spectrum are customers who are clueless. They do not search because they do not know which questions to ask, where to find the answers, or how to interpret the information if it arrives. For instance, I once had a friend call on a Saturday morning and ask for help buying a bicycle. Fine plan (when you are clueless ask for help), but poor execution (I knew nothing about bikes). Not wanting to disappoint my friend, I accompanied her to the bike store, where we explained that she needed a bike, and I was there to help with her decision. The enthusiastic young salesman started describing the technical differences between the bikes. This went on for 20 minutes until my friend’s eyes were completely glazed over. She stopped the salesman, looked him in the eye, and said, “I want a red bike.” The salesman responded, “Well, I better take you to the red-bike section then.” We walked out 10 minutes later with a red bike. Think for a moment about the implications of this example for a bike manufacturer. Technical innovation would not increase the chances of a sale to this customer, no matter how much value the innovations created. More generally, the risk for companies is that they invest in innovations that customers cannot recognize.

Consumers’ Inference Process

When search is incomplete, we shift to forming inferences; we use what we can observe to infer what is too costly or too difficult to search for. We have already seen one example of a cue that customers can use to form inferences: using dents in taxis to infer the quality of the driver. McDonald’s Corp. offers another illustration. The fast-food chain has long emphasized to franchisees the importance of keeping a restaurant’s parking area clean.6 Why, you might wonder? Customers do not really care about the parking area. What they care about is the cleanliness of the kitchen, and perhaps the bathrooms. But what can customers see when they drive past? They may use the cleanliness of the parking area (what they can see) to infer the cleanliness of the kitchen and bathrooms (what they care about). What is most surprising is that when you ask customers why they did not choose to stop, they often cannot tell you; they just did not feel comfortable stopping at that restaurant. In other words, this is not a conscious thought process; it is operating at the subconscious level, which makes it both pervasive and powerful.

We can illustrate this principle through a visual example created by vision science professor Edward H. Adelson.7 (See “The Checker Shadow Illusion.”) On the left is a checkerboard, with a black square labeled “A” and an apparently white square labeled “B.” In fact, squares A and B are both the same shade of grey. We demonstrate this on the right, where we have removed the surrounding context. Even more remarkable, now that you know this is true, look back to the left and try to tell your eyes that the A and B squares are the same color. You can’t. Our eyes are leading us to inferences; we do not know that it is happening, and even if we did, we could not do anything about it.8 Although purchasing decisions are a different neural process than this visual process, a similar phenomenon occurs when customers are evaluating different products or services. Customers often do not realize they are forming inferences, and even if they do, they are powerless to stop it.

The Role of the Brand

The most common cues we use to infer product quality are price and brand. In a business-to-business setting, signaling information about quality is essentially the only role of the brand. Even in consumer markets, this signaling role is hugely important.9 However, in some consumer markets, brands may signal more than just product quality; consumers can also use brands to signal information about themselves.10

In many circumstances, the more effectively customers can search, the less they will rely on the brand. Moreover, their perceptions of the brand will change quickly as new information comes in. Both factors diminish the importance of the brand. However, in markets where customers cannot search easily and effectively, they are forced to use the brand to make purchasing decisions. For example, if customers want to buy a reliable laptop computer with state-of-the-art features, they have two options. If they are knowledgeable, they can search computers’ spec sheets and reviews to evaluate the computers’ features. Alternatively, they can form an inference based upon the computer brand. For customers who lack the expertise to evaluate computers’ features, relying on the brand may be the only option.

In general, customers prefer to search. It is usually only when they are unable to search effectively, either because they lack expertise or because the cost is too high, that they rely upon the brand. Second, the role of the brand may vary across customers. A novice computer buyer may lack the expertise to search and be forced to rely upon inference. Most computer buyers have enough expertise to simply engage in search.

Similarly, we can expect the brand to play a more prominent role for prospective customers than for existing customers. This became apparent to a consulting firm that was innovating in its consulting processes. In a professional-services setting, as in any services setting, there is generally no spec sheet, and prospective customers are often forced to rely on inference. As a result, the firm’s prospective customers are unlikely to reward it for its process improvements, although they may infer the quality of the firm based on the strength of its brand. This contrasts with the firm’s existing customers, who experience the process improvements firsthand.

Third, the role of the brand may vary across product features. Features that are on the spec sheet, such as the size of a laptop’s hard drive, are typically features than can be discovered through search. However, features that do not appear on the spec sheet, such as reliability or ease of use, are not easily discovered through search (although reviews from trusted sources can be helpful), and thus it is these features for which the brand’s role will be more prominent.

The Impact of the Internet

How has the Internet affected the role of the brand and the way that customers make purchasing decisions? We should first recognize that for some products, or at least some product features, the Internet has had relatively little impact on how customers make purchasing decisions. For example, the Internet does not help us search for information about future events, such as how Volkswagen AG will respond to future product recalls, or how well security software will protect against the next generation of online threats. It is also less helpful for product attributes that require physical inspection, particularly when the needs are specific to the user (such as the fit and appearance of a swimsuit).11

The Internet has had a profound impact on the way that customers evaluate many other products, however. It has done so in two ways. First, it has lowered the cost of search by making information more accessible.12 The initial impact of this change was perhaps smaller than anticipated in some markets. When customers shop in physical stores, they do not have ready access to a laptop or desktop computer. It was not until the advent of smartphones that the Internet substantially lowered the cost of search in many markets.

The second impact of the Internet has been to broaden the number of product features that are searchable. User-generated content, including blogs and product reviews, now mean that customers can search for information about features that were previously unsearchable, such as the quality of a golf course or the fairness of a contractor.

Consider how this has changed the restaurant market. It used to be that tourists in a new city had little choice but to ask the hotel concierge for a recommendation or choose a restaurant that had a big national brand. This was a good outcome for chains, including the Hard Rock Cafe. How do tourists make restaurant choices now? They pick up their smartphone and query or They can compare prices, menus, location, even how politely they will be greeted by the maître d’. This has sharply diminished the role of the brand. Now small, innovative restaurants offering great food and service become easier to discover, while the Hard Rock Cafe has closed some of its locations.13

Does this mean that the Internet has increased or decreased price competition? If customers can now recognize greatness where they previously could not, companies with innovative products will benefit, as customers will pay a larger premium for those products. On the other hand, in cases where customers were paying a premium for differentiation that was perceived rather than actual, the Internet will foster competition and undermine that premium.14 For example, if a viral video revealed that national brand and private-label vitamins are identical products made in the same factory, some consumers would probably stop paying a premium for the national brands.

What Should Companies Do Differently?

For companies engaged in innovation or product development, the implications are clear: They need to ensure not just that they create products that create value for customers but also that customers can recognize this value. We recommend that companies focus on three sets of questions. (See “Will Customers Discover Your Innovation?”)

First, are customers motivated to search? Do they recognize that there could be a better solution, and are they willing to invest effort to find that solution? Recall the Aqualisa example: Because customers did not know a better shower was possible, they were not motivated to look for it. How much does the solution differ from existing options? If this is the first major innovation in the industry for 20 years, customers are less likely to be searching for new alternatives than if the industry has had a steady stream of major innovations. A concrete measure of how motivated customers are to search is the length of their procurement or decision process. For example, customers on average spend 15 to 20 hours searching for information when buying a new car.15 However, few customers will invest this amount of time when choosing detergent for their dishes.

Second, are customers able to search effectively? Traditionally, the litmus test for this question was whether the information was listed on a spec sheet. Recall the laptop computer example: The size of the hard drive is on the spec sheet and searchable, but the reliability and ease of use are not. As long as customers have the expertise to interpret the spec sheet, features on the spec sheet are generally searchable. If the decision is important enough but customers lack expertise, they may still be able to search by turning to expert advisers. Examples of expert advisers include doctors, financial advisers, real estate agents, insurance brokers, IT consultants, and workplace-benefits consultants. In some markets, customers can use customer reviews to search for information that is not on the spec sheet. In this case, the spec sheet is no longer as good a litmus test of whether customers are likely to search rather than infer. One factor that has not changed: Customers typically adopt a decision process, and changing this decision process is difficult. For example, if customers have always searched for the best deals on cars by waiting until the end of the model year, then convincing them to purchase earlier in the model year will be difficult.

If customers cannot search, companies need to understand what cues they will use to infer the absent information. It is an indication of how well McDonald’s manages its restaurants that it knows not just that customers infer restaurant cleanliness from the state of the parking lots but also that customers do this subconsciously. You may have to create cues to help customers with this inference process. For example, automobile manufacturers would like to convince customers that the engines and transmissions in their cars are precisely engineered. Because the quality of the engineering of these components is not observable to customers, car companies instead highlight engineering features that are observable. Recall the car advertisements in which ball bearings roll along door seams. Most customers don’t really care how precisely ball bearings track on their car’s door seams. However, the manufacturer has provided a cue to infer the quality of its engineering. Customers can use this cue to evaluate the engineering of parts that they cannot observe or do not have the expertise to evaluate.

Helping Customers Recognize Innovation

Developing great products is not enough to succeed in business — companies have to develop great products that customers can recognize as great. Fortunately, the way that customers collect information and make purchasing decisions is now understood. Rather than merely asking what customers need, companies have to understand how customers will evaluate which products will satisfy their needs.

If customers are motivated to learn about products and have the expertise to interpret what they learn, then we can expect the search process to play an important role in their decisions. This is a welcome situation for innovative companies, as customers are more likely to recognize their innovations. However, when customers are either not sufficiently motivated or not sufficiently informed, then search will give way to inference. This makes it much less likely that customers will recognize innovations. The implications for companies are clear: Focus development on innovations that your customers will easily recognize, or find ways to alert them to innovations they may not detect on their own.


1. E. Anderson, S. Lin, D. Simester, and C. Tucker, “Harbingers of Failure,” Journal of Marketing Research 52, no. 5 (October 2015): 580-592.

2. C. Isidore, “Boeing Wins $14.7 Billion Jet Order From United,” July 12, 2012,

3. Y.E. Moon and K. Herman, “Aqualisa Quartz: Simply a Better Shower,” Harvard Business School case no. 9-502-030 (Boston, Massachusetts: Harvard Business School Publishing, 2002, revised 2006).

4. See L.P. Banville, “Taxi Cab Accidents in New York: 1999–2014,” December 10, 2014,; and New York City Taxi & Limousine Commission, “2014 Taxicab Factbook,”

5. S. Moorthy, B.T. Ratchford, and D. Talukdar, “Consumer Information Search Revisited: Theory and Empirical Analysis,” Journal of Consumer Research 23, no. 4 (March 1997): 263-277.

6. J. Pepin, “Burger Meister Ray Kroc,” Time, Dec. 7, 1998.

7. See E.H. Adelson, “Checkershadow Illusion,” 1995,

8. For another surprising example of these visual effects, see R.B. Lotto and D. Purves, “The Effects of Color on Brightness,” Nature Neuroscience 2, no. 11 (November 1999): 1010-1014.

9. Prominent studies of the signaling role of the brand include T. Erdem and J. Swait, “Brand Equity as a Signaling Phenomenon,” Journal of Consumer Psychology 7, no. 2 (April 1998): 131-157; and B. Wernerfelt, “Umbrella Branding as a Signal of New Product Quality: An Example of Signaling by Posting a Bond,” Rand Journal of Economics 19, no. 3 (autumn 1988): 458-466.

10. This second signaling role of the brand is particularly important if consumption is conspicuous. For example, when we wear a Rolex watch, drive a BMW vehicle, carry a Louis Vuitton bag, or talk on an iPhone, our consumption is conspicuous to others. In these settings, consumers may enlist brands to convey signals about themselves. Wearing a Rolex watch signals success and perhaps good taste — personal characteristics that are desirable to communicate, but objectionable to mention explicitly. See Y.J. Han, J.C. Nunes, and X. Drèze, “Signaling Status With Luxury Goods: The Role of Brand Prominence,” Journal of Marketing 74, no. 4 (July 2010): 15-30.

11. Rajiv Lal and Miklos Sarvary draw a distinction between what they term digital attributes, which can be searched online, and nondigital attributes, which cannot. See R. Lal and M. Sarvary, “When and How Is the Internet Likely to Decrease Price Competition,” Marketing Science 18, no. 4 (November 1999): 485-503.

12. F. Zettelmeyer, F.S. Morton, and J. Silva-Risso, “How the Internet Lowers Prices: Evidence From Matched Survey and Automobile Transaction Data,” Journal of Marketing Research 43, no. 2 (May 2006): 168-181.

13. C. Tice, “Hard Rock Cafe Hits Some Sour Notes But Keeps Rolling,” March 2, 2010,

14. John G. Lynch Jr. and Dan Ariely demonstrated this point in a clever study of wine markets. They found that making it easier to obtain information about quality reduces price sensitivity for differentiated wines. However, when the Internet revealed that the wines were undifferentiated, price sensitivity among customers increased. See J.G. Lynch Jr. and D. Ariely, “Wine Online: Search Costs Affect Competition on Price, Quality, and Distribution,” Marketing Science 19, no. 1 (February 2000): 83-103.

15. B. Ratchford, D. Talukdar, and M.S. Lee, “The Impact of the Internet on Consumers’ Use of Information Sources for Automobiles: A Re-Inquiry,” Journal of Consumer Research 34, no. 1 (June 2007): 111-119.

i. Anderson et al., “Harbingers of Failure.”

3 Comments On: Why Great New Products Fail

  • Sunday Sinyinza | March 15, 2016

    Profound and refreshing insights. Well stated!

  • Leon Zurawicki | March 15, 2016

    What seems to be missing in the article is the observation that individual consumers by nature differ in their curiosity factor and the amount of energy they put in the info search. Marketers have started looking into how different personalities affect the consumers’ information search process.

  • DAVID L MILLER | March 26, 2016

    Absolute classic by Prof Simester. “The risk for companies is that they invest in innovations that customers cannot recognize”. This is a major issue that plagues most firms. A further interpretation is that ideas are not being vetted with external audiences, either functional department end-users or paying customers.
    Brain games play a major factor in pricing and all sorts of consumer behavior.

Add a comment