“Markets that don’t exist can’t be analyzedâ€¦.The only thing we may know for sure when we read experts’ forecasts about how large emerging markets will become is that they are wrong.” –Clayton Christensen, The Innovator’s Dilemma
One of the most difficult tasks in market research is guiding product development. The tech industry’s bipolar view of the future dominates its handling of new product research. The visionary companies tend to reject all product-related market research. They rely on their own ideas and instincts.
On the other hand, reactive companies try to make all their product decisions through research. When you do this it’s very easy to use the wrong sort of research. As I’ve mentioned before, if you use a focus group to make product decisions, you might as well flip a coin, because there’s no way to know if the group represents customers as a whole.
It’s better to use quantitative research â€“ at least then you’ll know you have a representative sample of customers. But there are still two major drawbacks to this sort of research, which I call the possibility gap and the blender.
The possibility gap. The visionaries are right on this point, customers usually don’t know what they want until they see it. If you ask an existing user for product ideas, they’ll take what’s wrong with the current product and dress that up as ideas for the future. For example, for years I looked at research on PC users, and they always asked for computers that are cheaper, have more memory, and run faster. Why? Because those are the barriers the users run up against most often.
In 1995, almost no customers in PC research studies were asking for high-speed network connections and photo-realistic 3D graphics, yet those turned out to be probably the most important new PC features in the following decade. To catch those opportunities, you would have needed a much deeper understanding of user psychology and of technology trends.
Hold that thought.
The blender. In most product feature studies, people are given a long list of features to evaluate, and the features that get the highest average score are the ones selected for the product. The problem with this is that it turns the customers into a single average, as if they had been dropped into a blender. The only features that score highly will be the lowest common denominator ones that affect everyone — things like weight, size, and ease of use. If you have a feature that’s beloved by some customers but hated by others, the two groups will cancel one-another out.
There’s a good example of this in the mobile phone world. If you survey mobile phone users about feature desires, the issues that rise to the top are smaller size, lower cost, and longer battery life. Those are the things that irritate almost all phone users. More advanced features, like built-in e-mail, end up close to the bottom of the list.
Despite this, two of the hottest advanced phones in the US today are Research in Motion’s BlackBerry and Palm’s Treo, both of which combine phones and e-mail features. They’re not attractive at all to most mobile phone users, but are beloved by the10% of mobile phone users who are so obsessive about communication that they want their e-mail with them all the time.
Very often, at least in technology products, the biggest opportunities are products that some people love but others hate. So what you want to look for in feature research isn’t the blended average, it’s the lumps that are in the mix before you turn the blender on. What feature requests cluster together? Do the people asking for those features have personalities or demographics in common? What problem do they share that drives them toward wanting those features?
The right way to guide products with research
I think the way to get past the blender and the possibility gap is not to try to design the actual products through research. Instead, focus on understanding the needs and psychology of the customers, so you can anticipate the way they’d react to new features. How do they live their lives? What do they care about? What are they trying to accomplish? What challenges do they face that you might be able to help with? Once your product engineers absorb these issues, they’ll start to more or less automatically design the right products.
Let’s take mobile phones again as an example. When you look in depth at the motivations of mobile phone buyers in the US and Europe, it pretty quickly becomes clear that a majority of them — about two thirds, actually — care only about basic voice and maybe text messaging. In the US, they buy the cheapest service plan they can, and take the free phone that comes along with it. In Europe, they’re usually on a very low-cost pay as you go plan (where they add money to the phone account as they go, rather than paying a flat monthly fee), and they often turn the phone off in order to hold down their bills.
These people have no interest in any advanced features or services. If you’re doing a study on advanced phones and you keep them in the research mix, their sheer numbers will make you conclude that there’s little hope for any sort of advanced phone. And, in fact, that just what some mobile companies have concluded.
But if you exclude those basic users from your study, you find that about one-third of mobile phone buyers actually are interested in advanced features of various sorts. One-third may sound like a small number, but keep in mind that about 700 million mobile phones were sold worldwide in 2005. A third of that is about 230 million phones a year, enough to attract almost any company’s attention.
The problem with these advanced users is that they don’t all want the same thing. If you apply the blender principle and mix them together as a group, you’ll find that on average they are moderately interested in almost every feature imaginable. This has led a lot of companies to create “smart phones” that are basically kitchen-sink bundles of features lumped together. These products usually don’t sell very well, because in the process of trying to be everything to everyone they become too big, too expensive, and too complex for anyone to love.
The products remind me of politicians trying to assemble the largest possible coalition of voters by not offending anyone. That sometimes works in politics because the voters have only a handful of parties to choose from, and a politician has to assemble a majority vote. But a product has an unlimited number of competitors, and 10% share might be a huge win. Better to please some people intensely and piss off everyone else than to get a lukewarm reaction from everyone.
Instead of trying to attack the engaged users all at once, you need to look for segments within them. Are there groups of people who want certain features in particular?
When you do that with advanced phone buyers, three groups emerge. One group gives high ratings to all communication-related features — e-mail, instant messaging, built-in fax, etc. Basically, they’re communication junkies, and they’ll pay extra for a communication-enhanced phone. These are the people buying RIM Blackberries and Palm Treos today.
The second group gives high ratings to information-related features — large memory, document display, databases, etc. These are people in information-intense jobs who need a mobile memory supplement. Think of a doctor looking up drug dosage information on the go, or a lawyer trying to find a case reference in court.
The third group responds best to entertainment-related features: music, video, games, and other ways to have fun. These entertainment-focused users tend to be younger than the others, and don’t want to give up their electronic lifestyle even as they enter the job market.
Segmenting the market isn’t a new idea; the auto industry has been doing it for more than 70 years (think sports utility vehicles and sports cars). But although the idea of segmentation is straight from Marketing 101, and is heavily used in established industries, it’s very hard to do in a new industry or product category. Market segments are only obvious after they have been proved by a successful product. Until someone builds that first e-mail phone or SUV, the natural human tendency is to either dismiss the existence of the market, or to lump the customers together and try to hit a home run with all of them at once.
You need to resist that temptation. Products designed to please all segments almost always fail, and if you wait until someone else validates the market, you’ll be fighting with 20 other companies to dislodge a competitor rather than running ahead of the pack.
Next time I’ll talk about how to segment the market for a new product.