Excerpts from Stop Flying Blind

Mapping the Future

Nobody can predict the future of a market or business perfectly. If I could do it, I’d be retired someplace living off my stock income. If the authors of all those management books could do it, they wouldn’t have to write books and give speeches for a living.

Part of the problem is that the world’s very complex, and any absolute prediction is bound to break down as unexpected things happen. But the biggest challenge is that the future doesn’t yet exist. It’s no a single deterministic thing, it’s a set of possibilities. We change the future every day with our own decisions. So what we need for the future isn’t a prediction. We need a map, showing all the possibilities and consequences of various decisions: if you go here you’ll end up in a valley, if you go there you’ll end up in the mountains — and if you go over there you’ll run off a cliff.

The better you draw the map of possibilities, the better your company can choose a good future for itself and its customers.

Mapping the future of a market or industry requires input from three different perspectives. You need to know first what’s going on with the customers. Not just what they’re doing today, but how they think, what they want out of life, and how they’d react to changes that might happen in the future. You need to know the insides of their heads so well that you can speak for them reliably.

Second, you need to know how technology is going to change, since that determines what your company can create. I’m using “technology” in a very broad sense here, meaning not just physical hardware like computer chips and paint formulas, but also processes companies can use to deliver services. The internet, for example, is a technology change that’s changing business processes in almost all companies, and creating a lot of new opportunities. The telephone did the same thing in the early 1900s.

And third, you need a good understanding of what your competitors will do. It’s not enough just to know their products and org chats, you need to understand how they think and operate, and what their basic personalities are, so you can anticipate what they’ll do in future situations.

Lots of other information can also be useful for predicting the future. For example, it’s very helpful to factor in future changes in government regulations (if you know what they’ll be). But I think customers, competition, and technology are the most important factors in mapping the future, and they need to be brought together very intimately because there’s so much synergy between them.

To understand how a future map is used, picture yourself as a Roman general leading your army to a winter camp. To the north there’s a sheltered valley that would be perfect for your needs. The fastest path to the valley leads across a river most people think is impassable, but your bridge-builders say they can do it, so you set them to work. You know the barbarians from the west are also searching for winter shelter in the same area. You don’t want them to reach the valley before you. Your scouts have identified a hill that dominates the road from the west. You send your archers to fortify the hill immediately, to cut off any advance.

The valley’s a potential market your customer research team found. The people building the bridge are your advanced technologists. The scouts who found the western road and the hill above it are your competitive analysts.

None of these people, working alone, could have drawn the map and told you where you could go on it. But once you had information from all three, you could see the likely future, plan out where you wanted to travel, and prevent the other guy from getting there first.

Unfortunately, this sort of map-making doesn’t happen naturally in the business world. In most of the companies I know, the people doing competitive analysis, market research, and advanced technology mix together like oil, sand, and water. Good market researchers are practical and methodical, deeply grounded in data and in the processes by which they gather it. They’re very uncomfortable with future speculation and unfounded predictions. Good competitive analysts are intuitive, they specialize making predictions based on small amounts of evidence. They hate being tied down by process. And true advanced technologists often have very fixed ideas about the world, ideas that are linked to the intellectual problems they want to research (ie, I want to work on speech recognition, therefore I believe that many important problems can be solved with speech recognition). They can be very impatient with anyone trying to impose customer or competitive realities on them.

On top of the basic differences in outlook, the people who gravitate to these teams come from different academic backgrounds, so they often have different vocabularies and different professional standards. The work also attracts different personality types, which often don’t mix well naturally.

Because of the differences, these teams often have pretty low opinions of one-another, sometimes bordering on contempt.

To make a good map of the future, you have to figure out how to mix data with intuition, to blend science and art. To do this, you first have to understand and appreciate each of the groups separately. Then you have to teach them to work together…

Competitive Analysis Done Right

The fall of competitive intelligence
Once upon a time, back in the 1990s, competitive intelligence was a hot area at many companies. They invested heavily in creating competitive intelligence teams. A professional group called the Society of Competitive Intelligence Professionals claimed that CI was the fastest-growing corporate discipline. SCIP had more than 3,000 members in 1996, and was growing by more than 100 new people a month.⁴ Competitive intelligence consulting firms did big business, and if you visit any good research library you’ll find whole shelves of books about Competitive Intelligence, most of them written in the 1990s.

But when the .com bubble burst and companies started cutting costs, many of those competitive intelligence groups were wiped out. “CI units are being eliminated… At least half of the CI functions in place today have suffered significant cutbacks, or will face them within the next six months.” That was from the SCIP’s own newsletter in 2003.⁵ One of the leading promoters of competitive intelligence in its heyday is now writing about how to apply the Talmud to business decisions.

Why the retreat? Conditions vary from company to company, but I think there are five main reasons:

1. The case for CI was grounded in fear. Much of the urgency behind creating the function was driven by fear of foreign companies, which were said to practice competitive intelligence aggressively. Japan in particular was described as a hotbed of competitive spying, and it was implied that this played a key role in Japan’s economic rise. The rhetoric was frightening, and it seemed likely that any company failing to create a competitive intelligence unit was doomed to fall to foreign conquerors.

But then the perceived Japanese “threat” to American business receded, and so did interest in Japanese business practices (when’s the last time you heard someone quote from the Book of Five Rings?).

2. The role was never completely defined. CI was a very new discipline, and there hadn’t been time for a consensus to develop on exactly what the role was and how to organize it. As a result, different books and consultants gave conflicting advice. In the lack of clear expectations, I think many competitive intelligence groups were never given well-defined charters. When a company’s under financial stress, a poorly defined function is an obvious thing to cut.

Inevitably, some of the advice was also damaging. For example, one prominent book said the CI role is like being a court jester for your company. The idea was that the CEO resembles Shakespeare’s King Lear, surrounded by liars and flatterers. The jester is the guy with bells on his hat who tells the king the truth, mocking the egotists and exposing the liars.

It’s true that someone in a competitive role must be unafraid to say exactly what the data indicates, even if it’ll upset people. But beyond that I’m uncomfortable with the jester analogy because it implies a completely negative role, and one that focuses only on influencing the CEO. In most of the companies I’ve known, to make change work you need to influence the whole management team, and beyond. You can’t do that if you speak only to the CEO, and besides you won’t win much respect from the organization if all you do is point out the flaws in other people’s work. When the CEO is replaced (which happens a lot more often than the death of a king), it’s likely that the one agreement among all the remaining executives will be that they want to strangle the jester.

Which, I believe, is what happened at the end of King Lear.

If you need an analogy for the competitive role, it’s better to think of the scout for a wagon train, forging ahead in the wilderness to identify dangers and find the easiest path for everyone. The scout’s not the manager of the wagon train, but he’s a leader with a unique and valued role. And he never wears bells.

3. The focus was on intelligence, not analysis. Much of the CI literature focused on how to gather and verify facts about the competition’s activities. It’s right there in the name — the function collects intelligence on what the other guys are up to. You can find entire books just listing various intelligence-gathering techniques, down to obscure things like taking the competition’s factory tour with two-sided tape on your shoes, so you can collect microscopic samples of the materials they’re using. The problem is that basic intelligence collection is becoming less important as the Internet grows and people change jobs more often. The Web is awash with competitive rumors, and chances are that if you can’t find the information you need online, one of your former coworkers is now working for the competitor and will sing like a canary if you buy them lunch. There’s simply less need for full-time employees who ferret out tidbits of intelligence.

Relative Value Chart

What companies do need is insight on what the flood of information means — how it adds up, and what it says about the competition’s thinking and future behavior. This is why I prefer the term “competitive analysis” rather than competitive intelligence. But that sort of predictive analysis is a very different discipline than collecting data, and you’ll do it best when competitive analysis is teamed with market research and advanced technology research. So competitive analysis isn’t really valuable as a standalone function.

4. The wrong people were hired for the function. This probably relates back to the lack of a clear charter for the CI role. When you’re not sure what a function will do, it’s easy to imagine that anyone can do it. Many of the people I’ve seen working in the field were marketing or sales people who had been dropped into the competitive role without much preparation, or much inclination for the work. They floundered around trying to figure out what to do, and produced very superficial reports.

Because of the flood of how-to books on competitive intelligence, I think some people formed the impression that anyone could do CI if they followed a few simple steps. That’s a little odd; I don’t know of any other field in business where the expectation is that anyone can be good at it. You don’t try to turn randomly-selected employees into engineers, or PR specialists, or salespeople. You look for people who have talent in that area. The same is true for competitive analysts. It’s a specialized field, and not everyone can do it well.

5. Competitive Intelligence is not mission-critical in the short term. A company has to have salespeople or no one eats. You have to have engineers or products just don’t get built. But if you don’t have a competitive team…well, the company keeps going just fine, thank you. For a while.

This means, in practice, that a competitive team has to be more than competent in order to survive. It has to be superb, delivering great value to the company in a visible way, so no one would think of living without it. In practice, that means just being a service group, delivering good information to clients in the company, is not enough. The group has to solve serious business problems and help close sales. Rather than being a source of competitive information, the group needs to be a source of competitive leadership.

Late one evening in 1989, I entered one of Apple Computer’s office buildings in Silicon Valley. Although Apple called its headquarters a “campus,” it was actually a series of buildings sandwiched between homes and stores over several square miles. The company had rented them haphazardly as it grew.

The building I went to was inconspicuous, two stories tall and tucked behind a screen of trees. It wasn’t the usual place for executive meetings, but an important meeting had been held there earlier in the day.

It was after sunset when I entered the building, and the place was very quiet. The building didn’t house a lot of engineers, so most of the employees had gone home. I went to a darkened conference room, where an IBM personal computer stood in one corner. It was a PS/2 Model 80, a hulking floor-based tower that was the leading edge of PCs at the time. After checking to make sure no one was nearby, I turned on the computer and watched it start up.

It launched a pre-release copy of Microsoft Windows version 3.0. I saw the software come up on the screen, played with it for a couple of minutes, and immediately knew Apple was in deep trouble.

To understand why, you had to know the history of Microsoft up to that time. This was back in the days when PC companies like Apple, Microsoft, Lotus, and Word Perfect viewed one another as peers. The dominant behemoth was IBM, and we were all dancing around them. Microsoft was the clever operating system company that had ridden the IBM standard to prominence, butno one really respected its ability to innovate in software. Its efforts in applications were a joke — Microsoft Word was something like the #6 word processor on the PC, and even Microsoft’s software for the Macintosh had numerous competitors, many of them viewed as technically superior to Microsoft’s products.

Microsoft Windows was the biggest joke of all. Its first two versions had been crude, extremely hard to use, and didn’t excite anyone. In some ways, they probably helped Apple by validating the idea of a graphical interface for a computer, without providing one that was good enough to steal away many customers.

Windows 3.0 changed that. It looked nice. The graphics were pleasant, the icons were reasonably well laid out on the screen, and it worked fairly well. There were still some rough edges, but it was good enough that I could picture a PC user installing it and not being embarrassed a week later. Windows was, for the first time, usable.

For reasons I still don’t know, Microsoft had decided to come down and give a demo of the unreleased software to Apple’s executives. I worked in the company’s competitive analysis department at the time, and as the only people in the company who had IBM PCs, we were asked to provide one for the meeting.

I wasn’t invited to the meeting, for obvious reasons, but I stayed late that night until I was sure it was over. As it turned out, when the Microsoft people left the meeting, they hadn’t erased the software from the PC. Now it was mine. I somehow carried the very heavy PS/2 tower out of the building and put it the trunk of my car. The next morning we started testing the software, trying to learn as quickly as we could just in case Microsoft came back and asked us to wipe the hard drive.

They never did.

With a pre-release version of Microsoft’s new product in hand, we were in a good position to prepare Apple for the upcoming competition. And in many ways we did — we documented our competitive advantages, educated the engineers about the improving competition, created marketing collateral, and generally tried to prepare the company for a fight. But the preparation turned out to be harder than I expected, in part because of resistance from above.

Spreading bad news about a competitor can be very disruptive to a company. It distracts employees, causes people to question their current plans, and generally hurts efficiency. The news is especially hard to deliver when a competitor has a history of screwing up, and most of the people in the company don’t use the competitor’s products. It’s seductively easy to rationalize that the competition is going to blow it one more time.

Sure enough, soon after we started raising a red flag about the software, my boss called me into his office. He said we were upsetting too many people, and told me to tone down the message. “After all,” he said, “it’s just another version of Windows.”

Maybe Apple was destined to lose anyway. Apple’s refusal to license its software to other companies meant it couldn’t establish a competing software standard, and its failure to produce new innovations that would make Windows obsolete meant it couldn’t hold onto many of the customers it had. But I think another cause of Apple’s fate was its inability to picture how the world would change. Apple didn’t really understand the minds of PC customers, and couldn’t see how Microsoft’s new software would act on them. And so despite a free preview from Microsoft, Apple never fully rose to the challenge of Windows 3.0, and Microsoft went on to cement its dominance of the PC industry.

By the traditional rules of competitive intelligence, I ought to feel at peace with my role in this. I did everything I could do legally to get advance information, our team turned out the best analysis we could, and we reported it as aggressively as we were allowed to. But I think that’s a cop-out. My company screwed up on a competitive issue. Therefore I’m partly to blame.

My experience with Windows taught me the most important rule of competitive analysis — your role is to make sure your company wins competitively. It’s not enough to deliver a great report and then wash your hands of the situation. If the company doesn’t act on your information, you failed.

You need to drive this principle into everything a competitive group does…

Are we there yet? Knowing when you’ve crossed the chasm

If your company’s launching a product or building a new market, one of the most important questions you’ll be asked is “where are we on the adoption curve?”

The adoption curve is a graph that shows how groups of people tend to adopt innovations. It was drawn by Everett Rogers in the book Diffusion of Innovation, and repopularized by Geoffrey Moore in his book Crossing the Chasm.

Adoption Curve Graph

Rogers based his work on cases like the adoption of improved seed among farmers, and the diffusion of sanitation practices among rural villagers. He found that an innovation moves through a predictable pattern of adoption in a group of people. First, it’s tried by a small percent of the population, which he called the Innovators, who are willing to try almost anything new just to be different from everyone else. It’s easy to get them to adopt an innovation, but because they like being different from others, the Innovators are not the greatest champions to get others to adopt it. In fact, sometimes other people will avoid things the Innovators like because they don’t want to be seen as weird.

(If you’re having trouble picturing this type of person, think back to high school and substitute the word “geek” for “Innovator.” Got it?)

If you want others to adopt an innovation, the Early Adopters are the ones to focus on. They’re a little more numerous than the Innovators, and they’re social leaders, the people that the rest of the population looks to for guidance. Unlike the Innovators, Early Adopters crave social status, and they get it by being the first to popularize new innovations. Think of that guy on your block who was the first to buy a flat-screen TV. Everyone else comes over to his house to look at the TV, ask questions, and so on. He gets social status, and the rest of the neighborhood gets to see what the TV’s like before they buy one.

Rogers’ work was very important because, for example, it told social workers that that rather than trying to teach new sanitation practices to everyone in the village at once, they should focus on the early adopter leaders in a village. Once they’re on board, everything else would follow more or less automatically. Rogers also cautioned that it’s not very useful to get the Innovators to adopt something new, because they’re viewed as weirdos by the rest of the village. If they adopt something, it’ll actually be less appealing to the rest of the population. (This explains the entire history of the game Dungeons and Dragons, by the way.)

This advice has been adopted aggressively by marketeers trying to sell new products. The idea’s simple — focus first on the early adopters, and use them to help sell an innovation to the rest of the population. In theory, marketing of an innovation should work like the flow of water through a multi-tier garden fountain: First you fill the small bowl at top, then water spills down to fill the second bowl, then water spills down to the third, etc.

What Geoffrey Moore added to the mix was the idea that in high tech, it’s pretty hard to get the water to flow from the early adopters to the mainstream. You have to budget extra time and investment, and make sure your product has very practical benefits, in order to get across the “chasm” to mainstream adoption.

That means for tech companies selling to consumers, knowing where you are on the adoption curve is obsessively important. If you’re still selling to early adopters, you need to keep investing patiently and refining the product. If you’re across the chasm, you can scale back your investment and focus more on harvesting profit and defending market share.

I’ve tried very hard to use this methodology to manage and track adoption of my companies’ products. Unfortunately, after spending a lot of time and research money on it, I’ve found that it’s much easier to use the adoption curve to explain things after the fact than to use it as a guide when you’re trying to make decisions. In practice there are several different adoption curves, and it can be very hard to determine which one you’re on. It’s disturbingly easy to convince yourself that you’re selling to mainstream users when in fact you aren’t, or that your market is close to saturated when in fact it’s about to take off.

Misread the adoption curve and you can easily kill off a product long before it’s destined to decline, or pour money into something that’s doomed to fail. I’ve seen three main sources of problems: Determining the shape of the adoption curve, determining who’s really an early adopter, and determining which of several adoption curves you’re on.

What shape is the adoption curve?
The first and most important question you need to ask yourself is whether you’re selling a consumer or enterprise product. I’m using the words “consumer” and “enterprise” a little differently from the way most people use them, so here’s a quick definition:

By “consumer,” I mean any product that’s selected by the person who’ll use it. That includes just about anything in your house (unless it was a gift, a situation I’ll discuss below). By my definition, consumer products also include a lot of business-related items — your suit, your cellphone, your briefcase. If your company lets you choose which car you rent on a business trip, or which airline you fly on, I’d call those consumer purchases as well.

An “enterprise” product isn’t selected by the user. It’s selected by an employer and assigned to a user, who has little or no say in the product choice. If there’s a computer on your desk at work, chances are it was an enterprise purchase. So was the phone, and the desk itself for that matter. Most other things at work are enterprise purchases, including the corporate servers in a closet somewhere, the building your work in, and the corporate health care plan.

The main difference between these two worlds is who’s doing the buying. In consumer products, the person who’ll use the product picks it out. In enterprise products, the product is chosen by a corporate buyer, on behalf of the whole organization. For high tech products, that buyer is a member of the Information Technology (IT) staff.

Because of this difference in buyers, the adoption curves for consumer and enterprise products are vastly different, which means you need to design and sell the products completely differently. I can’t over-emphasize how different these worlds are. A corporation’s culture, language, brand image, and business practices all have to be tuned to one or the other. They’re so different that it’s very rare to find a high tech company competent to sell both consumer and enterprise products. For example, IBM and Sun are both enterprise companies. The old IBM PC was an exception, but as we saw IBM couldn’t sustain the business. Apple is consumer. It tried for more than a decade to make itself into a corporate supplier, before Steve Jobs came back and told them to face reality (and focus on their strengths). Nokia too is a consumer company, although it’s trying mightily to build up an enterprise business.

The only high tech companies I can think of that manage to be both consumer and enterprise have set up separate divisions to do it. HP is an example, as is Microsoft. But they’re exceptions rather than the rule.

Consumer vs Enterprise

If you’re selling a consumer product, even a high tech one, I’ve found that Rogers’ classic adoption curve applies pretty well. Early adopters try out the products, share their findings with others, and if a product is good, adoption will move fairly smoothly to mainstream users. The challenge in consumer products seems to be more about understanding exactly why your product’s being purchased, and therefore which market segments you’re really operating in. More on that below.

For enterprise products, the smooth continuous adoption curve simply doesn’t exist. This is primarily what Crossing the Chasm was about, and if anything I think it understates the differences between consumer and enterprise sales.

First, let’s look at the motivations of the buyer. Someone buying a consumer product is usually motivated for the benefits it brings to himself — usually a mix of increased productivity (“If I bought that circular saw, I could finish paneling the den a lot faster”), increased status (“I can’t wait to se the look on Dad’s face when he sees the new den”), and increased pleasure (“Man, what a cool saw.”)

A corporate buyer cares about none of those things. In fact, they are trained to ignore them. Employees in the company are constantly lobbying for nicer corporate goodies — plusher cubicles, better computers, nicer office chairs. The buyer’s job is to defend the company against all those demands, and instead do something that’s sensible. That doesn’t mean user desires are completely ignored, but it’s more important to stay within budget, and to avoid new support costs.

Support costs are a big issue for high tech products — if a product is cheap to buy, but is so complex or flawed that it generates a lot of user questions, the cost of answering those questions can easily be greater than what you saved on the product. The tech consulting firm Gartner Group has for years argued that for PCs, the cost of training and support is higher than the total purchase cost of the hardware.

But the most important priority for a corporate buyer, exceeding all other requirements, is avoiding a mistake that could get you, the buyer, fired.

This means that corporate buyers as a group are extremely conservative, much more so than the average consumer. They’ll tend to buy only from companies they know, and they’re very leery of new products. Even a proven increase in productivity may not be enough to motivate them to buy. After all, most IT departments are not rewarded for increasing the productivity of the company. They’re service groups, rewarded for controlling costs and not allowing any major technical disasters.

When I was at Palm, we surveyed corporations to determine their attitudes toward purchasing new technology, especially handhelds and smartphones. Instead of the usual bell-shaped curve, we found a two-humped curve:

How corporate technology buyers describe their companies’ adoption of new technology

About 30% of the companies surveyed said they were aggressive technology buyers — they viewed the use of advanced technology as part of their competitive advantage. As a deliberate corporate policy, they tried to stay on the leading edge of new products and services. These companies tended to decentralize technology purchasing, allowing individual departments to buy rather than forcing them to work through a central purchasing group.

The other 70% of companies were more conservative about technology. They didn’t try to be at the leading edge, and in fact were fairly reluctant to integrate new technology. They were more likely to have centralized control over technology purchases. When we looked specifically at corporate-funded purchases of handhelds and other mobile devices, we found that the early adopter companies were moving ahead aggressively to broad deployments across the company, while the other 70% of corporations were still doing small trial deployments, if anything at all. The early adopter companies accounted for 68% of all corporate handheld purchases, even though they were only 30% of companies.⁶

(Incidentally, this split adoption pattern explains why you sometimes see schizophrenic press coverage of the market for mobile devices. One article will say the market is taking off fast, another will say it’s dead in the water. Both are true; it just depends on which companies you’re looking at.)

We also found some interesting differences in technology adoption by industry. Some industries were biased toward early tech adoption, while others tended to be more conservative. For example, engineering and research firms, communication, and wholesale all tended to be aggressive tech adopters. Meanwhile, companies in education, government, and health care were much more likely to be laggards. (In health care, remember that we were looking at adoption of new computing technology, not new medical technology.)

Corporate handheld purchases tended to mirror these adoption preferences, with one exception. Health care has a large installed base of handhelds even though it’s slow to adopt new information technology. Apparently doctors find handhelds so useful for tracking drug and other information that they’re forcing the devices into medical corporations whether IT wants them or not. This is an interesting case of a consumer (userled) adoption cycle driving corporate adoption. Something similar happened when the first PCs entered corporations in the late 1970s.

In most companies, we found that the actual purchasing decision was controlled by two parties — the departmental manager (who’s paying for the product) and the corporate IT manager (who enforces corporate standards, supports products, and often negotiates the actual purchase). The relative strength of these two groups varies from company to company, but in most cases both of them have some level of veto power.

Because of all this, you can’t sell an enterprise technology product the way you do a consumer one. In an enterprise sale, you usually need to generate two different types of demand in two different places. First, you need demand from the managers of the department that will use your product. Usually they need to be convinced that they’ll improve productivity or make their employees happier. Then you also need IT management to at least acquiesce to the purchase. They don’t have to love it, but they must at least be willing to put up with the product, or they’ll find a way to stop the sale.

I have vivid memories of watching a focus group with IT managers while I was at Apple. One of them said adamantly, “I have devoted my career to keeping Macintosh computers out of my company.” No matter how many cool and creative ads Apple came up with, no matter how much the employees wanted Apple’s computers, they were not going to be bought by that company.

As Geoffrey Moore has pointed out, this means companies need to use a two-step selling process for enterprise technology products. It’s relatively straightforward to sell to the corporate early adopters, because they make it their business to try everything that’s new. But to sell to the other 70% of corporations, you have to look closely at the adopter profile of the particular vertical you’re selling to. You have to prove not just that your product is useful in that industry, but you also have to demonstrate that it’s not going to cause trouble for the IT staff (a much higher hurdle, and one that takes time).

This means the deck is stacked against small startups selling new technology products to corporations. If a similar or even somewhat inferior product is offered by a large company that IT already knows, the IT buyers will try to steer purchases toward their trusted supplier.

For years IBM used this effect to lock out competitors, by pre-announcing copies of a competitor’s innovation, so customers would hold off buying from the upstart. Microsoft uses similar tactics today.

Some markets mix consumer and enterprise. Not all markets can be cleanly sorted into either the corporate or consumer bucket. A very good example is mobile phones. Most mobile phones are bought by the people who will use them, and so I’d classify that as a consumer market. But if you’re a mobile phone manufacturer, you don’t sell directly to most of those users. Instead, you sell phones to a mobile phone carrier like Verizon or Tmobile, which in turn sells the phones to users. From my perspective, that starts to look more like an enterprise market.

If you’re running a mobile phone company, do you focus on phone users, or on the buyers at mobile phone companies?

The answer is that you agonize about it a lot, and different companies come to very different conclusions. Some companies market focus mostly on users, counting on demand from them to force the carrier to offer their phones. Nokia is notorious for doing this in Europe. Other examples are Motorola’s Razr slimline phone and Palm’s Treo smartphone. But some other major mobile phone companies don’t worry as much about appealing to users, instead trying to produce whatever the carrier wants, and counting on the carrier to push that to users. The Korean company LG, one of the fastest-growing mobile phone companies in the world, uses this approach.

What about gifts? These are a special case because they’re not bought by the user or a corporate buyer, but instead by someone trying to express love or make an impression. Gifts don’t follow the normal adoption curve. Instead, they’re driven by fashion and herd thinking. The fashion cycle moves very quickly — a hot gift one year is likely to be a doorstop a couple of years from now. For example, PDAs were a raging gift item in the late 1990s; gift giving accounted for about a quarter of Palm’s sales. A few years later, the rage had moved on to digital cameras.

Gift sales grow explosively and decline just as fast. If you’re selling a consumer product, it’s important to track what percent of your sales are gifts. Restrain your expectations (and the expectations of your investors) if you see a lot of gift-giving. Think of the revenue as a onetime upside event rather than sustained demand. The hot gift period won’t last.

Are your early adopters actually mainstream buyers?
In a stable rural town or village, it’s pretty clear who the high-status people are, and if you’re a leader in one aspect of village life, you’ll tend to be a leader in most aspects. But in the modern world, most of us belong to a series of different villages — a neighborhood, a job, interest groups around town or on the Internet. You might also be a member of a social club or school alumni association. It’s pretty common to be an early adopter in one area but a laggard in another (for example I’m a geeked-out innovator when it comes to computers, but a late adopter in cars).⁷

This makes it much harder for a marketer to proactively identify who the early adopters are for a new product. There are some people who just have that early adopter personality type and tend to lead in everything they do, so you can try to focus on them. In high tech, we try to find technology early adopters, assuming that someone who’s quick to buy a flat screen TV will also be quick to buy a new mobile phone. But in that case it’s easy to end up accidentally marketing only to the technophile innovators, and they’re a dead end in terms of driving sales to others.

When you’re not sure who the early adopters are, it’s also hard to track your position on the adoption curve. For example, say you have sold a product to 10% of the population. That might mean that you’re just now finishing with the early adopters in a market that will eventually include 90% of the population (curve B in the chart below). Your sales are about to explode. Or it might mean that you’ve just saturated a market that’s destined to top out at 12% of the population (curve A). Your sales are about to plummet.

Early adopter curve

It is surprisingly hard to tell the difference between these two cases when you’re in the middle of them. Generally if a product has reached 10% of the population, a lot of other people will be thinking about buying it, just because it generates some buzz. It’s very difficult to sort out the difference between this buzz and actual demand until after the fact, when you look back at what happened to real sales.

The better you know your market(s), the better your chance of avoiding this problem. This is why you have to be very serious in asking the question, what problem are you solving for your customers? Why are they buying from you? Often a company will have a lot of noble and interesting reasons why customers are supposed to buy its products, but when you look intensely at the customers, that’s not why they’re buying. Once you know the real reason why people are buying, the next question is, how many people have this problem? Or more to the point, how many people have this problem and care about it so deeply that they’re willing to spend money to solve it?

Even if you determine that you really are solving world hunger or some other universal problem, you then need to use market research to map out your demand funnel — how many people are aware of your products, what percent of those people think about buying, and what percent actually do buy. Where are you losing the most people? These two products both sell to 15% of the market, but they face very different challenges:

Product A vs Product B Chart

Most people have heard of product A, but they generally think it’s not for them. Doing a lot more marketing for this product is not likely to produce a rise in sales, unless the product has some incredibly compelling secret feature that no one currently knows about.

Even then, you’ll be fighting against the current impressions people have about your product, which can be very hard to change.

More marketing might help Product B; the biggest barrier to its sales is that people just haven’t heard of it. When a company sees a chart like Product B’s, it’s pretty common to assume that if you increase awareness, consideration and purchase will also rise as well. That might happen, but you can’t take it for granted. Probably the people who have already heard of your product are the most enthusiastic customers for it, and you can’t count on everyone else reacting to it the same way.

The overall point here is that additional marketing won’t necessarily create more demand, until you know why people aren’t buying today.

Which demand curve are you on?
Many high tech products are flexible, and can be used for multiple purposes. The Internet’s a great example — its first mass-market use on PCs was for transferring e-mail. Then it was used for browsing information. As the new medium grew, companies created additional uses for it — online auctions, sharing music, e-commerce, and so on. If you’re a company selling Internet access or Internet hardware it’s almost impossible to plot Internet demand on a single curve. Instead, your demand is a composite of a lot of different curves — one for e-mail, one for music, one for auctions, etc. Some of those curves are probably close to saturation (e-mail, at least in the US). Others are still in the early adopter stage.

In the case of PCs, demand has gone through several curves as different uses for the PC emerged. PCs started mostly as appliances for word processing and spreadsheets. Desktop publishing came along in the late 1980s and created a new surge in demand. Games and the Intenet drove much higher penetration of PCs into homes in the 1990s. Other applications have created their own smaller surges in demand along the way.

Looking back at the development of the market, it’s easy to see how these overlapping demand curves worked, but at the time it was very hard to predict them. For example, in the mid-1980s it was very easy to predict that PC sales would soon plateau, as usage of spreadsheets and word processing saturated. In reality, a new wave of growth was about to start.

Today we face some of the same questions. Although PCs have high penetration in the US, they are not as ubiquitous in Europe, and are downright rare in developing countries like China and India. As those countries’ economies grow, will PC ownership approach US levels? Or will other products, like advanced mobile phones, take over some functions of the PC in other countries?

No one knows.

Although multiple usages are commonplace in high tech products, they’re not unheard of elsewhere. For example, as low-carb diets took off, beef became a diet food.

Overlapping adoption curves of a hypothetical product with multiple usages. The first usage gets the product launched, and it also becomes a popular gift item in the fourth year. The second, more popular, usage doesn’t get started until year seven.

Sales based on adoption curves
Here’s what those three adoption curves could do to the sales of our hypothetical product. At any point on the curve, it’s extremely difficult to predict what will happen next. In this case, I have imagined a happy outcome for the company and its investors — they are patient enough to get to the second wave of growth. In reality, most companies today make savage resource cuts at the point I’ve labeled “management team fired,” and wouldn’t have enough money to fund the second wave of growth. You must understand your market segments, and where you stand in them, or this sort of demand hiccup is likely to cripple your company.

This chart shows the real penetration of various products into American households over time (I assume the line for “airplane” indicates percent of families who have ridden on airplanes, not the percent of homes that have been penetrated by them). To me, the most striking thing about the chart is how many glitches and reversals there are in the curves. All the lines eventually go up and to the right, if you’re willing to wait 100 years. But if you were actually living at a particular point on one of those curves, you couldn’t reliably predict the size and timing of future growth. For example, imagine yourself as a telephone executive, 60 years after the invention of the phone. Telephone penetration has been dropping for the last five years, and is now back down to where it was twenty years ago. Would you predict that it was about to start going up again? What would shareholders do to an executive making a prediction like that today?

[Chart: Federal Reserve Bank of Dallas Annual Report 1996. http://www.dallasfed.org/fed/annual/1999p/ar96.pdf]

In the handheld market, we had a terrible time figuring out where we were on the demand curve. When we researched our customers, we found a situation that diffusion theory says shouldn’t exist. Our penetration into technology early adopters was okay but not very high. A lot of early adopters were still thinking about buying our products. This should have meant we were still in the early stages of growth.

On the other hand, a lot of people in our installed base looked like mainstream and late adopters. You’re supposed to get those people only after you saturate the early adopters. So what were they doing buying our products while most of the early adopters hadn’t bought yet?

Looking back, I think two things were going on. The first is that there were two demand curves for handhelds. One demand curve was for the use of a handheld to track your calendar and address book. In that market, handhelds were well past the early adopter stage, and in fact were approaching saturation. There are only so many people in the world who are so obsessive about their calendars that they want an electronic tool to track them, and most of them have already bought, at least in the US. That’s why there were so many technology late adopters in the installed base.

But the second use of handhelds is as a more generalized information management device. Add software, and a handheld can become a medical database for doctors, or a flight computer for a pilot. There are software programs for almost any vertical market or hobby. The growth in awareness of these programs is much slower than the growth in calendar and address book, because the add-in software is hard to find and isn’t bundled with devices. So that demand curve is still in the early adopter stage. I think that’s why there was still a lot of consideration among early adopters.

The second factor complicating the demand curve was that during the bubble years, handhelds became a popular gift item. They were the “in” thing to give Dad for father’s day or Christmas, even if he didn’t really need or want one. These gift sales made the market look bigger than it really was.

Looking back, after the fact, it’s possible to tease apart all of these threads and see how they fit together. At the time, it was almost impossible to see. The analysts were exuberantly predicting sales would rise from 10 million units a year to more than 60, and it was very hard not to get caught up in that — especially when some of the research seemed to support it.

Unfortunately, in the real world handheld demand plateaued at about 20 million units a year.

Lessons

What can you learn from all of this? I think the message isn’t to abandon the demand curve, but you have to be very careful in how you use it.

Know if you’re selling an enterprise or consumer product. Even if your product is designed for use in businesses, if he buyer is the user, you’re going to see a consumer buying pattern, and you need to set up your sales, marketing, and product design for that. On the other hand, if the buyer is an IT manager or other corporate official, you need to organize to sell to both the departmental manager and the corporate buyer. That takes longer than winning over a consumer, because you’re asking the corporate buyer to put his or her job on the line when buying from you. On the other hand, once you win over corporate buyers, their conservatism can make them very loyal customers.

Make sure your investors and management team understand the dynamics of the markets you’re after, very early in the game. The better you set their expectations, the better the chance that you’ll be given the time and money you need to develop the markets. Few things are more unpleasant, and more likely to destroy your credibility, than spending two years on product development, and then going back to the board of directors to ask for a huge marketing fund because you’ve switched target markets.

Know your segments. The better you understand the actual usages of your product, the better you can apply the demand curve. Understand who’s buying your products, and what their motivations are. If there are multiple motivations, investigate whether those point to distinct segments, each of which will have its own demand curve.

It’s easy to get tremendously confused when you lump multiple segments together. For example, the research that has been done on advanced phones implies strongly that there will be at least three different segments for such devices — one for communicators, one for entertainment phones, and one for information phones. The market for communicators could easily saturate before the market for entertainment phones even takes off. I think it’s very likely that these out of phase adoption curves will lead to big swings in enthusiasm for the advanced phone market in the next few years, with some people predicting it’s about to explode and others predicting it’s about to die. The reality is, there isn’t a single market to forecast.

Avoid the self-fulfilling prophecy. You need to accept that you can’t completely plot the demand curve until after the market has saturated. Your position on the curve depends a lot on how attractive your products are and how well you market them. Do a better job, and the market will grow. You can change the curve.

Many of the biggest mistakes I’ve seen companies make were misjudging where they were on the adoption curve.

For example, in the early 1990s many analysts concluded that the Apple Macintosh was a dead end, with a saturated market destined to die. They encouraged Apple to pour hundreds of millions of dollars into other new businesses — servers, new devices, applications, new operating systems. Most of them didn’t pay off. Meanwhile, investment in the Mac was constrained. When Steve Jobs returned to the company, he put the focus back on the Macintosh, and revived demand for it. You can still make a good argument that the Mac will eventually be a dead end, but it’s impossible to say if “eventually” will be in five years or twenty. A lot depends on what Apple does. It’s up to them. There were other problems at Apple, of course; I’m oversimplifying the situation. But one of the company’s biggest burdens was a belief that the adoption curve was a prophecy rather than a problem to be fixed.

About Stop Flying Blind

Everyone agrees that companies should focus on competing in the future rather than just reacting to what’s happening today. But how do you actually do that? How do you determine what a market’s going to be like when the market doesn’t yet exist? How do you predict what your competition’s likely to do before they even know it themselves? How do you spot the turning points that your company can use to change the rules of its industry, before anyone else sees them?

Most companies fly blind on these issues, but they don’t have to. By combining a variety of different perspectives — competitive analysis, market research, and advanced technology research — a company can map the possible futures, pick out the one most favorable to it, and help bring that future into being. Stop Flying Blind tells how to do that. For more information, visit www.mikemace.com.

Footnotes:
4 For a complete discussion, see the book Competitive Intelligence by Larry Kahaner. [back]

5 Written by Bill Fiora, principal of Outward Insights, a CI consulting firm. http://www.scip.org/news/v1i22article1.asp [back]

6 If you’re interested in methodology, this was a survey of 440 randomly-selected technology buyers/approvers in US companies having 100 or more employees. The survey was conducted in 2003.[back]

7 Consultant Peter de Jager does a nice job of explaining this in an essay on his website: http://www.technobility.com/docs/article032.htm [back;]

Download a PDF of this whitepaper.

Categories: Perspective