NTT DoCoMo: Marketing i-mode

To ease the congestion problem, in 1997, DoCoMo gave Enoki—a lifetime employee who had worked his way up through NTT’s engineering ranks—a mandate to convince DoCoMo subscribers to begin using their cell phones in a fundamentally different way. More specifically, the mandate was to build a wireless Internet service that would create demand for sending/receiving text-based data via cell phones. Enoki decided to use DoCoMo’s existing packet switched network for the new service, which minimized the overall investment associated with the project.3 One advantage of such a network was that instead of having to keep a wireless circuit open for the duration of a phone call, it was possible for DoCoMo to interweave text-based data packets into a voice stream—allowing multiple subscribers to “share” a single circuit.
In building a team to help him create the mobile Internet service, Enoki began by recruiting two outsiders who he believed would help him break free of the rigid, technical mindset that permeated
NTT’s engineering-driven culture. The first was Mari Matsunaga, a marketing guru who had achieved a reputation as an editor-in-chief extraordinaire at Recruit, a Japanese magazine publishing house. Given the fact that Matsunaga had zero technical expertise (she didn’t even own a cell phone at the time), Enoki’s decision to bring her on board took many in the company by surprise. Enoki explained: In traditional Japanese companies, you’re supposed to hire from within. But because my job was to build a wireless data service, I felt the focus had to be on the content itself, rather than on the technology per se. NTT had many wonderful engineers, but no one who knew about content. That’s why I wanted Matsunaga; she’s a genius when it comes to content.
Enoki’s second outside hire was Takeshi Natsuno, a Wharton MBA who had returned to Japan in
1995 to join Hypernet, an Internet advertising company. A high-octane entrepreneur, Natsuno had spent the past couple of years refining his ideas about how to create an economically viable business model around a network-based system. Enoki jokingly recalled:
The first time I met Natsuno, he was bursting with insights about how to make a wireless
Internet business work. I immediately wanted him on my team. The problem was, he wasn’t available, he was still working at Hypernet. So every time I ran into Mari-san, I would ask her,
“Has Hypernet gone bankrupt yet?” because I was so keen to hire him.
Eventually, the trio formed a charismatic team: There was Matsunaga, the chic, forty-something, former editor with the infectious laugh and the finely-tuned marketing intuition; there was Natsuno, the Internet whiz kid with the over-the-top personality and the tight focus on economics; and there was Enoki, the industry veteran and the team’s paternal leader. Together, the trio shook up
DoCoMo’s corporate culture, not only by creating an environment where blue jeans were acceptable and brainstorming sessions could be held over beer, but more importantly, by making a series of controversial business decisions that totally belied conventional industry wisdom at the time.

Target Market Selection
The team’s first decision had to do with target market selection. Almost all the higher-ups at
DoCoMo had assumed that “i-mode” (the name given to DoCoMo’s new mobile Internet service) would be targeted toward business people, not only because of their relative price insensitivity, but also because it was generally assumed that business people would be the most likely to value a

3 To ease the congestion problem, NTT had earlier built the packet-based network. Enoki did have to make some investments

in upgrading the functions of the network to make it compatible with the i-mode service.

502-031

premium, “always-on” Internet service. As evidence, most overseas operators were already targeting
(or planning to target) business people with their respective mobile data services.
Contrary to these expectations, the i-mode team decided to focus on an entirely different demographic: trendy young people in their teens and twenties. What made this decision particularly counterintuitive was the fact that, because of low home PC penetration (about 12%), less than 10% of
Japanese consumers had Internet access in 1997, and the few who did tended to be businessmen.4
Nonetheless, the team felt strongly that the youth market was the way to go. Matsunaga recalled:
At first, a lot of people didn’t understand our decision. But I knew that even if they weren’t familiar with the Internet, young people had very strong communication needs; they loved to be in touch with their friends. Plus, they were attracted to new things, they liked innovation. I felt that if we could capture this group of people, everyone else would follow their lead.

Third Party Content
The team’s second major decision had to do with content. Whereas most wireless telecom carriers in Europe and the United States were focused on developing or purchasing proprietary content, the imode team decided to create a portal that would direct subscribers to content provided by thirdparties. This portal—which would be accessible via a mini-browser preloaded on i-mode handsets— would be structured as a hierarchical menu system; the menu would serve as a starting point for greater Internet access. In other words, i-mode would control the platform, but it would rely almost entirely on others for the content itself. Natsuno, who was the architect of this decision, explained his logic: The mindset at most telecom operators is to control the entire value chain, from the content to the network. But we decided early on to reject the “telecom way of thinking” and adopt an
“Internet way of thinking” instead. We knew that we were going to make money simply by increasing data traffic. The challenge was to figure out how to offer third-parties some kind of
“win-win” incentive so that they would want to participate as well.
Based on his experience at Hypernet, Natsuno knew that it was a struggle for web-based content providers to create robust revenue streams. One reason for this was it was difficult for content providers to impose even small fees for access to their websites; Internet users hated being billed for tiny amounts (particularly in Japan, where people rarely used credit cards for small purchases). Most content providers also found the cost of creating and maintaining a billing system to be prohibitive.
DoCoMo, on the other hand, had a billing system that was set up to handle large numbers of small transactions (i.e., phone calls); it could thus handle “micropayments” at a minimal cost. Given this, Natsuno had an idea: Why not offer to handle the billing for content partners who wished to charge subscription fees for access to their sites? The more Natsuno thought about it, the more sense it made. A centralized billing service would make the payment process seamless for subscribers, since they would be able to receive a single bill that combined their DoCoMo wireless fees with the subscription fees imposed by third-party sites. More importantly, Natsuno realized that this kind of billing service—for which DoCoMo would eventually charge content providers a 9% commission— had the potential to create exactly the kind of “win-win” scenario he was hoping for.

4 Irene Kunii, “Amazing DoCoMo,” BusinessWeek (International Edition), January 17, 2000.

NTT DoCoMo: Marketing i-mode

Technology Standards
The next decision the i-mode team faced had to do with technology standards, most importantly, which language protocol the new service would support. Enoki and Natsuno began by evaluating the Wireless Application Protocol (WAP), which was being supported by an industry consortium led by Nokia, Ericsson, and Motorola. WAP sent data faster and more efficiently over wireless networks than HTML (the protocol used on the web), and many European and U.S. mobile operators were already adopting the WAP standard.
The problem with WAP was that it required Internet content providers to produce two versions of their content—an HTML version for PC browsers, and a WAP version (using HDML/WML) for mobile phone browsers; the pair worried that this would deter third-parties from offering i-mode content. A second problem was that although many players were pushing to make WAP the global standard, the specifications for WAP were not fully in place. As a result, many incompatible variants of WAP existed.
Ultimately, the two decided to forego WAP and adopt a compact form of HTML instead.
Compact HTML (cHTML) optimized existing webpages for mobile viewing by eliminating certain graphics, tables, and frames. In the process, webpages were stripped down to a maximum size of 5k, allowing for quicker downloads. Natsuno believed that the decision to adopt cHTML would make it easy for content providers to modify existing websites for i-mode compatibility; however, as he recalled, DoCoMo engineers were appalled that we had chosen what they believed was an inferior technology. In the “telecom mindset,” the assumption is always that the most advanced technology is the best technology. In contrast, I believed that the important thing was to be compatible with the existing network.
Using the same logic, the i-mode team decided to adopt widely-used formats for graphics (GIF) and sound files (MIDI). In the case of the latter, the fact that Japan’s hugely-successful karaoke industry used MIDI was a key factor in the decision.
The next task was to select an email standard. There were two primary email methods to choose from. The first method, which was familiar to most PC users, relied on a central server to collect messages that users could retrieve at any time. This method allowed users to send messages of almost any length, and “save” old messages on the server. The second method sent messages directly from one user’s handset to another. This method allowed users to read incoming messages immediately (since they didn’t need to be retrieved from a central server), but it limited the length of those messages to 250-characters and only allowed for minimal storage of old messages on the handset itself (typically, 30-100 messages, depending on the handset’s memory capacity5). After much debate, the team decided to go with the latter option. As Matsunaga described it,
It was the speed of the direct method that convinced us. While the sender’s phone was still transmitting, the receiver’s phone was already receiving! The machine gun-like delivery was mind-boggling. And the small message size was perfect for short bursts of communication that didn’t necessarily warrant a phone call. I thought these types of messages would create a powerful effect because of their immediacy and frequency.
Matsunaga’s staff subsequently came up with a set of symbolic characters that were ideal for condensed messages. The idea was that, by pressing certain keys, users would be able to transmit
5 By 2002, handsets could store up to 1,000 messages.

4

502-031

images that expressed particular sentiments, such as “I love you” (a symbol of a heart), “hopping mad,” “brokenhearted,” “happy,” “fed up,” etc. Altogether, the team embedded 200 such symbols in the i-mode handsets.6

The Hardware: “Under 100 grams and 100 cc”
Because access to the i-mode service would only be possible with a specially-configured phone, the i-mode team worked closely with DoCoMo suppliers to design the i-mode handsets. Enoki’s first request was that the handsets be “under 100 grams and 100 cc, and run on batteries for as long as existing models.”
These simple specifications created a bit of an uproar among DoCoMo’s suppliers, however, who argued that the handsets had to be heavier, bigger, and consume more battery power if they were to include top-of-the-line features such as a large color screen. Enoki, however, refused to back down, even if it meant producing a phone with a tiny black-and-white display and minimal PDA-like functionality: “I knew it was important that the handset look like a phone. It didn’t need to look like a PDA or an electronic planner because that wasn’t what we were selling; we were selling a phone.”
Fortunately for Enoki, in recent years DoCoMo’s dominance in Japan’s wireless market had given the company significant leverage to push suppliers—including NEC, Panasonic, Mitsubishi, and
Fujitsu—to meet its hardware specifications. As a result, DoCoMo handsets were typically sleeker and more advanced than those available in other overseas markets. In addition, DoCoMo was usually able to dictate the timing of new model launches. (For its part, DoCoMo assumed much of the inventory risk for handsets, and subsidized a sizable portion of their cost to consumers. See
Exhibit 2 for additional information about the cellular handset supply chain.)
It was perhaps not surprising, then, that i-mode’s hardware partners ended up producing handsets that met Enoki’s strict specifications. (See Exhibit 3 for examples of some of DoCoMo’s handsets.) The handsets—stamped with the DoCoMo/i-mode brand—had two-inch screens, weighed just 3.5 ounces, offered 400 hours of standby battery life, and transmitted data at 9.6 kbps
(versus the 56 kbps of PC modems).7 A single button on the phone immediately switched the display to i-mode, and there were navigation and scroll keys to allow users to operate i-mode with one hand.
These first-generation handsets cost users roughly ¥20,000–¥30,000 ($167–$250),8 about ¥5,000–¥10,000 more than a conventional handset with no i-mode capability.

Marketing the i-mode Service: “Anytime, Anywhere, Anything”
Whereas Natsuno was largely responsible for the business model, Matsunaga was responsible for developing a marketing plan that would convince consumers to adopt the i-mode service. As she constantly reminded her staff, the idea was to create “the information equivalent of a convenience store: anytime, anywhere, anything.” She explained:

6 Typing on a handset was generally cumbersome; to type a single character, users had to press a given key up to four times.
7 The only visible reference to the manufacturer was a single letter that appeared in front of the handset’s model number. Most

European and U.S. handsets at the time were significantly heavier, with much shorter battery life.
8 All currency calculations herein are based on an exchange rate of $1 = ¥120.

NTT DoCoMo: Marketing i-mode

I’m not a technically savvy person, so the i-mode concept had no appeal to me from a technological standpoint. But I knew the power of information, so that was what I focused on—providing delightful snippets of information that people would find helpful and would brighten up their day.

The Content: Fulfilling the “Happiness” Criterion
In selecting content providers, the team focused on two criteria. The first was utility; the content had to provide some real benefit to subscribers. This involved signing up partners in the financial services, airlines, hotel, retail, and news industries. The second criteria was more whimsical. As
Matsunaga explained, it was important “to make people feel happy when using the i-mode service.”
This involved signing up lots of “fluff” sites, such as those that offered horoscopes, games, musical snippets to create personalized “ring tones,” and graphics for customizing the handset’s “look and feel.” By the time i-mode launched, the team had signed up 67 “i-mode menu” content providers. (See
Exhibit 4 for the different types of i-mode content.) Services ranged from the serious (e.g., banking) to the frivolous (e.g., screensavers). As Matsunaga recalled,
Right after i-mode launched, Bandai, one of our partners, sent me a message on my i-mode phone that had an image of Mickey Mouse saying, “Congratulations on the launch.” I was so thrilled to get that message, the Mickey Mouse was so cute, so endearing, and the message was so personalized. At that moment, I knew the service was going to be commercially viable.
Because I knew that if i-mode could be associated with content that elicited that kind of emotional response from our subscribers, the word-of-mouth would be huge.

The Marketing Campaign: “Warm, Sweet, and Human”
In preparing the advertising campaign for the launch of i-mode, Matsunaga’s goal was to make the technology behind the service as invisible as possible. She thus decided that the ads would make no mention of words like “Internet,” “Web,” or HTML.” Rather, the “i” in i-mode would stand for information, interactivity, and independence. A lower-case “i” was chosen because it looked less threatening and more friendly than an upper-case “I”, and the logo itself was designed to evoke the familiar symbol that adorned information booths in Japanese subways and airports. Matsunaga also rejected more stylish, cutting-edge logos in favor of a yellow design with a round, anthropomorphic look because, as she put it, “I wanted to have something warm, sweet, and human.”
The ads—which appeared on tv/radio, magazines, newspapers, and billboards—featured one of
Japan’s most popular young actresses, the 18-year-old Ryoko Hirosue, and the ad copy focused almost entirely on the informational benefits of an i-mode subscription (see Exhibit 5). Because of this, the spots not only won the appreciation of content partners (many of whom were featured prominently), but they also kept the focus exactly where the i-mode team wanted it—on the content.

300 Yen: The “Impulse” Price
The most contentious element of the marketing plan was price. Some members of the team believed that the cost of a basic subscription to i-mode should be comparable to the cost of subscribing to a PC-based Internet Service Provider (ISP), which typically ran about ¥2,000 ($17) a month. Matsunaga, however, argued for a much lower price:

6

502-031

I knew from my experience in magazines that certain prices created hesitation in consumers, and certain prices didn’t. For example, there was a big difference in magazines that cost ¥500 vs. those that cost ¥300; just a ¥200 discount transformed a thoughtful purchase into an impulse purchase. I wanted i-mode to feel like a “throwaway” purchase, something that cost so little that people didn’t even bother to think about it. No expectations, no regrets.
Ultimately, the team agreed, and by the time of i-mode’s launch, the basic subscription price had been set at ¥300 ($2.50) a month, plus ¥0.3 (one-fourth of a penny) for every data packet sent or received.9 No dial-up was required to access the service; rather, users would be online 24 hours a day. Moreover, they would be charged on the basis of how much content they transmitted, regardless of how long it took to transmit it.
Under this pricing scheme, a single email message with 40 characters would cost ¥1 (just fourfifths of a penny), much less than the typical cost of a one-minute wireless voice call (¥20). And if users chose to pay extra to access sites that charged a subscription fee, DoCoMo would handle the billing so that they would receive a single monthly bill for all of their i-mode-related activity.

An Overnight Success Story i-mode officially launched in February 1999. By all accounts, the service was an immediate success. After six months, it had attracted over 2 million subscribers; by late 2001, that number had topped 30 million. (See Exhibit 6 for i-mode subscriber growth from 1999 to 2001.) As one analyst summed it up: “They lit a match and saw it turn into a bonfire.”10
When the service launched, about half of i-mode users were in their 20s or younger; they tended to use the “always-on” service multiple times a day, often during “niche” times (e.g., while waiting in line). Within a year of launch, it was not uncommon to see groups of Japanese youth silently riding the subways, clutching their i-mode phones in one hand, their eyes fixated on the tiny displays. By now, almost all of them sported handsets that featured their favorite cartoon characters (Hello Kitty was particularly popular), and emitted pop tunes in lieu of the standard ring tone.
Compared to DoCoMo’s other customers, i-mode users were less likely to switch to other operators, spent more on voice calls, and generated more revenue overall.11 And although many sites offered free content, about half of all users paid extra (¥100 to ¥300/month) to subscribe to at least one “premium” site. Users were most willing to pay for ring tones, games, or anything that personalized their handsets. Moreover, users tended to “float in and out of” subscriptions, subscribing for a month or two at one site, then dropping the subscription for a few months to try out a competing site, before perhaps rejoining the original site. i-mode’s seamless billing structure made this kind of trial easy and low-risk. (See Exhibits 7-11 for more on subscribers, usage, and ARPU.)

9 DoCoMo’s packet switching network, in which digital data was set in discreet chunks over shared circuits, made it possible

for DoCoMo to charge subscribers based on the number of data packets sent or received, as opposed to the amount of time spent online. A data packet is equivalent to 128 bytes.
10 Frank Rose, “Pocket Monster,” Wired, September 2001., p. 131.
11 By the end of 2001, the average revenue per i-mode user (ARPU) was ¥8,540: ¥6,960 for voice calls, ¥300 for the i-mode

subscription, ¥1,280 in data transmission fees.

502-031

NTT DoCoMo: Marketing i-mode

Content Providers
As the number of i-mode subscribers grew, so did the number of i-mode content sites. By the end of 2001, i-mode had over 3,000 “i-mode menu” content partners, and although the company was fielding a couple hundred new applications a month, the team remained very selective about who it allowed into the i-mode content community. As Natsuno explained, “This is our way to exercise quality control.”
In order to become an “i-mode menu” partner, content providers had to meet rigid quality standards. Chat rooms and bulletin boards were unacceptable, partners could not offer links to sites outside the i-mode portal, and there were limits on how much advertising space they could sell on their sites. In exchange, partners were given a spot on i-mode’s hierarchical menu. The menu system itself was very democratic; sites were listed by popularity so that sites with broad consumer appeal were rewarded. About one-third of i-mode’s partners took advantage of DoCoMo’s billing service to charge users subscription fees.
Small, entrepreneurial companies benefited the most from an i-mode partnership. One example was Index Corporation, a company that offered (for ¥170/month) romance-oriented fortune-telling on its “Goddess of Love” site. Yoshimi Ogawa, COO of the company, marveled:
When we launched our i-mode service, we didn’t even have a Web presence; it was just a few of us building a tiny site. Today, we operate over 70 different sites, we have over 2 million subscribers, and we are listed on the Japanese stock exchange. And we’re still the number 1 fortune-telling site on i-mode!
Another example was Giga Networks. Prior to i-mode, Giga had manufactured and marketed karaoke systems. Nobunori Kanazawa, Giga’s President, recalled:
When i-mode asked me to offer karaoke over the i-mode network, I was skeptical. But two things persuaded me to give it a try. One, I realized that I didn’t have to make any big investments in new technology; I could just use our existing music files at very little cost. Two,
I liked the billing service they offered. In the end, I agreed to go along, although instead of offering karaoke, I decided to offer ring tones. Today, we have about 15% share of the i-mode ring tone market—that’s over 5 million subscribers—and the business brings in over 6 billion yen a year. Which just goes to show you what a good manager I am! (laughing)
The opportunity was also compelling for existing Internet companies, since an i-mode partnership gave them access to consumers who didn’t even own PCs. “This was a way for us to gain traction across a much broader customer base,” explained Atsushi Kunishige, Director of DLJdirect, an online securities firm. He elaborated:
Just two months after launch, 15% of our customers were executing their trades via i-mode.
We started calling them “washroom traders,” because these were people who wanted to check the market during office hours, but couldn’t use the PCs at their desks. Now that they had the i-mode service, they could sneak into the washroom with their mobile phones!
Even more impressive was the growth in “voluntary” i-mode sites; by 2001, over 50,000 voluntary sites were in existence, and a significant percentage of i-mode traffic was going to these sites.
Voluntary content providers had difficulty charging users for access and had no visibility on the imode portal; however, users could visit these sites by typing a URL directly into their mini-browser, bookmarking the location for return visits. The most popular sites tended to feature chat rooms and party lines for dating services.

8

502-031

Domestic Challenges
Despite the tremendous success i-mode had experienced over the past three years, by early 2002,
Enoki knew the business was facing significant challenges. For one thing, i-mode’s subscriber growth was falling off, and new subscribers (“newbies”) appeared to have lower usage rates than existing subscribers. In addition, competition was getting increasingly intense. i-mode still claimed
60% of Japan’s wireless Internet market but it had two strong competitors, both of whom had entered the market after i-mode’s launch. The first was KDDI, which had about a 24% marketshare and offered a wireless Internet service called “EZWeb.” It was based on a proprietary WAP-based standard (CDMAOne) that provided higher-speed data services (up to 14 kbps) than i-mode’s 9.6 kbps service. The second was J-Phone, a subsidiary of Japan Telecom, which offered a wireless
Internet service called “J-Sky” that used a standard similar to HTML; its 17% marketshare was growing, largely at the expense of KDDI. Japan Telecom was controlled by Vodafone, the world’s largest wireless operator.
Both KDDI and J-Phone were currently marketing handsets that could access “voluntary” i-mode sites, in addition to their own proprietary sites. (By now, many Japanese content providers were providing similar services to all three carriers.) Moreover, J-Phone was focused on trying to be first to market with innovative services; for example, it had recently launched a handset with a built-in camera that allowed users to send photos over the Internet using their cell phones. This handset had been particularly popular among young people, and there was evidence that J-Phone’s hip brand image was also helping the carrier gain traction among this demographic. According to Enoki:
Both KDDI and J-Phone have copied our business model. KDDI is a pretty predictable competitor. J-Phone is a more interesting operation, because they’re trying to target young people now that we’ve become so mainstream. I’m keeping my eye on them. If they continue to make inroads in the youth market, they could become a threat to us down the road.
Complicating matters for Enoki was the fact that i-mode had recently committed to a shift in its business model. Over the past couple of years, the Japanese government had been flooded with complaints about DoCoMo’s market power from content providers who had been unable to achieve coveted “i-mode menu” status. So in response to government pressure, DoCoMo had recently announced that i-mode would allow other Internet service providers to gain open access to its network, beginning in November 2002.
Under the open system, two things would change. First, the i-mode portal would no longer be the default starting point for i-mode subscribers; rather, subscribers would be able to choose any service provider to serve as their default portal. Second, DoCoMo would reduce its basic monthly subscription fee to ¥200, although subscribers who retained i-mode as their default ISP would still pay ¥300.
Among i-mode’s content providers, the response to the “open network” announcement was mixed. Many of the entertainment sites were already struggling with extremely high churn rates
(averaging about 15% a month) due to fierce competition.12 As Giga Network’s Kanazawa cautiously put it: “It’s unclear what the impact of opening up the system is going to be. Certainly we’re concerned about the increase in competition. Plus, until now, all of our growth has come from imode growth. So we also have to prepare for the fact that that is probably going to slow.”

12 “Churn” here refers to the percentage of subscribers who drop their subscription in a given month.

9

NTT DoCoMo: Marketing i-mode

The FOMA Upgrade
In this context, the key challenge facing Enoki was maintaining i-mode’s momentum and continuing to grow the business. To a large extent, this involved managing the migration of i-mode’s existing customer base to subsequent generations of the i-mode service.
In the three years since i-mode’s launch, DoCoMo had periodically introduced new handset models with upgraded capabilities as a mechanism for spurring growth (see Exhibit 12). For example, in December 1999, DoCoMo had been the first carrier to introduce a series of handsets with color displays and improved audio output for better ring tone quality (see Exhibit 3b). Then in June
2000, DoCoMo had been the first to introduce Java-enabled handsets that allowed i-mode users to view animations on their handset. Since Java-based content consumed much more packet traffic than non-Java content, the Java service (branded “i-appli”) had provided DoCoMo with a significant revenue boost. (See Exhibit 13 for an i-appli example.) Next on the horizon was a series of secondgeneration Java handsets that would increase the transmission speed of the i-mode service from 9.6 kbps to 28.8 kbps and allow for larger downloads. (See Exhibit 14 for the estimated download size of various types of content.)
In addition, in the fall of 2001, DoCoMo had become the first carrier in the world to offer a viable
“third-generation” (3-G) service—branded “FOMA” for “Freedom to Multimedia Access”—that transmitted data at 386 kilobytes per second, a speed that was theoretically comparable to PC-based broadband access. Bandwidth crowding had once again become a problem for DoCoMo, and the company was hoping that the release of its 3-G service, which ran on a different network, would alleviate the problem. For the most part, the plan was to graft the i-mode business model onto the faster 3-G network, adding high-bandwidth functionality such as video and music streaming. In prelaunch trials of the 3-G service, however, two handset recalls had been necessary, certain technical limitations (such as a short battery life) had come to light, and participants had complained about poor connection rates. Still, DoCoMo had forged ahead with an October 1, 2001 launch in limited coverage areas.
To date, adoption of FOMA had been modest (90,000 users by March 2002) and DoCoMo was still struggling with questions about how to position, price, and promote FOMA relative to its existing imode service. For example, there were questions about whether FOMA should be targeted toward corporations or consumers. (By now, corporate subscribers made up 15% of DoCoMo’s i-mode subscriber base. This was up significantly from three years ago, and DoCoMo was devoting more resources to its i-mode corporate sales force.) More generally, it was unclear how aggressively
DoCoMo should be marketing FOMA at all, given that not much high-bandwidth content was available yet, and technical “kinks” were still being worked out of the system. (See Exhibit 15 for
FOMA subscriber information. See the Appendix for more on the FOMA service.)

International Challenges
On the international front, Enoki was confronted with a different set of challenges. Globally, the mobile Internet market was stagnant; in fact, it was estimated that i-mode subscribers alone accounted for as much as 30%-35% of the total global mobile Internet installed base. So over the past two years, DoCoMo had spent billions establishing minority stakes in other mobile carriers around the world, including a 16% stake in AT&T Wireless, and a 15% stake in the Dutch KPN. The team was now working with these partners to bring the i-mode model to the U.S. and European markets.

10

502-031

Enoki recognized that there were significant differences between these markets and Japan, including the fact that because no single carrier dominated these markets, no carrier had the clout to dictate handset specifications to global suppliers such as Nokia, Ericsson, and Motorola.
There were also differences in consumer behavior. In Europe, mobile phones were even more prevalent than in Japan; some countries had over 70% penetration rates13 (Exhibit 16). But despite the ballyhooed introduction of WAP-based Internet services by a variety of European carriers over the past couple of years, the wireless Internet had yet to catch on. One problem was that such services were expensive (European operators collected a per-minute fee for WAP usage); another problem was lack of content (content providers were reluctant to cooperate because they didn’t see a clear profit potential). In contrast, a simple, text-messaging service called “SMS” had become extremely popular among European consumers because of its low cost relative to voice calls.
As for the U.S., American consumers were less likely than Japanese consumers to use a cell phone, but more likely to access the Internet via a PC; they thus tended to be skeptical about the benefits of accessing the web via a two-inch screen. Americans were also accustomed to being billed a “flat” monthly rate for unlimited Internet access, rather than a fee based on usage. When it came to cell phone service, they were accustomed to being billed based on airtime.
Finally, in both markets, there were differences in marketing strategies. In both the U.S. and
Europe, operators tended to target business users and other “early adopters” when a new service was introduced. Moreover, carriers were generally not responsible for collecting revenue on behalf of content providers, and were reluctant to open up their network to too many content providers. (As one AT&T Wireless executive put it, “We have enough good content. Do you need 30 ways to check
Mariners scores, or is three or four enough?”14)

Looking Ahead
As Enoki considered the challenges—both domestic and international—facing the i-mode group, he paused to reflect on how quickly things had changed at i-mode during his tenure. Indeed, in the past couple of years, both Matsunaga and Natsuno had written best-selling books about their i-mode experiences. For her part, Matsunaga—who had since become one of Japan’s most high-profile businesswomen—was no longer with the team; to Enoki’s regret, she had departed to pursue other interests. Perhaps most significantly, the “skunk works” atmosphere that he, Matsunaga, and Natsuno had worked so hard to create had clearly become a thing of the past. The i-mode business was simply too big, and too strategically important, for DoCoMo to let “fly under the radar” anymore. Decisions about how to market i-mode to overseas markets, as well as decisions about how to market FOMA to
Japan, could no longer be made by a small handful of individuals; rather, they had to be considered within the context of DoCoMo’s overall strategic objectives. And yet despite all this, Enoki couldn’t help but feel energized and optimistic, for while he appreciated the difficulty of the challenges that awaited, he felt certain that new opportunities awaited as well.

13 The introduction of “prepaid” mobile phones had created a boom in usage among teens; by 2001, prepaid services made up

more than half of the total European subscriber base.
14 Frank Rose, “Pocket Monster,” Wired, September 2001.

502-031

NTT DoCoMo: Marketing i-mode

Exhibit 1

Installed Base of Fixed-Line, Cellular, Internet, and Mobile Internet in Japan

Number of Subscribers (Thousands)

80,000
70,000
60,000
50,000
40,000
30,000
20,000
10,000
0
Mar-91 Mar-92 Mar-93 Mar-94 Mar-95 Mar-96 Mar-97 Mar-98 Mar-99 Mar-00 Mar-01

Fixed-Line Phone Installed Base

NOTE: Total Japanese population = approx. 127 million

Cellular Installed Base
DoCoMo's Customer Base
Internet (via PC) Installed Base
Mobile Internet Installed Base

Source: Company estimates

12

Exhibit 2

502-031

The Cellular Handset Supply Chain

Japanese cellular operators typically offset the cost of handsets to consumers by paying subsidies to dealers. They also paid (both one-time and ongoing) sales commissions to dealers. Estimates for
DoCoMo’s handset subsidies and commissions can be found in the figure below.
The handset-related subscriber acquisition cost (SAC) for Japan’s cellular operators was often quoted to be about ¥50,000. However, it was generally assumed that DoCoMo’s handset-related SAC was lower than its competitors. [For example, by the end of 2001, DoCoMo’s average SAC was estimated to be between ¥30,000 and ¥35,000, below both KDDI (average SAC of ¥50,000) and J-Phone (average
SAC of ¥35,000–¥40,000).]
In addition, DoCoMo’s monthly churn rate was about 1.16%, whereas KDDI and J-Phone’s churn rates were estimated to be about 3%.

handset

handset

•

Handset manufacturers handset

Handset subsidy Cellular
Operators
(Inventory risk)

(¥5,000¥30,000)
One-time
commission
(¥20,000¥23,000)

Dealerships

Users

Includes primary dealerships
(inventory risk), secondary dealerships, and operator specialty stores Ongoing commission (¥200/mo. for 5 years)

Source: Casewriter research, analyst reports (UBS Warburg)

13

Exhibit 3a

NTT DoCoMo: Marketing i-mode

i-mode Handsets: First-Generation (501i Series)

Source: NTT DoCoMo

Exhibit 3b

502-031

i-mode Handsets: Second-Generation (502i Series), Released December 1999

Source: NTT DoCoMo

15

Exhibit 4

NTT DoCoMo: Marketing i-mode

The Different Types of i-mode Content

Transactions:

Information:

•
•
•
•
•
•

•
•
•
•
•
•

Money Transfer
Balance Check
Security Trading
Ticket Reservations
Retail Sales etc. News Updates
Weather Forecasts
Sports Scores
Stock Quotes
Business News etc. Databases:

Entertainment:

•
•
•
•
•
•

•
•
•
•
•
•

Search
Phone Directories
Restaurant Guides
Dictionary Services
Cooking Recipes etc. Character Download
Ring Tones
Horoscopes/Fortunes
Games
Screensavers
etc.

Source: NTT DoCoMo

Exhibit 5

502-031

i-mode Advertisements (1999)

Source: NTT DoCoMo

NTT DoCoMo: Marketing i-mode

Exhibit 6

i-mode Subscriber Growth

DoCoMo users (non i-mode) i-mode users (non-JAVA) i-mode users (JAVA)

45,000

40,000

Number of Subscribers (Thousands)

9,000

35,000

9,453
10,669
12,332

30,000
14,331

25,000
19,700
19,702

20,000
23,753

20,609

15,000

20,994
23,849

20,097

10,000
12,000

10,480

5,000
7,160
5,603
3,995

0

1,598
48

Mar-99

Source:
Note:

Mar-00

//

Mar-01

Jun-01

Sep-01

Dec-01

//

Mar-02
(Est)

Company Records
CAGR from 1999-2001 = 22.8%

18

502-031

Gender Distribution of i-mode Subscribers (May 1999 vs. December 2001)

Exhibit 7

May 1999

December 2001

18%

42%
58%
82%

Male

Male

Female

Source:

Female

Company Records

Age Distribution of i-mode Subscribers (May 1999 vs. December 2001)

Exhibit 8

May 1999

December 2001

unknown

unknown

40+

40+

35-39

35-39

30-34

30-34

25-29

25-29

0-24

0-24
0%

Source:

10%

20%

30%

40%

0%

10%

20%

30%

40%

Company Records

Description of i-mode Usage Patterns (December 2001)

Exhibit 9

Transactions

i-mode
Menu Sites

Database
Information
Entertainment

Voluntary
Sites

Games/Horoscopes
Ringtones/Screens

0%
Source:

10% 20% 30% 40% 50% 60%

0%

10%

20%

30%

40%

Company Records

19

NTT DoCoMo: Marketing i-mode

Exhibit 10

DoCoMo’s Average Revenue Per User (ARPU), Per Month

Monthly Aggregate ARPU (¥)

14,000

i-mode ARPU

12,000

Voice ARPU

10,000
120
880

8,000

1,470

7,770

7,300

1,510

1,580

1,540

7,060

6,960

6,960

12,500

6,000

10,700
9,200

4,000

8,620

2,000
0
Mar-97 Mar-98 Mar-99 Mar-00 Mar-01
Monthly ARPU, by Year

Source:

//

Jun-01 Sep-01 Dec-01

//

Mar-02
(Est)

Monthly ARPU, by Quarter

Company Records

Exhibit 11

DoCoMo’s Monthly ARPU, Compared to Other Cellular Operators
Monthly ARPU
(voice + data)

Data-Only
ARPU

DoCoMo
J-Phone
KDDI

$71
$65
$68

$13
$8
$12

Verizon Wireless (U.S.)
Cingular (U.S.)
Vodafone UK
Orange UK
Orange France

$48
$52
$33
$30
$29

25 cents negligible $3
$3
88 cents

Cellular Operator

Source: Company estimates

20

Exhibit 12

502-031

DoCoMo Handset Introductions

Date

New Features

February

501i
(¥22,000)

First generation i-mode handset

December

1999

DoCoMo handset models

502i
(¥25,000)

Second generation i-mode handset. Added color display and improved music download for ring tones (4 chords max)

2000

June

2001

January

209i
(¥19,000)

Upgrade of the standard DoCoMo handset; all DoCoMo handsets now i-mode compatible.
Music download improved to 16 chords max.

503i
(¥29,000)

April

First JAVA-enabled handset. Allowed for animation, JAVA-enabled applications, SSLcompatible for better security
210i
(¥20,000)

October

All 2xx handsets (DoCoMo’s standard model) now equipped with color display.
N2001,
P2101
(¥40,000–
(¥60,000)

November

211i
(¥18,000)

All 2xx handsets now equipped with SSL for better security
N2002
(¥40,000)

2002
(plan)

Spring

504i

i-motion services launched for FOMA-enabled handsets Second-generation JAVA handset. Will allow for larger JAVA application downloads and faster i-mode transmission speeds (28.8 kbps, instead of 9.6 kbps).

251i

Will include built-in CCD digital camera and picture upload service.
---

Source:

First FOMA-enabled handsets. TV phone, video streaming, and music streaming.
Battery life: 55 hours standby, 100 minutes talk time.

Second-generation FOMA-enabled handset.

Casewriter research

NTT DoCoMo: Marketing i-mode

Exhibit 13

An Example of a Java-Enabled (i-appli) Application

The Java-enabled i-appli service allows users to access dynamic content and download stand-alone applications, which then run automatically. In the example above, an i-appli application keeps track of stock prices.
Charts and stock movements are updated automatically, and users can change the time interval between automatic updates.
Source:

NTT DoCoMo

Exhibit 14

Estimated Download Size of Various Types of Content

Type of Content
Ring Tones
Characters
Screen Savers
Photos
Games
Songs
Video Clips
Source:

Size (KB)
5–10
5–10
10–30
50+
10–100+
100–1000+
100–1000+

Casewriter estimates

Note: Maximum file size for i-mode downloads equals 100 KB, as of early 2002.

Exhibit 15

502-031

Breakdown of FOMA Subscribers, as of November 30, 2001

Corporate/Individual

Male/Female
11%

31%
69%
Corporate

Male

Individual

89%

Female

Age Distribution
60s
50s
40s
30s
20s
Teens
0%

10%

20%

30%

40%

Source: Company records

Exhibit 16

NTT DoCoMo: Marketing i-mode

Penetration Rates for PC Ownership, Internet, Mobile Phone (2001)
PC Ownership

Internet (via PC)

Mobile Phone

Japan
South Korea
Singapore
Hong Kong

42%
68%
64%
59%

27%
58%
60%
56%

50%
59%
72%
83%

USA

65%

52%

42%

France
Germany
UK
Italy
Finland

34%
48%
46%
41%
51%

20%
35%
38%
34%
42%

56%
63%
72%
79%
77%

Source: Company records (mobile phone data), casewriter research adapted from Jupiter, Nielsen/NetRatings

Appendix

502-031

The FOMA Service

DoCoMo originally intended to launch its 3-G FOMA service in May 2001 but technical problems forced the company to impose a five-month trial period. During the trial period, 50% of the trial participants reported being “satisfied” with the fast-speed data access capability, and about 60% rated the new TV phones (which allowed for video streaming) “satisfactory.”
In addition, the average daily packet consumption during the FOMA trial period was 3,600 packets, significantly higher than the 220 packets consumed by typical i-mode users (Java and nonJava combined). Corporate customers used the service even more heavily, as indicated in the graph below: Average Daily Packet Consumption, by Type of Subscriber*
Corporate FOMA user
Individual FOMA user
High-Speed i-mode/Java user i-mode user (Java) i-mode user (non Java)
0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

Data Packets Transmitted Per Day
*Note:

FOMA user estimates are based on the FOMA trial period. “High-Speed i-mode/Java” refers to users of secondgeneration JAVA handsets, which allow for faster i-mode transmission speeds (28.8 kbps, instead of 9.6 kbps), and larger Java applications.

DoCoMo’s FOMA service operated on a W-CDMA network, which DoCoMo hoped would eventually become the global 3G standard.15 In order to subscribe to FOMA, users had to purchase a
DoCoMo W-CDMA handset. At launch, the only data services available to FOMA subscribers (other than standard i-mode content) was a TV phone service. In addition, DoCoMo had launched a new service called “i-motion” in November 2001; “i-motion” allowed the download of files up to 100 kb and video clipping.
The fees for voice calls on the FOMA network were similar to those charged to current DoCoMo subscribers. With respect to data communication, there were four optional basic fees, ranging from
¥2,000 to ¥15,000 per month, depending on the number of “bundled” free calls. By year end, the monthly data-ARPU of the average FOMA user was ¥10,400 (compared to a monthly data-ARPU of
¥1,580 for i-mode users).

15 Both KDDI and J-Phone were also planning 3G services for 2002. J-Phone had adopted the W-CDMA standard, while KDDI had adopted a cdma2000 standard developed by U.S.-based Qualcomm. Fierce global competition was expected between cdma2000 and W-CDMA, and industry analysts were viewing the Japanese market as an early test of the two standards.