What is Samsung thinking?

This is an interesting time for tablet computing fans, with the HP Slate (link) being announced today and a revised B&N Nook (link) supposedly being announced next week. Meanwhile, I'm still coming to the terms with the pricing Samsung announced this week for its upcoming Galaxy Tab.

I had a very strong negative reaction to the price, but I wanted to wait a couple of days to see how I'd feel after I had time to think about it. So now I've thought about it, and here's my reaction:

$600 for a seven-inch tablet?? Are you freaking kidding me? A whole netbook costs about $400. Why does it cost $200 extra just to remove the keyboard?

I don't understand Samsung's strategy. A $400 device is maybe an impulse buy for a rich person at Christmas. A $600 device is a carefully considered investment for most people, especially when all the most enthusiastic tablet buyers have already been siphoned off by Apple.

I got a chance to play with a Galaxy Tab at CTIA. The interface is very cool, but I kept asking myself what I'd actually use it for. What problems does it solve that you can't solve with a smartphone? Samsung appears to assume that Apple has created a market for generic tablets to do, you know, tablet stuff. But has it? Or has it created a market for iPads that seamlessly handle lots of content and unique applications?

And although the design of the Galaxy Tab looks nice, I think the ergonomics of it are questionable. Despite what Samsung's publicity photos show, the device is a bit too wide to hold comfortably even in my dinnerplate-sized hand. To hold it securely, I needed to put my thumb on the front of it. But the margins around the screen are so narrow, and the back case is so slippery, that I felt like I was going to drop it when I put my thumb alongside the screen. The weight of the device also put uncomfortable pressure on my thumb (it's a lever effect). My grip felt more secure and comfortable if I put my thumb on the screen, but then I would accidentally press icons and interfere with the interface.

Although Samsung likes to talk about itself as a leader, in practice it's usually a fast follower -- give it a device to copy and it'll turn out its version faster than just about anyone else on the planet. If the device sells, great. If it doesn't, Samsung just moves on to the next device. My guess is that's what it'll do with the Galaxy Tab.

I'm hoping for better from other new products, although I'm not encouraged by what I'm hearing about the HP device (for one thing, Friday is a terrible day to announce a product because your news coverage gets cut off by the weekend). But I'd like to get my hands on that one before I make up my mind about it.

(Note: This post was modified on 10/22 to correct the announcement date for the HP Slate.)

What's really wrong with BlackBerry (and what to do about it)

Just a couple of weeks after Research in Motion turned in a good earnings report, the death watch over the company has resumed, with Business Week magazine running a long article that mocks co-CEO Jim Balsillie (even picking on his duck-emblazoned tie) and saying that RIM needs to learn how to market as well as Apple (link).

Business Week quoted Balsillie at a press briefing:

"There's tremendous turbulence in the ecosystem, of course, in mobility. And that's sort of an obvious thing, but also there is tremendous architectural contention at play. And I'm going to really frame our mobile architectural distinction. We've taken two fundamentally different approaches in their causalness. It's a causal difference, not just nuance. It's not just a causal direction that I'm going to really articulate here -- and feel free to go as deep as you want -- it's really as fundamental as causalness."

OK, he deserves to be mocked for that. But Business Week goes on to conclude that his quote captures the whole dilemma of the company -- technical sophistication coupled with incoherent marketing.

Business Week has joined a large and distinguished group of experts taking jabs at RIM. Morgan Stanley recently downgraded RIM's stock, saying it's going to lose share faster than previously expected (link). Gartner reported that Android had passed BlackBerry to become the most popular smartphone OS in the US (link). And CNET said RIM is about to be kicked out of the enterprise market (link).

I've been getting very tired of the criticisms of RIM, because most of them seem superficial and some are petty. Yes, Android is doing well, but neither RIM nor Apple is giving away its operating system, so it was close to inevitable that Android would eventually get the unit lead. It's the default choice for most smartphone companies, so of course it moves a lot of units in aggregate. But there is room in the market for several mobile platforms to succeed. The companies Android is hurting most are Microsoft, Access, and others that were hoping to sell mobile operating systems.

Yes, RIM's not good at sexy marketing, but it has always been that way. People have been predicting its imminent doom for as long as I can remember (do you recall when Microsoft Exchange was supposed to destroy it?). My guess is that the folks at RIM are shaking their heads at all of the bad press and assuming it will once again blow over in a quarter or two.

I think that would be a serious mistake. In my opinion, RIM is indeed in danger, probably a lot more danger than its executives realize. But I don't agree on the reasons most people are giving for why RIM is in trouble, and I think most of the solutions that are being proposed would make the situation worse, not better.

The fault lies not in our ties, but in our selves. In my opinion, RIM's real problems center around two big issues: its market is saturating, and it seems to have lost the ability to create great products. This is a classic problem that eventually faces most successful computer platforms. The danger is not that RIM is about to collapse, but that it'll drift into in a situation where it can't afford the investments needed to succeed in the future. It's very easy for a company to accidentally cross that line, and very hard to get back across it.

There's a lesson in RIM's situation for every tech company, so it's worthwhile to spend some time understanding what's happening.

How a computing platform dies

To explain RIM's challenges, I have to give you a little tech industry history. When I worked at Apple, I spent a lot of time studying failed computer platforms. I thought that if we understood the failures, we might be able to prevent the same thing from happening to us.

I looked at everything from videogame companies to the early PC pioneers (companies like Commodore and Atari), and I found an interesting pattern in their financial results. The early symptoms of decline in a computing platform were very subtle, and easy for a business executive to rationalize away. By the time the symptoms became obvious, it was usually too late to do anything about them.

The symptoms to watch closely are small declines in two metrics: the rate of growth of sales, and gross profit per unit sold (gross margins). Here's why:

Every computing platform has a natural pool of customers. Some people need or want the platform, and some people don't. Your product spreads through its pool of customers via the traditional "diffusion" process -- early enthusiasts first, late adopters at the end.

It's relatively easy to get good revenue from the early adopters. They seek out innovations like yours, and are willing to pay top dollar for it. As the market for a computer system matures, the early adopters get used up, and the company starts selling to middle adopters who are more price-sensitive. In response to this, the company cuts prices, which results in a big jump in sales. Total revenue goes up, and usually overall profits as well. Everybody in the company feels good.

Time passes, and that middle portion of the market gets consumed. Eventually demand growth starts to drop, and you make another price cut. Sales go up again, sometimes a lot. With revenue rising, you and your investors talk proudly about the benefits of reaching the "mainstream" market.

At Apple, when we hit this point we called our low-cost products the Macintosh Classic and Macintosh LC. At Palm, it was the M100.

What you don't realize at this point is that you're not "reaching the mainstream," you're actually consuming the late adopters. Unfortunately, it's very difficult to tell when you're selling to the late adopters. They don't wear signs. Companies tend to assume that because the adoption curve is drawn as a smooth-sided bell, your demand will tail off at the end as gradually as it built up in the beginning. But that isn't how it works. At the start, you are slowly building up momentum from a base of nothing. That takes years. But by the time you saturate the market you have built up huge sales momentum. You have a strong brand, you have advertising, you have a big distribution channel. You'll gulp through the late adopters really rapidly. The result is that sales continue to grow until they drop suddenly, like a sprinter running off the edge of a cliff.

The chart below illustrates how the process works:

Until you get close to the end, your revenue keeps rising, enabling you to tell yourself that the business is still in good shape. But eventually you reach the dregs of the market, and sales will flatten out, or maybe even start to drop. You cut prices again, but this time they don't increase demand because there are no latent customers left. All the cuts do is reduce further the revenue you get from selling upgrades to your installed base. The combination of price cuts and declining sales produces a surprisingly rapid drop in revenue and profits. If you want to make a profit (which your investors demand), your only choice is to make massive cuts in expenses. Those cuts usually end up eliminating the risky new product ideas that are your only hope of re-igniting demand.

At Apple I called this the platform "death spiral" because once you get into it, the expense cuts and sales declines reinforce each other. It's almost impossible to reverse the process, unless you're Steve Jobs and you get very lucky.

The best way to survive is to stay away from the cliff edge in the first place. But that means you need to be hyper-attentive to small changes in sales growth and gross margins. Which brings us back to RIM's situation.

Dissecting RIM's financials

At the top level, RIM's financials look utterly fantastic:

RIM Revenue and Profit

Fiscal years. Dollars in millions.

Since fiscal 2003 (when it turned profitable), RIM has grown from $500m revenue to over $15 billion. That's 30X growth in eight years. The BlackBerry subscriber base has grown from 500,000 people to about 50 million. Throughout that period, the company's net income has hovered at between 15% and 22% of revenue.

This is one of the most impressive business success stories of the last decade, and most CEOs in any industry would kill to have that sort of results. Considering how much turmoil there is in the smartphone market, RIM's senior managers must feel extremely proud of their success, and more than a bit bewildered that people keep criticizing them.

And that's exactly my point. Looking at the high-level financials can lull you into a false sense of security if you're managing a computing platform. You have to really dig to find the warning signs. That's especially hard to do in RIM's case because the company has several different sources of revenue: device sales, service revenue, and enterprise server revenue. The overall results they report are mashup of all three revenue streams. To understand what's really happening, you have to tease them apart. Here are some key data points.

First, let's look at the total number of BlackBerry subscribers:

Total BlackBerry Subscribers

RIM's fiscal quarters. Units in millions.

Pretty impressive growth. But remember, we're looking for subtle signs of saturation. Let's look at the number of subscribers added per quarter...

Net New Subscribers Per Quarter

RIM's fiscal quarters. Units in millions.

This is where you get the first little twinge of discomfort. Until a year ago, the rate of growth of BlackBerry subscribers was itself increasing every quarter. In other words, RIM added more new subscribers each quarter than it had added in the previous quarter. But for the last four quarters, RIM's subscriber growth has plateaued at around 4.7 million net new subscribers a quarter. The company's still growing, but it looks like the rate of growth may be flattening. That might imply the beginning of saturation.

Next let's look at net new subscribers as a percent of total BlackBerry units sold.

New Subscribers Added Per Unit Sold

RIM's fiscal quarters.

This one's a little disquieting as well. Five years ago, RIM was getting .7 new subscribers for every BlackBerry sold. In other words, most of its sales were to new users. Today, RIM is getting .37 more subscribers per BlackBerry sold, and that figure is at an all-time low. To put it another way, RIM now has to sell more than two and a half devices to get one more subscriber. Either RIM is selling most of its units to its installed base, or it is having to bring in a lot of new customers to replace those who are leaving for other devices. My guess is it's a mix of both.

If you look closely at that chart, you'll notice a curious bump in the line at Q4 of 2009. The percentage of new subscribers went back up all of a sudden. What did RIM do to produce that growth? A look at device gross margins tells you.

Device Gross Margin Percentage

RIM's fiscal quarters.

[Note: RIM does not report separately the gross margins it gets in the devices business, so I had to estimate this number using the company's hardware revenue and the total cost of goods sold across all of its businesses. Most of RIM's total COGS are hardware expenses, but they also include some server costs associated with providing e-mail service. That means my calculation understates RIM's device margins by a bit. But as the company grows, server costs should go down as a percent of overall costs (because you get better economies of scale). So apparent hardware margins should be going up over time. That makes the fact that they're declining all the more ominous.]

RIM increased new subscriptions by substantially cutting the profit it makes per device. What happened is that the BlackBerry Bold, Storm, and Curve all came to market with increased features, replacing older devices that were much cheaper to build. That should have produced only a one-time hit to margins, though -- they should have gone back up as component costs on the new phones declined. Instead, margins have stayed down ever since. Why? Let's look at the what RIM gets paid for each BlackBerry it sells:

RIM's Revenue Per BlackBerry Device Sold

RIM's fiscal quarters. Hardware revenue per unit sold.

This chart shows the average price the carriers pay to RIM per phone, prior to the discount they put on the phone when you sign up for a contract. The line looks pretty flat, and in fact through the middle of fiscal 2009 RIM's price per unit was very stable. Then in Q3, with the introduction of the new devices, RIM gets a temporary spike in revenue per unit. The new phones are selling at a premium. But that goes away in the next two quarters, and then about a year ago, RIM started cutting prices. Today the company gets about $50 less per unit than it usually did in the past.

When you assemble the big picture, it looks like this: To keep growing, RIM has been forced to reduce margins and prices. Despite the cuts, the rate of growth in subscribers appears to have flattened out. And more and more of the sales mix is going to existing users, or user replacement, rather than new users. RIM starts to look like a company that's working harder and harder just to stay in one place.

The picture gets more ominous when you look at some recent surveys of smartphone user satisfaction. In JD Power's 2010 smartphone satisfaction survey, BlackBerry finished near the bottom, with below average ratings in every category except battery life (link). Just three years earlier, as the iPhone was coming to market, BlackBerry had the highest satisfaction ratings in the industry (link). I don't love JD Power's methodology (for reasons that are too long to explain here), but no way should RIM's rating be declining like that.

The low satisfaction is starting to threaten RIM's future sales. In June of this year, Nielsen released some tidbits from a survey of the future purchasing plans of smartphone users (link):

OS Preferences of People Planning to Replace Their Smartphones

The chart shows US smartphone users who were thinking about buying a new device in Q1 of 2010. More than half of the BlackBerry users considering a new smartphone were leaning toward a different OS.

If I were working at RIM, that chart would scare the crap out of me.

The company is by no means dead, but the symptoms of a stalling platform are definitely there. If you work at RIM and are reading this, here's what I want you to understand: Your company's at risk. Your great financials mask that risk, and give you lots of logical-sounding reasons to avoid making the changes that need to be made. RIM is like a 53-year-old man who has high blood pressure and cholesterol but tells himself that he's OK because he can still run a half-marathon. You are indeed fine, right up until you have the heart attack. Then it's too late.

Here's what you need to do:

How to avoid the cliff

To keep a platform viable, you need to focus on two tasks: Keep the customer base loyal, and add adjacent product categories.

Keeping the base loyal. This is transcendently important to a platform company. As your market matures, more and more of your sales will come from replacement devices sold to the installed base. You'll also depend more and more on a base of developers who add value to your products. If you can keep these people happy, you'll have a steady stream of replacement sales that you can build on. It won't be enough to produce the growth that your investors want, but it'll be a great foundation.

On the other hand, if these customers and developers drift away, there's virtually no way you can grow something else fast enough to offset their loss. The trick here is that the supporter base for a computing platform is like a herd of cattle. They move as a group. When the herd is contented, it tends to stay in one place. But if the herd gets restless, even a small disturbance can cause a stampede in which they all run away at once.

For example, this is the factor that HP failed to consider when it bought Palm. The Pre's small base of users and developers was a classic group of restless cattle. When HP bought the company, the first priority should have been to calm those people by promising a renewed commitment to the Pre and follow-on products. Even if HP didn't see smartphones as its long-term future, it should have focused on keeping the developers and users loyal until it had something else for them to buy and develop for. Instead, HP CEO Mark Hurd more or less killed the product line a day after the purchase (link):

HP won't "spend billions of dollars trying to go into the smartphone business; that doesn’t in any way make any sense....We didn’t buy Palm to be in the smartphone business. And I tell people that, but it doesn’t seem to resonate well. We bought it for the IP."

Ooookay, so if you're a Pre customer, do you buy again? Do you tell your friends to buy? If you're a WebOS developer, do you keep writing code while you wait for HP to decide what it'll do with that "IP" it bought?

The answer is, you run for the exit as fast as you can. HP bought a company for a billion dollars and then immediately trashed it.

Back to RIM. Your cattle are restless. If you don't believe me, go look at that Nielsen chart again. Your goal is to keep the cattle content, by feeding them a steady diet of delightful new products that deepen their commitment to the platform. RIM's record in this area is very mixed. There have been a lot of new BlackBerry products announced in the last few years, but most of them seem to be focused on copying things Apple has done rather than finding new ways to delight BlackBerry customers.

Some of the Apple imitation is probably necessary. Apple has turned a lot of features into checkoff items that are now expected from any smartphone -- a better browser, for example. If RIM didn't eventually add those features, the herd would at some point stampede away for sure.

But what I haven't seen from RIM is a vision for deepening the special features that made people bond with BlackBerry in the first place. The personal communication functionality of BlackBerry is about the same now as it was five years ago. Why in God's name was Apple the first North American smartphone company to really push video calling? As the communication beast, RIM should have led that years ago.

Instead, the latest BlackBerry devices feel a bit like an overbuilt ice cream sundae -- the original BlackBerry functionality is at the base more or less unchanged, and a bunch of gooey media toppings have been dumped on top of it. I see sprinkles, fudge, marshmallow, pineapple, whipped cream, a cherry, and a few gummy bears, but no significant improvement to the old, dried-out ice cream at the bottom of the bowl.

Inevitably, RIM can't implement those new media toppings as cleanly and elegantly as Apple did, because its platform wasn't designed for that. So what you get is a BlackBerry that endorses Apple's design direction but fails to fully deliver on it. Maybe that helps keep some BlackBerry users from leaving instantly, but it doesn't give them a positive reason to stay. Rather than playing to win, RIM is playing not to lose, and doing it poorly.

This is especially scary because RIM depends much more than Apple on mobile operators to help drive demand for its products (if you're in the US, ask yourself how many Verizon and AT&T ads you have seen for BlackBerry, versus how many ads you've seen from RIM itself). The operators follow customer interest, they don't create it. If they get the sense that BlackBerry users want to switch, they will be only too happy to facilitate that switch -- especially since they don't have to share service revenue with Android vendors the way they do with RIM.

What RIM should do. RIM need a product vision identifying a few new differentiators for BlackBerry that will resonate well with the busy knowledge workers who are at the core of its installed base. There should be no more than three of these features (because customers can't remember more than three), and they should not be copies of things that Apple is already implementing. RIM should focus on building them deeply into the product, so they are very well integrated with the rest of the device. My nominees are meeting planning, conferencing, and live document sharing.

Other smartphone companies will eventually copy these features, so RIM needs to create a pipeline of development in which it'll bring out another 2-3 new differentiators every 24 months.

Adding adjacent categories. Settling down the installed base is not enough. It's an enormous task, but all it'll do is stabilize the business. It won't produce the growth that investors expect. To get that, RIM needs to eventually add new types of product that expand its market.

Apple is a master at this process. When Steve Jobs came back, Apple had only the Macintosh. It refreshed that product line, securing the customer base. Then it added the iPod, iPhone, and iPad. Each of them targeted Apple's core market of creative, entertainment-loving people, and each of them leveraged Apple's existing software and hardware. This overlap made the new products relatively inexpensive to develop and market -- they could be sold to the same sorts of people, through the same channels, and they reused a lot of technology. Each new product line also tended to drag a few more customers back to the earlier products, so they reinforced each other.

These new products enabled Apple to grow its revenue rapidly without putting pressure on the Macintosh to carry the whole load. Apple could invest enough in the Mac to keep it a stable and very profitable business, while the new products produced the topline growth.

To understand how wickedly efficient Apple's business model is, take a glance at the R&D budgets of RIM and Apple.

Quarterly R&D Spending of Apple and RIM

R&D spending in most recent four quarters. Dollars in millions.

Although Apple has about three times the revenue, RIM's R&D spending is about two-thirds of Apple's. With just a third more money, Apple produces the Macintosh, iPod, iPhone, iPad, Apple TV, iTunes, App Store, custom microprocessors, and a suite of mobile services. RIM is producing a bunch of minute variations on a family of phones, an e-mail server, a new OS, and a suite of mobile services that also has to be individually interfaced to each operator. RIM puts much of its effort into infrastructure that has little or no impact on features that users can see and value.

Now RIM wants to add more product lines. Its first effort will be the PlayBook tablet in 2011. This will be a decisive test of RIM's ability to grow in the future, and so far the signs are worrisome. Unlike Apple's first announcement of the iPhone, the PlayBook announcement didn't show much functionality that looked fundamentally new compared to the competition (in fact, the interface looked to me a lot like a warmed-over version of Palm's WebOS). The pitch was almost all about enabling technology rather than user benefits. When you find yourself talking up the dual-core processor and symmetric multiprocessing in a consumer product, it's a sign of a serious lack of differentiation.

I'd be more hopeful about the prospects for the PlayBook if RIM had done a better job of evolving its BlackBerry products recently. Unfortunately, RIM's latest innovation flagship is the BlackBerry Torch, an overproduced heap of half-integrated features that ranks as one of the most disappointing mobile devices I've seen from a major manufacturer in years.

Yeah, I know there are some people who like the Torch. But there were also people who thought MS-DOS was easy to use.

Burned by the Torch. I recently bought a BlackBerry Torch for my wife, who needed a smartphone to manage work e-mail. We both wanted her to have something simple to use, with a keyboard that made her comfortable. She liked the Torch in the store, so we bought it for her.

The device was a usage nightmare. Even after years of working with touch screen technology, RIM hasn't managed to evolve its user interface to the point where the touch pad and the touch screen work together smoothly. Some functions are easier to perform on touch screen, and others are easier on touch pad, and so the whole interface feels muddled. But by far the more disappointing problem was that the huge number of new applications just added to the phone do not work together properly. I can't even list all of the problems we both had figuring out how to use them, but one vivid example should suffice. My wife entered a lot of contacts directly into the device's contacts app, but didn't bother to include the area code in the phone numbers. The BlackBerry didn't warn her about this.

Then she went to the messaging app and tried to send a text message to our daughter. When she tried to send the message, the app reported that it could not send to a contact without an area code. So she went back to the contacts app and added area codes.

Then she went back to the messaging app and again tried to send a text message. The messaging app reported once again that it could not send a message without an area code. It had apparently made a copy of the data from the contacts app when it was first used, and would not update the copy. So my wife then edited the contact information from within the contacts app (it lets you do that). But when she tried to save the updated contact, the phone responded that it could not accept external changes to the contacts, and deleted the change.

Next, she tried to send a message by typing our daughter's phone number, including area code, directly into the To: portion of a new message. When she tried to send that message, the messaging application did a lookup on its contacts database, changed the phone number back to the version without an area code, and then reported that it could not send the message because the phone number lacked an area code.

Using the BlackBerry Torch is like being trapped in a real-life version of "Waiting for Godot."

I've seen this sort of incoherent design before. It happens when you have several teams working on parts of the device, and you haven't done proper planning up front to make sure the apps will work together well. It is a symptom of an out-of-control development process. The fact that this happened on RIM's flagship product is deeply disturbing. If the same incompetent processes are applied to the PlayBook -- a much more complex product with a lot of new functionality -- it is almost certain to fail.

By the way, we returned the phone.

What RIM should do. To fix this problem, RIM needs to create rigorous up-front planning processes in its software team, with someone who has dictatorial power placed in charge of overall software integration for a device or OS release. Also, the product manager needs to be empowered (actually required) to delay shipment of a product if it's not right. I'm sure someone at RIM knew about the problems in the Torch. The fact that the company went ahead and shipped it is almost as disturbing as the problems themselves.

Rescuing RIM

To sum up, RIM is at risk because its natural market is saturating and many of its customers are considering a switch to other platforms. The company may be able to bumble along in this situation for years before the problem comes to a head, but once a migration away from BlackBerry starts it would be almost impossible to stop. So if the company wants to ensure its survival, it needs to act now. Two steps are needed:

--The BlackBerry line needs to be given a several fundamental, visionary innovations that will give its core customers a reason to stay; and

--The company needs to change its development process to guarantee proper design and integration in all of its products.

Given the time needed to create a new product, these changes will take at least 18 months to bear fruit, probably more like two years. During that time RIM will remain at risk of a platform collapse. What's worse, the company's engineers already have their hands full copying iPhone features, customizing phones for a huge range of operators, and simultaneously creating a new operating system and developing a new version of the current one. The sort of changes I'm suggesting would disrupt that work, forcing the cancellation of some projects and slips in the schedule for others. They would make the problem worse before they make it better. In the meantime, the company would lose serious revenue, and might even miss earnings projections for a quarter or two. The stock's value would be trashed, and there would be calls for firing management.

As the founders of the company, Jim Balsillie and Mike Lazaridis could probably pull this off without losing their jobs. And I know they have the courage to make big changes. But I doubt they can see the need, or especially the urgency. Their current processes and business practices got them to $15 billion in revenue; why should they change now? It's much more prudent to focus on making the numbers for next quarter.

That's probably just what RIM will do. And if it does, that's why the company will probably eventually fail.


[Edit: Since this post is still getting a lot of traffic, I wanted to let you know that I've posted a look at RIM's Q3 FY 2011 financials, with  updated charts and a deeper look at international sales.  I think the situation is both better and worse than I originally believed (link).  And you can see my take on their June 2011 layoff announcement here.]