Internet of Things

Puzzling Over Google’s Nest Acquisition


Looking past the glitter, big names, and big money ($3.2B), a deeper look at Google’s last move doesn’t yield a good theory. Perhaps because there isn’t one.

Last week’s Monday Note used the “Basket of Remotes” problem as a proxy for the many challenges to the consumer version of the IoT, the Internet of Things. Automatic discovery, two-way communication, multi-vendor integration, user-interface and network management complexity… until our home devices can talk to each other, until they can report their current states, functions, and failure modes, we’re better off with individual remotes than a confusing — and confused — universal controller..

After reading the Comments section, I thought we could put the topic to rest for a while, perhaps until devices powered by Intel’s very low-power Quark processor start shipping.


A few hours later, Google announced its $3.2B acquisition (in cash) of Nest, the maker of elegant connected thermostats and, more recently, of Nest Protect smoke and CO alarms. Nest founder Tony Fadell, often referred to as “one of the fathers of the iPod”, takes his band of 100 ex-Apple engineers and joins Google; the Mountain View giant pays a hefty premium, about 10 times Nest’s estimated yearly revenue of $300M.


Tony Fadell mentioned “scaling challenges” as a reason to sell to Google versus going it alone. He could have raised more money — he was actually ready to close a new round, $150M at a $2B valuation, but chose adoption instead.

Let’s decode scaling challenges. First, the company wants to raise money because profits are too slim to finance growth. Then, management looks at the future and doesn’t like the profit picture. Revenue will grow, but profits will not scale up, meaning today’s meager percentage number will not expand. Hard work for low profits.

(Another line of thought would be the Supply Chain Management scaling challenges, that is the difficulties in running manufacturing contractors in China, distributors and customer support. This doesn’t make sense. Nest’s product line is simple, two products. Running manufacturing contractors isn’t black magic, it is now a well-understood trade. There are even contractors to run contractors, two of my friends do just that for US companies.)

Unsurprisingly, many worry about their privacy. The volume and tone of their comments reveals a growing distrust of of Google. Is Nest’s expertise at connecting the devices in our homes simply a way for the Google to know more about us? What will they do with my energy and time data? In a blog post, Fadell attempts to reassure:

“Will Nest customer data be shared with Google?
Our privacy policy clearly limits the use of customer information to providing and improving Nest’s products and services. We’ve always taken privacy seriously and this will not change.”

What else could Fadell offer besides this perfunctory reassurance?  “[T]his will not change”… until it does. Let’s not forget how so many tech companies change their minds when it suits them. Google is no exception.

This Joy of Tech cartoon neatly summarizes the privacy concern:


The people, the brands, the money provide enough energy to provoke less than thoughtful reactions. A particularly agitated blogger, who can never pass up a rich opportunity to entertain us – and troll for pageviews – starts by arguing that Apple ought to have bought Nest:

“Nest products look like Apple products. Nest products are beloved by people who love Apple products. Nest products are sold in Apple stores.
Nest, in short, looked like a perfect acquisition for Apple, which is struggling to find new product lines to expand into and has a mountain of cash rotting away on its balance sheet with which it could buy things.
[...] Google’s aggressiveness has once again caught Apple snoozing. And now a company that looked to be a perfect future division of Apple is gone for good.”

Let’s slow down. Besides Nest itself, two companies have the best data on Nest’s sales, returns, and customer service problems: Apple and Amazon. Contrary to the “snoozing” allegation, Apple Store activity told Apple exactly the what, the how, and the how much of Nest’s business. According to local VC lore, Nest’s Gross Margin are low and don’t rise much above customer support costs. (You can find a list of Nest’s investors here. Some, like Kleiner Perkins and Google Ventures, have deep links to Google… This reminds many of the YouTube acquisition. Several selling VCs were also Google investors, one sat on Google’s Board. YouTube was bleeding money and Google had to “bridge” it, to loan it money before the transaction closed.)

See also Amazon’s product reviews page; feelings about the Nest thermostat range from enthusiastic to downright negative.

The “Apple ought to have bought Nest because it’s so Apple-like” meme points to an enduring misunderstanding of Apple’s business model. The Cupertino company has one and only one money pump: personal computers, whether in the form of smartphones, tablets, or conventional PCs. Everything else is a supporting player, helping to increase the margins and volume of the main money makers.

A good example is Apple TV: Can it possibly generate real money at $100 a puck? No. But the device expands the ecosystem, and so makes MacBooks, iPads, and iPhones more productive and pleasant. Even the App Store with its billions in revenue counts for little by itself. The Store’s only mission is to make iPhones and iPads more valuable.

With this in mind, what would be the role of an elegant $249 thermostat in Apple’s ecosystem? Would it add more value than an Apple TV does?

We now turn to the $3.2B price tag. The most that Apple has ever paid for an acquisition was $429M (plus 1.5M Apple shares), and that was for… NeXT. An entire operating system that revitalized the Mac. It was a veritable bargain. More recently, in 2012, it acquired AuthenTec for $356M.

With rare exceptions (I can think of one, Quattro Wireless), Apple acquires technologies, not businesses. Even if Apple were in the business of buying businesses, a $300M enterprise such as Nest wouldn’t move the needle. In an Apple that will approach or exceed $200B this calendar year, Nest would represent about .15% of the company’s revenue.

Our blogging seer isn’t finished with the Nest thermostat:

“I was seduced by the sexy design, remote app control, and hyperventilating gadget-site reviews of Nest’s thermostat. So I bought one.”

But, ultimately, he never used the device. Bad user feedback turned him off:

“[…] after hearing of all these problems, I have been too frightened to actually install the Nest I bought. So I don’t know whether it will work or not.”

He was afraid to install his Nest… but Apple should have bought the company?

So, then, why Google? We can walk through some possible reasons.

First, the people. Tony Fadell’s team is justly admired for their design skills. They will come in handy if Google finally gets serious about selling hardware, if it wants to generate new revenue in multiples of $10B (its yearly revenue is approximately $56B now). Of course, this means products other than just thermostats and smoke alarms. It means products that can complement Google’s ad business with its 60% Gross Margin.

Which leads us to a possible second reason: Nest might have a patent portfolio that Google wants to add to its own IP arsenal. Fadell and his team surely have filed many patents.

But… $3.2B worth of IP?

This leaves us with the usual questions about Google’s real business model. So far, it’s even simpler than Apple’s: Advertising produces 115% or more of Google’s profits. Everything else brings the number back down to 100%. Advertising is the only money machine, all other activities are cost centers. Google’s hope is that one of these cost centers will turn into a new money machine of a magnitude comparable to its advertising quasi-monopoly.

On this topic, I once again direct you to Horace Dediu’s blog. In a post titled Google’s Three Ps, Horace takes us through the basics of a business: People, Processes, and Purpose:

“This is the trinity which allows for an understanding of a complex system: the physical, the operational and the guiding principle. The what, the how and the why.”

Later, Horace points to Google’s management reluctance to discuss its Three Ps:

“There is a business in Google but it’s a very obscure topic. The ‘business side’ of the organization is only mentioned briefly in analyst conference calls and the conversation is not conducted with the same team that faces the public. Even then, analysts who should investigate the link between the business and its persona seem swept away by utopian dreams and look where the company suggests they should be looking (mainly the future.)
There are almost no discussions of cost structures (e.g. cost of sales, cost of distribution, operations and research), operating models (divisional, functional or otherwise) or of business models. In fact, the company operates only one business model which was an acquisition, reluctantly adopted.”

As usual — or more than usual in current circumstances — the entire post is worth a meditative read. Especially for its interrogation at the end:

“The trouble lies in that organization also having de-facto control over the online (and hence increasingly offline) lives of more than one billion people. Users, but not customers, of a company whose purpose is undefined. The absence of oversight is one thing, the absence of an understanding of the will of the leadership is quite another. The company becomes an object of faith alone.  Do we believe?”

Looking past the glitter, the elegant product, the smart people, do we believe there is a purpose in the Nest acquisition? Or is Google simply rolling the dice, hoping for an IoT breakthrough?


Internet of Things: The “Basket of Remotes” Problem


We count on WiFi and Bluetooth in our homes, but we don’t have appliances that provide self-description or reliable two-way communication. As a result, the Internet of Things for consumers is, in practice, a Basket of Remotes.

Last Friday, I participated in a tweetchat (#ibmceschat) arranged by friends at IBM. We discussed popular CES topics such as Wearables, Personal Data, Cable and Smart TV, and the Internet of Things. (I can’t help but note that Wikipedia’s disambiguation page bravely calls the IoT “a self-configuring wireless network between objects”. As we’ll see, the self-configuring part is still wishful thinking.)

At one point, the combined pressures of high-speed twittering and 140-characters brevity spurred me to blurt this:

Remotes Basket Case

A little bit of background before we rummage through the basket.

In practice, there are two Internets of Things: One version for Industry, and another for Consumers.

The Industrial IoT is alive and well. A gas refinery is a good example: Wired and wireless sensors monitor the environment, data is transmitted to control centers, actuators direct the flow of energy and other activities. And the entire system is managed by IT pros who have the skill, training, and culture — not to mention the staff — to oversee the (literal) myriad unseen devices that control complicated and dangerous processes.

The management of any large corporation’s energy, environment, and safety requires IT professionals whose raison d’être is the mastery of technology. (In my fantasy, I’d eavesdrop on Google’s hypergalactic control center, the corporate Internet of Things that manage the company’s 10 million servers…)

Things aren’t so rosy in the consumer realm.

For consumers, technology should get out of the way — it’s a means, not an end. Consumers don’t have the mindset or training of IT techies, they don’t have the time or focus to build a mental representation of a network of devices, their interactions and failure modes. For example, when my computer connects to the Net, I don’t have to concern myself with the way routers work, how the human-friendly gets translated into the IP address.

Not so with a home network of IoT objects that connect the heating and cooling systems, security cameras, CO and fire sensors, the washer, dryer, stove, fridge, entertainment devices, and under-the-mattress sleep monitoring pads. This may be an exaggerated example, but even with a small group of objects, how does a normal human configure and manage the network?

For an answer, or lack thereof, we now come back to the Basket of Remotes.

I once visited the home of an engineer who managed software development at an illustrious Silicon Valley company. I was shocked, shocked to see a basket of remotes next to the couch in front of his TV. ‘What? You don’t use a programmable remote to subsume this mess into one elegant device and three of four functions, TV, DVR, VoD, MP3 music?’

‘No, it’s too complicated, too unreliable. Each remote does its separate job well, with an easy mental representation. These dumb devices don’t talk back, there’s no way for a unified remote to ask what state they’re in. So I gave up — I have enough mental puzzles at the office!’

Indeed, so-called “smart” TVs are unable to provide a machine-readable description of the commands they understand (an XML file, also readable by a human, would do). We can’t stand in front of a TV with a “fresh” universal remote – or a smartphone app – touch the Learn button and have the TV wirelessly ship the list of commands it understands…and so on to the next appliance, security system or, if you insist, fridge and toaster.

If an appliance would yield its control and reporting data, an app developer could build a “control center” that would summarize and manage your networked devices. But in the Consumer IoT world, we’re still very far from this desirable state of affairs. A TV can’t even tell a smartphone app if it’s on, what channel it’s tuned to, or which devices is feeding it content. For programmable remotes, it’s easy to get lost as too many TVs don’t even know a command such as Input 2, they only know Next Input. If a human changes the input by walking to the device and pushing a button, the remote is lost. (To say nothing of TVs that don’t have separate On and Off commands, only an On/Off toggle, with the danger of getting out of sync – and no way for the TV to talk back and describe its state…)

Why don’t Consumer Electronics manufacturers provide machine self-description and two-way communication? One possible answer is that they’re engaged in a cost-cutting race to the bottom and thus have no incentive to build more intelligence into their devices. If so, why build unbearably dumb apps in their Smart TVs? (Korean LG Electronics even dug up WebOS for integration into its latest TVs.)

A look at Bang & Olufsen’s Home Integration page might give one hope. The video demo, in B&O’s usual clean luxury style, takes us through from dining to sleep to waking up, opening curtains, making coffee, morning news on TV, and opening the garage door. But it only provides a tightly integrated B&O solution with the need for one or more IT intervention (and it’s expensive — think above $100K for the featured home).

This leaves middle class homes with an unsolved, mixed-vendor Basket of Remotes, a metaphor for the unanswered management challenges in the Consumer IoT space.