At the age of 38, having unrooted myself twice to cross first the Pacific Ocean and then the Atlantic, I've learned to embrace uncertainty and to enjoy unexpected satisfaction more than most people with my training.
To be sure I'm not talking about those glorified travel anecdotes recounted in 2nd-rate travel books or blogs, those about getting lost in the country side of Toscana and bumping into a quiet but brilliant wine maker, against all odds.
On the contrary, I would stress that one is doomed to be disappointed more often than not when they embrace the uncertainty. If it's not the case, religion and philosophy would have been left without any audience long ago. However, it's the rare satisfaction of getting much more out of a limited expectation that encourages us to take the risk time and time again, whether it involves sacrificing a (worthless) day in our life or an arm.
And without a doubt this is my exactly feeling about attending the Connected Conference last week held by Liam Boogar and his widely followed Rude Baguette last week — unexpected volume of inspiration from a first-edition event.
I was only able to attend the Wednesday sessions due to my HEC obligation to the High Fashion Brands & Management course by Professor Rossella Beato that turned out to be just as enriching and inspiring. However, that whole day at the low-key venue of La Halle Freyssinet got my brain running fast way into the night that I arrived at Prof. Beato's class the next morning sleep-deprived.
The undoubtly highlight sessions of the day for me would be the brilliant presentation given by the former GM/VP Europe of Apple and a fellow HEC MBA, Pascal Cagni, as well as the equally great presentation of Barbara Belvisi from Elephants & Ventures.
Barbara's presentation was mainly about this 10-year window of opportunity of funding the hardware startups and is well recapped at this Rude Baguette coverage of the ensuing panel discussion.
I'm gonna, however, talk about some of my thoughts that arised during Mr. Cagni's presentation, which is by the way in the most Apple way with all the eye-catching simplicity and pointedness.
Having had a wonderful ride with Apple from 2000 to 2012, Monsieur Cagni has since left Apple and become an angel investor in London. In the highly compressed presentation he gave at the Connected Conference, he portrayed a wonderful future by 2020 when the some tens of billions of devices in our daily life will be connected to the internet and make our life more intelligent and efficient, creating a market in the order of trillion US dollars.
Just like most of Steve Job's presentations, the well designed and properly targetted Apple-styled presentation was so convincing that one almost wanted to belive that any decent investment in really useful connected devices would no doubt have a healthy expected return. However, having been in the semiconductor industry for 10+ years and witnessing so many distorted — at least from an R&D's point of view — distribution of the value along the chain, I couldn't help but to approach Mr. Cagni after his presentation, introduce myself and pose my grand questions.
Specifically I recounted the rather distorted distribution of value between the semiconductor vendors who seemingly did most of the heavy-lifting and Apple itself in the case of all generations of iPhone, where the necessary standardization of semiconductor technologies is paramount to shooting oneself in the foot while Apple was able to reap the biggest part of the hardware gross margin — a ridiculous 50% historically — through its smart yet proprietary eco-system. Will the same thing happen again in the burgeoning industry of Internet of Things (IoT) or more specifically connected objects?
Casual tech news follower would point to the homeruns in Google's $3.2B acquisition of Nest and subsequently Nest's $555M acquisition of Dropcam. Unfortunately I feel these high valuations exactly blinded entrepreneurs into the good old thinking of "build it and XXX will buy us" — plug in any of the tech giants you like in the blank.
Why is this a problem if the future of connected objects is really so bright, like TRILLION-USD bright?
I would argue that when one looks at any new IoT device and dreams about it being the next big thing, he should really figure out first "who are the buyers?" In other words, is it the end consumers or some service companies the services of which your devices have to be tied to in order to function?
A bit of the history that I'm familiar with to give a bit of perspective on this subject. One of the most unhappy stories in the recent history of semiconductor is ADSL, the first modem technology in the history that really brought that last mile of internet connection to the homes of the users.
To be sure, American users can genuinely argue Cable Modem came at similar time but it suffered the same fate of ADSL that we could basically see them as unfortunate twins.
So let's get on with the ADSL story. Prior to its introduction in the late 1990's, the mainstream internet connection was telephone modem. Users who are old enough or who started early enough must still remember the famous humming tones of telephone modems after the dialing — and maybe the frustrating redialing actions as well. With telephone modems, the end consumers can buy any model they want off the shelves — they all work as long as you have a landline at home and a subscription with your phone service providers (remember AOL?).
With ADSL, this device neutrality was completely annihilated.
ADSL standard is a classical engineering gem in which the great minds in the semiconductor and telecom industry figured out a way to facilitate much higher data transmission over the same old crappy telephone landline. The standard involves a lot of signal processing that was made possible by not just the new coding/decoding theories but more importantly the advance of semiconductor technologies. Among all smart ideas the most genious of all lies in the letter « A » of ADSL — asymmetric. Basically what it means is that instead of asking the client-side device to do a huge chunk of processing in both transmission and reception, the standard sacrifices the transmission speed (upload speed) and focuses on combing the base-station computation power and clever signal conditioning on the client side to facilitate a truly impressive download speed, anywhere from 1~2Mbps in real throughput.
However, because of this combination of so many great ideas, it turns out the the client-side modem has to be in-pair with a counterpart modem at the base station. Interoperability was really not very satisfying that the whole industry quickly converge to a business model: ISPs would hold auctions among modem vendors to decide whom to purchase their current batch of modem pairs from.
An absolute nightmare for both the semiconductor vendors and modem manufacturers of ADSL.
The profitability of the modem manufacturers quickly dropped from the 20% range to below 10% in a matter of 2 or 3 years as more Taiwanese OEMs joined the fight. That of the semiconductor manufacturers held on a bit longer but not by much. By the time ADSL entered the mature growth stage with humongous annual unit volumes, not even the market leader Broadcom could make enough gross margins to justify the R&D investments.
"What the hack does this have to do with IoT?" you ask.
Well, take a seemingly fantastic IoT product, smart door locks, for example. (I won't name product or company names here as it's not really relevant — every maker of such device would face the same issue).
A smart door lock would allow the user to detect whether the your door lock has been opened or breached while you're away, either at work or in vacation. It will then send an alert through any wireless network available to your smartphone, letting you know of such an event.
Extremely useful, no? Who wouldn't want to know if a burglar is working on your lock while you're presenting your quarterly sales number to your manager that is dozing off in the 3-hr meeting?
The question that naturally follows turns out to be the most critical one — now that you know there's a breach, what do you do? Do you go home immediately and face the burglar fearlessly? Or do you call a cop? How long does it take for either of these two reactions to go into effect?
I guess by now some of you already know the point I am trying to make here: assuming that there are some dominant security service providers in a country just like there are only a handful ISPs in the same country, as a smart door lock manufacturer you can only sell to these security companies, who could quickly dispatch two former pro-wrestlers to the location of the alert.
And just like in the ADSL horror story, you will be naturally forced to bid against other smart door lock manufacturers when selling to these security companies. The price will keep dropping through each round of auction until your margin barely covers your flight and hotel fees of attending the auctions.
I hope the message is clear enough here. It's really not that difficult to understand what's at stake here but as an entrepreneur or an investor you really have to visualize how the eco-system of the ultimate service of your product is going to play out.
If that effort is not in place, you might as well buy the book below on your way home from a successful fund-raising because it's very likely you're gonna have to spend more time in sharpening your bidding skills than worrying about the look of your app UI.