Learning by Shipping

products, development, management…

Archive for January 2016

5 Ways to Compete With [Big] Incumbents

Japan, Jiu-Jitsu-KämpferIn The Stack Fallacy: Why Big Companies Keep Failing Anshu Sharma writes about how difficult it is for a [big] company to move up the stack to adjacent businesses/product categories by building on their successful base. If you are competing against one or more incumbents, even if you believe they will ultimately fail because of this fallacy, it is still an incredibly challenging competitive situation. Using some typical weaknesses as your competitive strengths can increase your potential for success when being the next part of the stack a big company takes on.

In a competitive environment, often a “checklist” battle dominates. This is especially true if you are competing with an enterprise incumbent. There are many ways to compete with a company that has more resources, existing customers, and access to broad communications channels. You can be systematic in product choices and communication approaches and increase the overall competitive approach.

You can think of these as the Jiu-Jitsu of the Stack Fallacy — using the reasons competitors can fail as your strengths:

  • Avoid a “tie is a win”
  • Land between offerings or orgs
  • Know about strategy tax
  • Build out depth
  • Create a job-defining solution

This post is mostly from an enterprise competitive perspective, but the consumer and hardware dynamics are very much the same. While some of this might seem a bit cynical, that is only the case if you think about one side of this battle being better than another — in practice this is much more about a culture, context, and operational model than a value judgment.

Avoid a “tie is a win”

The first reaction of an incumbent (after ignoring then insulting the competition) is to build out some response, almost always piggy-backed on an existing product in an effort to score a “tie” with reviews and product experts (in the enterprise this means places like Gartner). The favorite tool is the “partner” or “services” approach, followed by a quick and dirty integration or add-in. Almost never do you see first party engineering work to compete with you, at least not for 12–24 months following “first sighting”.

Their basic idea is to clear the customer objection to missing some feature and then “get back to work”. In enterprise incumbent-speak, “a tie is a win”.

The best way to compete with this behavior is to go head to head with the idea that a checkbox or add-in does away with the need for your service and worse such an implementation approach will almost always be insufficient over time and hamstrung by the need for integration.

Don’t worry about your competitor pointing out the high cost of your solution or the burden of something new being brought into the enterprise. Both of those will become your strengths over time as we will see.

Land between offerings or orgs

The incumbent’s org chart is almost always the strongest ally of a new competitor. The first step is to understand not only where your competitor is building out a response, but the other product groups that are studying your product and getting “worried”. Keep in mind that big companies have a lot of people that can analyze and create worry about potentially competitive products.

You can bet, for example, that if you have any sort of messaging, data storage, data analysis, API, or visualization and compete with the likes of Oracle, Salesforce, Tableau, or other big company that several groups are going to start thinking about how to incorporate your product in their competitive dialog.

You can almost declare success when you hear from your customers that your product has come up in multiple briefings from a single company. Perhaps the biggest loss in a large company is when a Rep loses a deal to a competitor and news of that travels very fast and drives tactical solutions equally fast — tactical because they are often not coordinated across organizations.

When you find yourself in this position, two things work in your favor. First, there’s a good chance you will soon find yourself competing with two “tie is a win” solutions , one from each org— white papers talking about partners who can “fill in the gaps” or add-ins that “do everything you need”, for example. No P&L or organization wants to lose a deal over a competitor.

Second, you will have time to continue to build out depth because the organizations will begin the process of a coordinated response. This just takes a long time.

The best thing that can happen at this point is if you have a product that competes with two larger companies. At that situation you can bet that you are the thing those companies care the least about and what they care the most about is each other. You might find yourself effectively landing between many organizations then and that spot in the middle is your whitespace for product design and development — go for it!

Know about strategy tax

Once an organization grows and becomes successful, one of the key things it needs to do is define a reason for the whole to be greater than the sum of the parts. The standard way incumbents do this is to have some sort of connection, go-to-market, feature, or common thread that runs through all the offerings. This defines the company strategy and the reason why a given product or service is better when it comes from a particular company (and also the reasoning behind a company being in multiple businesses).

In practice, the internal view of these efforts quickly becomes known as astrategy tax. From a competitive perspective these efforts are like gifts in that they make it clear how to compete. For example, your product might have integrated photos but your competitor needs to point customers to another app to deal with photos. Your product might be supported by channel partners but your competitor will only sell direct (or vice versa). This can go to an API level, particularly if you compete with a platform provider who is strategically wedded to a specific platform API.

A classic example for me was the Sony Memory Stick. If you were making any device that used removable storage then you were clearly going to use CF or SD. But there was Sony, marching forcefully onward with Memory Stick. It was superior. It had encryption. It had higher capacity (in theory). At one point after a trade show I left thinking they are going to add Memory Sticks to televisions and phones, and sure enough they did. What an awesome opening if you needed to compete with a Sony product.

A strategy tax can be like a boat anchor for a competitor. Even when a competitor tries to break out of the format, it will likely be half-hearted. Any time you can use that constraint to your advantage you’ll have a unique opportunity.

Build out depth

The enemy of “tie is a win” is product depth. Nothing frustrates an incumbent more than an increasingly deep feature set. Your job is to find the right place to add depth and to push the incumbent beyond what can be done by bolting capabilities into an existing product via add-ins, partners, or third parties.

Depth is your strength because your competitor is focused a checkbox or a tie, figuring out the internal organization dynamics of a response, or strategizing how to break from the corporate strategy. While you might be out-resourced you are also maniacally focused on delivering on a company-defining scenario or approach.

The best approach to building out depth is to remain focused on the core scenario you brought to market in the first place. For example, if you are doing data visualization then you want to have the richest and most varied visualizations. If you have an API then your API should expose more capabilities and your use of the API should show off more opportunities for developers.

There’s a tendency to believe that you need to build out a solution that is broad and to do that early on. The challenge here is that this takes you into the incumbent’s turf where you need to build not only your product but the existing product as well. So early on, push the depth of your service and become extremely good at that — so good that your competitor simply can’t keep up by using superficial means to compete. This example from Slackcrossed my feed today and shows the depth one can go to when there is a clear focus on doing what you do better than anyone else.

Your goal is to expand the checkbox and to move your one line of the checkbox to several lines. This is how you change the “tie is a win” dynamic — with depth and ultimately defining a whole category, rather than one item.

Create a job-defining solution

When building a new product and company, one of the most significant signs of success is when your product becomes so important it is literally someone’s job. Once you become a job then you are in an incredible feedback loop that makes your product better; you have an opportunity to land and expand to other parts of a big company; and you have an advocate who has bet a career on your product.

New products have a magic opportunity to become job-defining. That’s because they enter a customer to solve a specific problem and if that gets solved then you have an advocate but also a hero within the company. Pretty soon everyone is asking that person how they get their job done so much better or more efficiently and your product spreads.

The amazing thing about this dynamic is that it often goes unnoticed because rarely are you replacing entirely something that is already in use, but simply augmenting the tools already in place. In other words, the incumbent simply goes about their business thinking that your product just complements their existing business.

This obviously sounds like a big leap to accomplish, but it speaks to the product management decisions and how you view both the product and customer. With enterprise products it is almost always a two step process. First you solve the specific user’s problem and then you solve the problems the IT team has in using the product as part of a business process (i.e. authentication, encryption, mobile, management).

This works particularly well because your incumbent’s product has already achieved this milestone, and it is their product that (a) is not working and (b) is almost certainly some other function’s job software. It is another way of landing in the whitespace of the organization. Your competitor’s job is not looking for more to do, especially not someone else’s job, so you have some clear road ahead.

The challenge of existing winners breaking into new or adjacent businesses is real and difficult. Very rarely does this happen. The inherent obstacles, both technically and culturally, new products have specific entry points to compete.

Steven Sinofsky (@stevesi)

This post originally appeared on Medium.

# # # # #

Written by Steven Sinofsky

January 29, 2016 at 9:44 am

Posted in posts

Tagged with , ,


everlaw-600x600-squareAJ Shankar was busy working on his PhD thesis at the University of California, Berkeley in the prestigious Programming Systems Lab, where he published a number of important papers in OOPSLA and PLDI. As a big fan of side projects, he also caught the maker bug.

One of those projects was working as a technical expert for a leading Seattle-based law firm. It led AJ to ask what every entrepreneur asks, “How can this be improved with software?” That’s where Everlaw got started, in 2011.

The problem: A world of legacy software for lawyers

The legal profession — particularly the area of litigation and trials — is a costly, complex, labor-intensive, and, frankly, error-prone process. Beyond that, it is steeped in the complexities of individual courts and jurisdictions dictating, sometimes at the trial level, how technology can be used. Having personally worked through the transition from WordPerfect to Windows over the better part of a decade, I know the challenges of bringing technology to this highly knowledge- and people-intensive process are significant.

AJ and his co-founder, practicing lawyer Jeff Friedman, a former Assistant U.S. Attorney and corporate counsel, know these challenges well from their experiences. They set out to invent something that meets both the demanding technical needs of litigation along with the unique business requirements of law firms, which often do not have the resources or skills required to manage complex software deployments.

In fact, complex deployments of on-premises software defines the current state-of-the-art in litigation support software. Anyone familiar with modern software would look at this “state-of-the-art” and see architecture from another era. That’s not to say those solutions do not provide value and make money, but AJ and Jeff see a far better way.

There is also a need for modern solutions to deeply technical problems — such as searching terabyte corpora for relevant documents (the state-of-the-art is mostly keyword search) or identifying clusters of relevant documents based on machine learning techniques (versus relying on humans to manually sift through and connect millions of documents). Historically, an industry vertical with such a legacy business model and architecture (i.e., very slow to change) would have a very hard time attracting top computer science talent to improve the space.

Law firms also need software to solve the modern problem of “big data.” In this context, big data can mean millions of email messages, chat transcripts, voice mail recordings, scanned documents, entire data sets and social media feeds, and much more. Those are the artifacts of the legal discovery process that flow across both sides of the aisle in ever-increasing volumes. These volumes are beyond what many law firms can deal with, and as some might know, producing large amounts of data can often be part of a legal strategy used against smaller firms.

Finally, the pace of change for software in the litigation industry needs to increase. The model of one-time, slowly updated on-premises software simply isn’t compatible with the fast-paced changes in technologies that can help legal. Part of the legacy world the legal profession faces is the same that any enterprise faces: A desire to move away from high, up-front product costs and transition to a cloud and software-as-a-service (SaaS) model.

The solution: Bringing cloud innovation to lawyers

Everlaw architected a solution that starts from customers: attorneys at small firms, large firms, state offices, and on both the defense and plaintiff side of cases. AJ’s technical background and Jeff’s real-world experience as an attorney proved to be a great place to start. To begin the journey, Everlaw assembled an engineering team of hard-core computer scientists, many from UC Berkeley.

In the Andreessen Horowitz pitch meeting, it turns out a lot of the former CEOs, execs, and founders have been involved in litigation. Our collective experience, especially as defendants, led to an immediate bond with AJ as he detailed the Everlaw solution. Many of us have been through the boxes of documents and questions from counsel about “discovered” documents. We knew how difficult the process was and we loved when AJ detailed Everlaw’s approach:

  1. Bring together core computer science experts from natural language, machine learning, and full-stack development to architect the system.
  2. Build innovative experiences that start with the process of ediscovery and provide a platform for an end-to-end solution for attorneys to collaborate as a case is developed.
  3. Deliver Everlaw as an incredibly secure, highly reliable, totally scaleable cloud-based SaaS service.

So far, their experience with customers has been amazing. Since most attorneys are part of the world of mobile and cloud experiences, as soon as they see Everlaw, they see how much easier, faster, and higher quality their trial preparation and work can be. In fact, customers usually say “why did it take so long” or “this is how it should work”. AJ has written a postthat includes more details on the company’s vision and the success to date.

At Andreessen Horowitz, we are always incredibly excited to see technology founders taking on the hard work of reimagining an industry. It is clear that mobile, machine learning, and cloud delivered via SaaS will revolutionize every vertical, including legal. We love the work that AJ, Jeff and the Everlaw team have done to bring such high-powered efforts to an incredibly important part of the economy.

For those reasons, we could not be more excited to be partnering with Everlaw and leading their Series A funding round, joining the existing investors. I am super excited to be joining the Everlaw board to support their ongoing work. Software eats legal.

Steven Sinofsky (@stevesi)

This post originally appeared on a16z.com

Written by Steven Sinofsky

January 14, 2016 at 10:23 am

Posted in a16z, posts

Tagged with

CES 2016—Observations for Product People

CES 2016—Observations for Product People


This is not me.

CES is the best place to go to see and learn about making products. In one place you can see the technology ingredients available to product makers along with how those ingredients are being put together and how they are interacted with and connected to customers.

I love going to CES and walking the show floor north to south, convention center to Sands and seeing and touching the products, including the way random show-goers perceive and question what is out there.

As much as I love attending, I also love taking a step back and thinking (and writing) about what I learned. Doing so provides great context for me in working with startups on their products, talking with enterprise customers about their needs, and partnering with bigger companies to enhance their go to markets.

As a reminder, CES is not a big electronics store nor is it a research lab. It is somewhere in between. While there are many ready-buy products on display, most are not yet ready to use. Many of the most interesting technologies are not yet in products. Most companies are working to put forth their best vision for where things are heading. It has always worked best for me to think about the show directionally and not as a post-holiday shopping excursion. Equally important is keeping in mind that I’m not the customer for everything one can see.

This is a long post. The breadth of CES is unprecedented. The show is not “consumer electronics” or even “home entertainment” but it every industry. Where else would you see booths from car companies, delivery services, film studios, computer makers, electronics component makers, cable tv companies, mobile phone carriers, micro processor and chip makers, home improvement superstores, and so on. From startups to mega-caps, from every country, from supply chain components to complete products everything is represented. The opportunity is unique.

CES has become a software show. Even the interesting hardware is dominated by firmware, cloud services, and connectivity. It is increasingly clear that if you’re interested in software you have to be interested in pretty much every booth. I’ve heard software is eating the world and that’s on display in Las Vegas.

The major observations impacting product makers and technology decision makers on display at CES 2016 include:

  • Invisible finally making a clear showing (almost)
  • Capable infrastructure is clearly functional (almost)
  • Residential working now, but expectations high and software not there
  • Wearable computing focusing on fitness
  • Flyable is taking off
  • Drivable is the battle between incremental and leapfrog
  • Screens keep getting better
  • Image capture is ubiquitous
  • Small computers better and cheaper for everyone
  • Big computers better but not game changing

Invisible finally making a clear showing (almost)

For many years much of the show floor was dedicated to the problems of where to store bytes, how to move those bytes around a network, how to type, or even how to convert bytes from one format or device to another. What’s most amazing is just how much of all of this is now simply invisible. The whole industry has moved up the stack.

If you go through all the winners of CES “best of show” (note, wow there are a lot of winners!) most all of them have a few things in common:

  • No local storage (for customers to deal with)—everything is cached from the cloud or streamed (i.e. no media servers, no hard drives, no formats to worry about, no backups to do). Yay!
  • No wires—everything is wireless. Even better, most everything is WWAN (mobile phones) Wi-Fi or Bluetooth. This is infrastructure that is now normal—meaning not a point of differentiation or confusion—the mobile ecosystem and supply chain all but guarantee this connectivity and capability. Almost nothing has an RJ-45 network jack and anything that might require one has some sort of wire-closet hub to separate the actual device from the wired connection. Most everything easily tunnels to your Smartphone via cloud services. Yay!
  • No buttons—everything has a touch screen and there are few buttons to deal with. When a complex user experience is needed, it is almost always done with a mobile app (more on that below). What was amazing was just how rare it was to even see a keyboard and certainly gone are rows of rectangular buttons. Yay!
  • Almost, no mains–a lot of focus is going to long life batteries, solar, and certainly wireless charging. Many of the winners such a Bluetooth location devices, cameras, and home automation/security operate on batteries lasting almost a year to two years. That’s long enough to probably never change the battery and just replace with the next generation! Put devices where you want and access them from anywhere. There’s a massive amount of cool engineering and clever approaches that go into being ultra low power. Yay!

This set of attributes represents the starting point for most any product. It is also a huge opportunity for consumers because it means the ability to adjust devices over time, even for residential equipment, is much easier than the past. Imagine when you move, you can just relocate your security camera, for example.

To be fair, there are some wires, but we are down to three: Apple lightning, USB C, and HDMI. USB was ubiquitous throughout the show and devices that should use USB C (like new PCs) and didn’t look like they missed out. Given wireless video casting, even HDMI cables will fade into the background for most people. I’m beginning to think more and more that Apple Lightning is looking more like Firewire in that it was superior at the time but the industry caught up faster than expected. It might even be the case that HDMI will move to the USB C connector form factor (not protocol). Going to/from these three cables is also easy, which is great.


USB C was everywhere.

One fun note is that quite often you see a product that seems clever and/or odd and then you see it again, and again. This year, I saw the identical USB charge station a dozen times. This is the China manufacturing and distribution system at work.

USB charge station seen all over the floor.

USB charge station for when you’re really serious about charging (C version on the way!)

Capable infrastructure is clearly functional (almost)

It is interesting to see the mobile supply chain’s relentless focus on continued integration drive very capable infrastructure into nearly every single device.

Going back, there would have been a CES where “wireless music” was a thing all by itself. Or maybe you recall a CES where just being able to have a camera was a big deal. Most probably remember when GPS was a “thing”.

CES 2016 shows that all of these scenarios have come together and basically in anything you want to make one can have all of these (and more) or pick and choose easily what capabilities to expose. From a base capability perspective this includes:

  • Attaching a camera and sharing captured images/video
  • Streaming audio
  • Controlling the power state and move it around
  • Locating the device
  • Alerting those nearby with sound, vibration or those far away with mobile alerts
  • Lighting the device with tiny LEDs of any color that never burn out and consumer little power
  • Uploading sensor data from the device
  • Sensing the environment

All of these are available to product makers and likely harder and more expensive to acquire discretely than they would be by essentially taking a mobile phone BOM and making a device. If you talk to the makers at the booths, most every device has more capabilities in hardware than is being exposed in the current release of software. Cameras are capable of 4K, SIM slots go unused, sensors collect but don’t share, and so on.

The big challenge is no surprise. Software development is unable to keep up with the hardware. What is going to separate one device from another or one company from another will be the software execution, not just the choice of chipset or specs for a peripheral/sensor. It would be hard to overstate the clear opportunity to build winning products using stronger software relative to competitors. Said another way, spending too many cycles on hardware pits you against the supply chain for most products.

Some of the devices that include most of these include a rubber duck (speaker, remote control), knit cap (music player), light bulb (speaker, camera, climate), walkie-talkies (location, camera), power strip (remote control, telemetry, power usage report), flower pot (soil water level, camera). The list goes on and on!

Screen Shot 2016-01-10 at 11.02.24 AM
Looks like a rubber duck, but it is also a remote controlled streaming media player with kids apps!
Looks like just a knit hat, but a music hat that streams music.

Residential working now, but expectations high and software not there

The most visible example of the ecosystem of components, manufacturers, and distribution coming together is in residential—products to control, protect, and monitor the home. There were dozens of companies showing what looks to be essentially the same product:

  • Wi-Fi or WWAN base station that connects to and controls the sensors in the home while also communicating with a monitoring service
  • Door/window open/close sensors to detect entry
  • Water sensor to detect floods
  • Motion sensor to detect intruders
  • Outlets and switches to control lighting and outlets
  • Smoke/Fire/CO sensors for safety
  • Thermostat for environmental
  • And so on.

In addition, there are more specialized (and harder to make) controllers for legacy home systems like garage door remotes, water heater, sprinkler, and so on.

Plus there are cameras for security monitoring and doorbells and locks to control entry, though many systems struggle with offering and integrating those.

The reality is that all of these basically just work and provide evidence of the supply chain at work. These are offered by startups, white labeled to many local distributors who will handle installation, and all the major home improvement stores carry them. You might have even seen the pitch from Comcast or AT&T for these as well. There were at least a dozen full service companies on the floor.

They are all essentially the same offering. Well, except for for the software and that is where they are all quite different and where the “ready for prime time” evaluation needs to be done. While they all have apps, for many scenarios some of these can prove quite awkward for some basic control—to the point where it is more annoying than helpful to use. For most customers, the app becomes secondary to more traditional keyfobs/dongles and PIN codes.

Once again, this shows where there’s opportunity to focus and potentially win.

Traditionally this has been an area where the reviews clamor for integration and synergy across device. A couple of things became clear this year:

  • Since everyone can offer everything (due to the supply chain) the viability of the company becomes more important than worrying if the company will offer a particular sensor/controller.
  • Integration is happening through a very traditional “consortium” and as nice as this sounds it isn’t clear it is working particularly well. First, much of what makes these easy to use is the way each maker handles out of box setup (which is mostly outside the standard) and adding additional sensors over time. Second, the UX for managing sensors and controllers integrated by third parties is usually least-common-denominator compared to first party.

In fact, this year saw a significant change in integration. Last year most all home automation was integrated with Nest. While that is still the case, as most would not the integration provided little useful capabilities and the “native” apps proved better. This year everyone integrated with Alexa from Amazon Echo. This made for compelling demos to turn lights on/off or adjust temperature. Time will tell if Alexa will be replaced next year or if Nest will up the level of integration.

IFTTT (an a16z portfolio company) was frequently used in demonstrations for conditional and multi-step scenarios. IFTTT replaces “custom installer” macros and other tools that have often plagued “home automation”.

Two great examples of this are programmable door locks and video doorbells. Both of these are logical integrations for the rest of a security system and while basic integration over z-wave is possible, for most scenarios (answering the door, programming new combinations) the vendor specific app is required. These are difficult to make products that need to fit in legacy infrastructure, so this is to be expected.

That said, because of the rising tide of infrastructure, the locks and doorbells have come a long way in the past year. Ring doorbell even released a battery operated (rechargeable) camera to accompany the doorbell (it is basically the motion-activated doorbell camera without the bell). Vivint has done good work to integrate Kwikset locks, a first party doorbell, as well as Amazon Echo to provide a more complete solution.

But for now, the base level capabilities are there and work across many providers. It is likely that these will further coalesce into a market where it is easier and better to get all the components from one company rather than trying to stitch them together. The good news is that this category is a pretty simple DIY project. The better news is that because of the SaaS revenue for monitoring, it is not hard to find an offering that comes with free installation (such as from Comcast).

Home integration happens with this in theory, though in practice the supply chain makes it easier to avoid cross-manufacturer integration if at all possible.

Example of one of many suppliers offering the full range of sensors and controllers.
Even at the low end, all the same sensors, detectors, cameras are available. 
Most cameras now combine motion detection and some machine learning to reduce false alarms. This camera is integrated into a traditional porch/doorway light so no extra wiring is needed.
In an example of the ecosystem at work, this same switch (based on the no-battery, no wires approach of enOcean) shows up in dozens of different systems.

Wearable computing focusing on fitness

The big news last year was all about “smart watches”. This year the focus of many of the same makers turned more to fitness and less about overall lifestyle.

There were certainly many connected measurement devices (body composition, weight, sugar levels, blood pressure, etc.) and every device is able to measure sleep (on your wrist, in your pillow, or in your mattress) or steps taken.

Unlike the home security sensors, there’s still a great deal of science to be done to correctly (accurately, precisely, reliably) measure humans and much science that is needed to make this information actionable. I continue to think we’re measuring more than we can consume and act on, especially on a constant basis.

It looks like the major band makers agree and this year became much more focused on the specifics of exercise. The biggest announcement came from Fitbit with the new Blaze wrist wearable, “smart fitness watch”.

Fitbit Blaze wrist wearable for fitness

In addition, Polar, Garmin, Under Armour and more all had new/improved bands dedicated to fitness. Much of the technology is about adapting algorithms to understand what the telemetry means depending on the sport (i.e. how do you measure fitness goals from your wrist when doing weights).

My view is that these bands are doing amazing work for people that are hardcore training in sports, but that the vast majority of people won’t benefit from the charts and graphs that come from doing a lot of work to set up using these. Speaking purely from the point of view of improving the average person’s fitness, a scale and blood pressure monitor seem more important. For most people, just walking for a fixed amount of time would be an improvement and a watch focused on timing laps for multiple sports is unnecessarily complicated and potentially demotivating. The support that comes from the community aspect for basic measurements and activities is documented and well-known to be a benefit, but that wasn’t the focus of the products on the floor.

The other aspect where these bands both differentiate and are still searching for broad fit is in software. With some sports sharing times (rides, trails, etc.) is a part of the hardcore enthusiast and so the community aspect is important. Again though that isn’t necessarily a mass consumer scenario.

I’m certain that the medical physiology (what measurements mean), sensor technology (how to measure), and medical research (how to act) will continue to evolve in this space. The longer term goal of a device that tracks meaningful body telemetry that regular people can act on themselves is not far off.

Fitness monitoring is not unique to humans. There were a number of products to help to monitor your pet with a pet wearable.

Connected Pet—monitor what your pet does during the day for better fitness.

Flyable is taking off

Drones were more numerous and more capable than last year. As much as the category is maturing, it is worth noting how early this really is.

There are two large players in Parrot and DJI who commanded a significant presence on the floor. Beyond that, once again we can see the supply chain at work as there were countless companies with largely similar products.

The most common experience in the drone booths would be to watch someone come up to a company rep and ask about the range and then follow up asking about the payload. I must have seen this 20 times and each time the person walked away disappointed, as if they where hoping this was the magic booth that had the drone that really could deliver groceries or fly cross-country. The other question was how autonomous the drone was and always the answer was disappointing.

The vast majority of what is going on is still in the realm of traditional radio controlled (RC) flight in new form factors with amazing cameras (made possible by the influence of the smartphone supply chain). Even the major vendors are still in the early stages of the basics of geofencing, route planning, and other scenarios focused on safety.

There’s clearly a product development cycle analogous to PCs v mainframes/minis happening. Drones are never going to be jets or general aviation, just as PCs were never going to be mainframes. When something sees a 10x increase in usage/adoption the new product is always much different at that scale.

On the other hand, things will not evolve so fast and loose the way PCs did because drones share the same airspace as jets (in a way that PCs never shared with mainframes). That’s why I think this evolution will see more “real” aviation get pulled into drones much sooner than we saw mainframes (i.e. servers and server hardware attributes) pulled down into PCs. Reliability, safety, and more will need to happen sooner rather than later. Piloting a drone will be a profession, not a hobby, until they can really pilot themselves (but even then…). With that there will be opportunities.

Like other categories, the difference between companies is not as much in the supply chain components or even the manufacturing/integration but in the software platforms. In the case of drones, it is more the minimal amount of software. There’s still a lot of “pro-sumer” work that needs to happen to get the full cycle of sensors, flight, data gathering all working. One example of this at work was Parrot demonstrating their work with senseFly (a Parrot company) for agriculture.

Another example was this complete “police surveillance operation” kit from Flymotion. It offered the full command center for monitoring in the case of disaster or other need.

IMG_0884Flymotion’s complete Police surveillance drone system.

One of the most crazy and unexpected drones was from EHANG, a China based company (co-founded by an ex-Microsoftie!). Their product is a single passenger drone — basically an autonomous Uber-drone. You get in it and it flies you somewhere. Totally mind blowing. Given the differences in regulatory climates, this product is making fast progress in China and is already airworthy. I don’t often post pictures of me, but here I am to give you a sense of scale of this one.

Here I am exiting after checking out the single passenger EHANG drone.
Another image of EHANG’s drone from their web site.

Next year is going to be an incredibly interesting year for drones. That is certain!

Drivable is the battle between incremental and leapfrog

Back down on earth and on the roads, the biggest battle in the global economy is over the next generation of “car” transport. Given the size of the market and the importance car companies played in the 20th century it is obvious why so much focus is on self driving cars or on alternative powered cars (or both at the same time).

All this coverage needs to be put in context of what was on display at CES. First, it is remarkable that car companies are using CES as a platform for announcing their autonomous work and general innovation in driving—while autos (and the Detroit supply chain) have been at CES for years it was always in the context of the after-market accessories or in building better premium “electronics”.

Second, while the whole North Hall of the convention center is devoted to cars, the vast majority of what is on display is traditional after-market customizations and even standard cars. FCA’s center stage was an interesting revamp of a Jeep interior, independent of autonomy or alternative power, for example.

The most interesting topic to ponder is really the nature of disruption that is taking place. Existing auto companies are seeing every aspect of their business upended. On the one hand all of their expertise in engines, interior design, drive trains is called into question by electric cars. On the other hand, autonomous driving challenges the fundamental business model of these car makers. Together these disrupt the entire process cars are built—a supply chain of parts makers, product managers, brand managers, dealer franchises, and more that has been built up over 100 years.

It is one thing for GM to show a Bolt, which by all accounts looks amazing. Or similarly for VW to show the BUDD-e van ( electric range of 373 miles, be capable of recharging to 80% capacity in about 15 minutes and would have a top speed of 93 m.p.h). But it will be quite another to deliver these at scale, sell them, and change the pricing and business models along the way. That’s just super hard for any company to do. As a reminder, FCA, Ford, GM combined sell light trucks for about 72% of their North America vehicles and those are more of their profits. Here’s a fascinating article on GM and the change underway there.

VW BUDD e electric van to be available in a couple of years.

The role software and hardware (again, the smartphone supply chain) will play and how companies execute those areas will almost certainly be determining factors. For example, it will be much more difficult to built a reliable car if the software and hardware systems are a combination of legacy and new; or if every car needs to be built to handle the “optional” autonomous or driver assist features. Will the car makers look to the existing supply chains in place or be able to make huge and difficult choices to trust new suppliers with new components?

An example of this is NVIDIA which is building out a significant and integrated suite of car electronics. Basically making a car SoC. NVIDIA is not Bosch or Delphi.

NVIDIAs car “SoC”.

While we were at CES Tesla updated their customers vehicles with the ability to summon your car. In a world where car makers still mail out DVDs or USB sticks to update maps, it is interesting to think about how things need to change inside those companies to enable that sort of customer experience.

If you think all of this is just being pro-Valley or cynical, then I would offer this counter example. Mercedes’ announcement at CES was that they intend to announce by the end of the year their strategy for electric cars. So in a year they will announce what they intend to do (of course many people are working on that now). The clear focus is on driver assist leading to autonomy (which they might be very advanced in).

For me, the most exciting transportation product was the Gogoro SmartScooter, which was also at the show last year. Think of the product as an electric Vespa with a max speed of 60mph and a range of about 60 miles at 40mph. But you don’t recharge it while parked, you pop out the battery and pop in a new one (or two) at one of many battery stations around town. You can own the scooter or potentially share them the way Divvy bikes are shared in major cities in the US. The company also has a home station to charge batteries in two hours.

This feels like a potential future of urban transport in most moderate climates.



Gogoro SmartScooter and public charge station.


Gogoro Energy Network showing charge stations in Taiwan.

Screens keep getting better

It used to be that the big (or flat and big) news at CES was about TV. Booths used to be filled with TVs. TVs are important but this year saw a greatly reduced push around smart TVs and a much bigger emphasis on overall image quality.

The reason for this is HDR and 4K. While most people gravitate to 4K (which debuted two years ago and is widely available now, including streaming content) the real news is HDR. HDR is “high dynamic range” or the ability to show more range of brightness. If you imagine scenes from Jessica Jones or Daredevil, HDR makes those scenes so much better, much more like what you would see in person. Unlike more pixels which we all know most people and most rooms can no longer discern, HDR is immediately visible to most viewers. Here’s a great thread on Stack Exchange about dynamic range.

Standard 4K image on left, HDR image on right.

All the major companies were showing off HDR displays. There’s a new industry acronym Ultra HD Premium which signifies an appropriate level of dynamic range. Netflix and other content providers will also be supporting HDR.

Take a moment to consider why this is not like the transition to HD and why it will happen much faster. HD required new content and going back to existing libraries of film and rescanning to make Bluray which you then needed to buy to play in your Bluray player. Network TV had to make the transition. Broadcast spectrum had to be allocated, and so on. Now this is all about software—recording is captured in RAW which has the information to make HDR (though more can be done in sensors for sure, which is a huge opportunity!) and re-encoding with enhanced metadata can be done as desired. Even distribution is no longer focused on just studios with new content coming from new players who have software perspective to bring.

Dolby is doing very exciting work to bring HDR to theaters and to home screens. They also showed some incredible work on sound called ATMOS which is a sound encoding that allows a single speaker bar to use a large number of drivers to deliver 7.1 sound. It was incredibly cool to sit there and hear sound coming from everywhere (Mad Max!) from one sound bar from Yamaha on the super cool LG HDR OLED display.

Still TVs continue to just get better. OLED continues to amaze and seeing a TV that is a sheet of glass is the star of the show. The Samsung SUHD 8K was the one to watch this year.

Samsung 8K HDR Quantum Dot display. Yowza!

In the magic of software and physics department, Sony was showing short throw laser projectors that were mind blowing. One was a 40″ image projected from a 4″ cube speaker essentially against the wall. The other was a 100″ image from about 12″ away. Amazing! (Super interesting how the digital sensor captured the image by the way — some insights into how the lasers work!)

Sony short throw projector. Image is about 40″ projecting from the cube speaker essentially next to the wall.
IMG_0966Sony short throw project. Image is about 100″ projecting from the floor console about 12″ from the wall.

Image capture is ubiquitous

Cameras are everywhere in products. Once again this is enabled by the supply chain create by the pull of smartphones. Incredibly high quality cameras can be integrated into very small places and draw very little power. If everything is connected, then you don’t even need to store images or have an interface to interact with them.

Cameras are gaining more resolution, working better in lower light or infrared, and offering new capabilities driven by software. In particular, motion sensing, face detection, and object recognition software capabilities are becoming key parts of cameras. Though cameras themselves as a stand-alone product are much less interesting than integrating them into environmental or people monitoring, smartphones, cars, doorbells, or industry/job specific functions (police cameras for example). As with home security, the supply chain makes it easy to have the camera, but software is what makes it useful.

A great example of this at work is the Blink camera. By using motion detection software and bluetooth LE this camera becomes completely wireless—it uses CR123A batteries that can last 6–12 months. It is like a completely wireless Dropcam (but not one you would look at all the time unless you wanted to change batteries). Netgear Arlo is a camera taking a similar approach. These cameras communicate over Bluetooth to a small powered base station that connects to a wired network

Blink camera operates completely wirelessly using batteries and Bluetooth LE.

In a dynamic similar to smart watches, the action cameras seemed to struggle this year. There was no pulling back from extreme sports or just extreme in general as the main purpose for GoPro, for example. It isn’t clear to me how much bigger this market can be. There were a vast number of GoPro-like clones out there.

There were quite a few “VR” cameras which were any number of cameras (depending on lens) designed to capture 360°. The playback would use Google Cardboard. A good example is the Nikon KeyMission 360 which captured 4K images.

Kodak’s Super 8 is a high tech Super 8 movie film camera. It is quite the hipster product. It shoots film, which comes in classic Super 8 cartridges that you ship back to Kodak where they are developed and then delivered as scanned footage. As a silver-halide-enthusiast, I find it very neat but I am a bit skeptical. What might have been more interesting it to build a camera that had the UX affordances of shooting with film but the convenience of digital (along the lines of a Nikon Df still camera).

Kodak’s Super 8 film camera and trusty Tri-X film!

One of the additions to CES this year was dedicated space for crowdsourced ideas/companies such as from Kickstarter. One project in this part of the show was the Enlaps camera. In one package the full range of the supply chain comes together — cameras, solar, mobile data, and cloud services. The product+service is incredibly cool and solves what is traditionally a very difficult problem which is capturing long term, time lapsed video from a remote location. While even phones have interval image capture, the ability to manage power, control the camera, and monitor what is going on is enormously complex (see http://ohioline.osu.edu/w-fact/pdf/0021.pdf for wildlife capture which is like time lapsed by motion triggered). I love the “set it and forget it” Enlaps product. It has two 4K cameras, solar power, and a web service that handles the complexity of time lapsed intervals so you can easily stream the results to your phone. You could use this for sunrise/sunset, changing seasons, tidal pools, wildlife migration, construction projects, crowd flow, events, and more.

Enlaps.io is a fully self-contained interval camera. It shoots 4K images at intervals and sends them back to your mobile device. The camera operates off solar power and can be placed remotely for as long as you need.

Once again our pets get some love from cameras too. In this case your pet can “Facetime” with you by pushing a very Pavlovian button. Of course it isn’t enough to just see your pet. With PetChatz you can release pet pleasing smells and treats using your mobile device.

PetChatz is video conferencing and treat dispensing for your pet. No, really it is.

Small computers better and cheaper for everyone

There is not a lot of news at CES for small computers as most companies save that news for GSM World. What is there shows the continued ability for the mobile supply chain to deliver all the components for a small computer and to now package them in ever-improving quality packages at ever-decreasing prices.

All of the vendor phones displayed the floor were of course Android. It is worth noting that one almost never saw Android being shown as part of products in booths unless the product is doing something that it probably shouldn’t be doing (i.e. root kit, peripheral, access to some low-level OS thing). The common thing I heard in the booth booths with Android would be “we’d like to do this but Apple won’t let us”. Personally, this is less of a call for Apple to open things up and more of a call for developers to think up different solutions, at least that’s my view of prioritizing “consumer electronics quality” over “get stuff done”. Most of the scenarios that were Android-only seemed somewhat dubious to me.

That said, there were dozens of phone makers with very high quality builds of phones. This one from nuu mobile is part of a line that goes from $99-$299 direct to consumer. There are lots of these companies differentiating mostly by channel approach (country, carrier, unlocked, rate plans) and less and less by software I think.

Top of the line nuu mobile Z8 retailing for US$299.

At the extreme low end, some of the China manufacturers still show some pretty old school stuff. I just liked how this ODM model had the generic “Brand” on it waiting to be picked up by a wholesaler.

Old school phone labeled “Brand” waiting to be picked up by a wholesaler.

And in a throwback to the smaller-is-better era, here is a full voice/text phone done as an earbud. Those are actual buttons. No word on talk time. I actually saw it work!

Full ear bud phone. No really.

Big computers better but not game changing

There were quite a few new Windows 10 laptops, all-in-ones, and big tablets announced this year (most with Spring availability).

The general trend is thinner, higher resolution screens, and Intel’s new Skylake processors. The Samsung 9 and the LG Gram garnered a lot of attention. The Samsung comes very close to the Macbook in form factor in a larger screen. As a Windows PC it has more ports of course. The LG is crazy light as you can see in the photo below. They were not quoting any battery life and there is no touch screen. Both skipped using USB C for power though which is disappointing in terms of specs.

LG Gram Windows 10 PC is super light.
Samsung 9 is a little thicker and heavier but features a touch screen.

HP, Dell, and Asus also had new PCs.

The area where PCs still currently lead phones is in graphics capabilities. But you can only experience this if you use the massive discrete cards from NVIDIA (primarily) and not with the integrated graphics on every laptop. If you want these insanely powerful graphics capabilities (say for deep learning experiments, bitcoin mining, CAD/3D, or just gaming) you have been stuck with a pretty hard to deal with tower. There’s help on the way in two neat PCs.

One is the MSI 27 XT all in one which takes a classic 27″ AIO and bolts on sort of a backpack for a PCIe graphics card. It isn’t pretty but it is a much more viable way to get the power you need assuming the display is good (which I was not able to see).

 Screen Shot 2016-01-10 at 11.10.16 AM
This MSI 27″ All-In-One has a discrete graphics card cage on the back for a PCIe card.

Razer which builds a great community of PCs and accessories offered up a pretty unique combination. The Razer Blade is a high end ultrabook stylized for gamers (colored LED keyboard lights). It runs high end Intel Skylake parts (Core™ i7–6500U) and has a great screen. It would be an accomplished Ultrabook on its own.

Via Thunderbolt 3 in a USB C form factor you can attach a mid-tower sized box with an external PCIe graphics card (as well as some additional ports). This turn the Ultrabook into a pretty high end workstation. I admit to taking a wait and see regarding the quality over time and security of all those kernel mode drivers via plug-and-play of thunderbolt so I’m looking forward to seeing how this evolves in real world settings.

Razer Blade ultrabook and Thunderbolt 3 connected PCIe desktop graphics Core accessory.

Finally, in the tablet form factor Samsung announced the TabS Pro for Windows 10. The most interesting thing is that it carries integrated LTE which you don’t see often. The tablet itself is a 12″ slab with a great sAMOLED display. It runs the updated Core M processor which runs everything anyone would need to run unless you’re running Visual Studio or full time CAD/CS. The specs are great. In practice the soft, z-fold cover is awkward (just watching the booth folks deal with it), doesn’t stay attached, and doesn’t support the 12″ screen while using touch.

Samsung TabPro S


It is tough to beat this story of entrepreneurial spirit. Meet 13 year-old Taylor Rosenthal. He’s an entrepreneur from Opelika, Alabama. As an avid team sports participant he has more than once run across the challenge of needing the right first aid gear for minor cuts and scrapes on the playing field. He developed a set of kits and a vending machine called RECMED. He made his way to CES to show off his company, which he told me is remaining independent even though he already received a significant buyout offer!

Way to go Taylor!

1-uh7jteCOrxrUoJaPmfktpw Taylor Rosenthal, CEO and Founder of Recmed.

Steven Sinofsky (@stevesi)

Written by Steven Sinofsky

January 11, 2016 at 10:59 am

Posted in posts

Tagged with , ,

%d bloggers like this: