BUSINESS ENVIRONMENT
Internet of Things Technology, Segmentation, Players, Economics, and Market Forces We are standing at the edge of the next era of Computing: One that will see the number of internet connected devices increase by a factor of 10 to more than 75bn over the next 5 years. This vast, rapidly expanding network of sensors and devices allows us to capture new information about our physical world, based on the ways in which we and the people around us - interact with them. This information empowers us to generate and surface useful insights that enable us to better understand our world, and how our behavior affects not only our physical and social environment, but also ourselves.
Ramnish Singh and ORS Rao
Introduction The present era of Computing, dominated by the Smartphone was won by companies who delivered the richest and deepest array of new experiences (Apps and Services) on their computing platforms (iOS, Android). This delivered an unprecedented amount of unforeseen value and created powerful new companies like Facebook, Twitter, AirBnB and Uber. The Smartphone’s benefits have also come at a price: human attention. We are getting drowned in information like never before. We check our phones obsessively, 150+ times per day. A recent Ipsos Mori poll of 16,000 people across 16 countries showed the majority of people feel they are constantly looking at screens and wish their lives were simpler (Mori, 2014). We are standing at the edge of the next era of Computing (see Table 1). One that will see the number of internet connected devices increase by a factor of 10 to more than 75 billion Š 2016 IUP. All Rights Reserved.
Internet of Things
49
Internet of Things: Technology, Segmentation, Players, Economics, and Market Forces
Ramnish Singh Director, Microsoft Corporation in US. E-mail: ramnish@gmail.com
ORS Rao Vice-Chancellor, ICFAI University, Jharkhand, India. E-mail: orsrao.icfai@gmail.com
over the next 5 years. This vast, rapidly expanding network of sensors and devices allows us to capture new information about our physical world, based on the ways in which we - and the people around us - interact with them. This information empowers us to generate and surface useful insights that enable us to better understand our world, and how our behavior affects not only our physical and social environment, but also ourselves. Ubiquitous distributed computing will enable the ambient intelligence that powers more personal contextual computing experiences. When we monitor and collect data about the physical world and our reactions to it on a large scale, small events and actions also become more meaningful and interesting. The Jawbone blog post titled “How the Napa Earthquake affected Bay Area Sleepers” is a simple example of this. A vibrant consumer IoT playfield fueled by the maker movement, venture capital (Wakabayashi, 2014), crowdfunding communities e.g. Kickstarter, IndieGogo, Seedrs, Crowd Supply, and a broad network of well-funded IoT accelerator programs world-wide is already shaking things up. Things are happening very very quickly. Gartner predicts that by 2017, 50% of IoT solutions will be from companies less than 3 years old (Rivera, 2015). All of these point solutions will provide new signals that have the potential to make the IoT world smarter, more aware and more useful.
IoT Technology Components At a fundamental level, IoT is central to ambient intelligence. It forms the connection between cloud and the physical world. The simplest scenarios are IoT endpoints driving data from the physical world to the cloud (i.e., alerts, telemetry like the on/off status of a light). The cloud can also provide data back to the endpoints as notifications, or it can control endpoints (e.g. translate a voice command to ‘turn my bedroom light off’). Data 50
April 2016
BUSINESS ENVIRONMENT Table 1: Era of Ubiquitous Distributed Computing
aggregated from heterogeneous data sources combined with the full power of the cloud, deep analytics and machine learning will ultimately enable true ambient intelligence (i.e. predicting my needs, so my bedroom light turns off when I put down my book and change my body position). So while IoT presents massive new opportunities with their own unique set of challenges, requirements and enabling technologies; at its core IoT is a natural evolution technology toward ubiquitous computing and ambient intelligence. Key IoT technology components are discussed below and represented in Figure 1: • Cloud Enablers: Consisting of Infrastructure as a Service (e.g. Storage and Compute) and Platform as a Service (e.g. Event Hub, Data Processing, Analytics and Machine Learning). • 1st Party Experiences: Consisting of the cloud side (e.g. Nest cloud backend, Nexia service) and client side (e.g. the Nest device UI) as well as Natural User Interface (NUI) experiences like speech and gesture control. • Device Enablers: Consisting of Component Software (e.g. Device OS, Firmware, APIs and Client Security) and Component Hardware (e.g. Reference Designs, Chipsets, Radios and Sensors). • Knowledge Graphs: IoT Graph (Physical world), Self Graph (Personal & Social), Org Graph (Office), World Graph (Bing, Google). • Cloud-Device Link: Connectivity such as Networking, Network Protocols, Bandwidth and Local Machine-to-Machine protocols. Device-Cloud enablers such as Identity, Provisioning and Security, and Tools for Configurability and Device Management. Internet of Things
51
Internet of Things: Technology, Segmentation, Players, Economics, and Market Forces Figure 1: IoT Technology Components
IoT Segmentation Considering IoT as a single market, obscures key differences in customer requirements. Conversely, narrowly segmenting the market by industry verticals or workloads uncovers unique factors, opportunities and challenges, but is unwieldy for the purpose of segmentation. For that reason, we consider 3 key segments as shown in Figure 2. • Consumer (e.g. Nest, Dropcam, SmartThings): Consumer IoT scenarios tend to have high-volume potential given that many scenarios are common to 'every household' (e.g. home heating) or ‘every person’ (e.g. wearables). This high-volume potential leads to high competitive intensity. Unlike commercial customers, there is little room for systems integration (no professional installers) and tech skills are generally low (limited ability to DIY), so simple and seamless experiences are at a premium. These dynamics support delivering seamless end-to-end experiences because they address key barriers Figure 2: IoT Segmentation
52
April 2016
BUSINESS ENVIRONMENT
to adoption and because the scenarios have sufficient scale to move the needle on critical mass for platform layers. • Commercial Business-to-Consumer (e.g., retail experiences, insurance profiling, utility pricing, smart cities, patient monitoring): While projected to comprise only 16% of connected devices, from a spend perspective Business-to-Consumer is larger than Consumer. This is fueled by businesses' ability to invest in large scale implementations (supported by System Integrators) in order to compete for consumers by delivering differentiated experiences. While Business-to-Consumer is a large segment opportunity, we believe it will be fragmented and characterized by a very high dependence on other data, specifically Consumer data. Players who are therefore strong in Consumer IoT have leverage to play in key Commercial Business-to-Consumer segments (e.g. Google offering Nest integration to utilities). • Commercial Business-to-Business (e.g., smart factories, field services, fleet management, wind turbines, airplane engines): While the segment is dominated by manufacturing (63% of spend) where solutions can be While Businessextremely niche, it also contains a number of breadth to-Consumer is scenarios like mobile workforce enablement and asset a large segment management. Long enterprise sales cycles and the need for opportunity, we deep engagements for implementation limit competitive believe it will be intensity. A higher percentage of Business-to-Business IoT fragmented scenarios can be justified on cost savings alone (than for say home security or a new retail experiences) which lowers overall barriers to adoption. Given the sophistication of enterprise + SIs, the segment may favor composable offerings to create highly custom solutions over generic plug-and-play devices, applications and services. While Consumer IoT strength translates into Commercial Business-to-Consumer, the benefit of Business-to-Business on Business-to-Consumer is more limited (i.e. even if BMW has adopted Business-to-Business IoT services from say IBM, BMW’s customers may force it to work with Google - even if BMW would prefer not doing so).
Niche IoT Players There are five major IoT players: Google, IBM, Apple, GE and Amazon. While there are other serious players, particularly in Asia - Alibaba, Baidu/Tencent, Samsung, Salesforce.com, Cisco Telcos, Semis - the five are chosen based on their focus on components and segments. These five are chosen to give a good sense for the types of competitive dynamics in the field of IoT, rather than trying to boiling the already vast ocean of IoT. • Google: Google is an early frontrunner in consumer IoT. They have major bets in all areas, including a recent push into 1st party devices. Google’s pattern of acquisitions suggests the $3.2 bn Nest and $550 mn Dropcam acquisitions likely represent just the beginning of their investment in IoT devices. Google’s Nest acquired smart home hub Internet of Things
53
Internet of Things: Technology, Segmentation, Players, Economics, and Market Forces
startup Revolv, which will collaborate with Nest’s open API program called “Works with Nest”to continue to unify the connected home. These targeted buys are complemented by other 1st party experiences in Chromecast and Google Fit, as well as market leadership with Android OS, Android Wear and Android One, enabling over billion mobile devices. • IBM: IBM recognizes the central importance of the data layer and was early to pivot toward IoT, launching Smarter Planet in 2008. Since then, software revenue growth (+30%) has significantly outpaced Services (+5%) and Hardware (-33%) and now accounts for nearly half of IBM’s profit. IBM has invested $17 bn on big data acquisitions and $7 bn cloud and security acquisitions. They have committed 2,000 FTEs to Watson and ~65% of their R&D is focused on data, analytics, and cognitive computing. • Apple: Apple owns leading candidates for the ‘uber-Gateway’ to consumer IoT early adopters (i.e. iPhone/AppleTV). While they have launched a number of ‘IoT platform’ plays like HealthKit, HomeKit and CarPlay, we believe these efforts are squarely aimed at retaining iOS share versus a larger cloud/data focus such as Google’s. • General Electric: General Electric has been an early evangelist making significant investments in Predix and their “Industrial Internet” initiative in a broad effort to transition from selling products to offering ‘equipment-as-a-service’. GE is focused on B2B devices and is generally seen as a very large customer for other parts of the IoT stack, like cloud platform, evidenced in their partnership with Pivotal and Amazon Web Services. Some interpreted GE's recent sale of its Home Appliances division to Electrolux as a signal that it is actively exiting the consumer space to focus on higher margin B2B lines of business. Through Wink - its partnership with New York crowdsourced invention engine Quirky - it is experimenting with a new reach into consumer. Wink is also the control hub for 100 partner products from 15 consumer brands working with Quirky and GE. • Amazon: Amazon views IoT as just another workload for their cloud platform. Similar to what they have done in Infrastructure as a Service, they use their customer traction and usage data to drive their higher level services offerings. On the consumer and devices side, with the exception of Kindle, their efforts, including Fire Phone and FireTV, have yet to gain significant traction. Amazon Echo may signal interest in the home automation space, but the first iteration of the product seems to focus on consumer entertainment and shopping scenarios that are a direct connection to its existing services.
IoT Economics and Market Forces Low Cost End points Greater information and control of the physical world can drive both cost efficiencies (more temp sensors lead to more optimal heating) and increased value (combining soil and water flow sensors with weather data to drive planting, irrigation and harvest dates for grapes). Therefore, addressable demand for more endpoints is only constrained by cost - the lower the cost of endpoints, the more use cases make ROI sense. We have seen this 54
April 2016
BUSINESS ENVIRONMENT
type of unlimited demand elasticity already with RFID tags: as the price of a tag went down, it became economical to move RFID's from a container level to a pallet and eventually to a SKU level (see Figure 3). On the supply-side, we are seeing rapidly declining cost of endpoints. Not just because of the normal decline of chip prices (Moore’s Law), but also due to the ability to have increasingly thin endpoints that rely on the cloud for intelligence (sometimes facilitated by a local gateway, e.g., SmartThings hub) as shown in Figure 4. Figure 3: Demand Elasticity of RFID Tags $1.20
4.5 4.0
$1.00
4.0
$1.00 3.5 3.0
2.9
2.5 $0.60
2.3
$0.57
2.0
2.0
$0.40
1.5 Tag Price Tag Voluem
1.0
$0.20
1.0
0.7
$0.17 $0.10
$0.00 1998
Tag Value
Tag Price
$0.80
2000
2002
2004
2006
2008
2010
201
0.5 0.0 204
Figure 4: Cost Trends Favor Cloud
Internet of Things
55
Internet of Things: Technology, Segmentation, Players, Economics, and Market Forces
Fragmentation of IoT scenarios Consumer IoT will be characterized by a few high-scale, breadth scenarios (i.e., 50M+ customers, such as home security and fitness trackers) and a long tail of lower-volume vertical scenarios like home gardening and blood sugar monitoring. This is observed today in mobile apps (see Figure 5), websites and biz SaaS. This long tail is expected to be even more pronounced in Commercial IoT. Commercial ‘things’ tend to be much more vertical-specific and lower volume. For instance, there are a total of 43,000 jet engines in service worldwide; and the broader category which GE terms ‘Big things that spin’ (incl. industrial motors, pumps, generators, etc.) totals only 3.2M devices. Even non-vertical applications are relatively low scale, e.g., the elevator and escalator market delivers 750k new units per year against a worldwide installed base of 12M. Figure 5: The Long Tail of iPhone Apps
Relative Shift to Cloud There are two important principles here. First, the importance of cloud in IoT will further increase over time. This is driven by cloud economics, which will drive cloud-side costs down faster than device-side costs, enabling ever cheaper, thinner IoT endpoints, and the fact that cloud needs to provide the connective tissue between data from heterogeneous IoT endpoints. This connective tissue aspect will be increasingly critical because, moving scenarios up the value curve toward full ambient intelligence generally requires increasing connections between silo’ed data sources. Let’s take sleep improvement as an example. To do this well, you first need data on sleep quality/patterns from a Fitbit, Jawbone Up, Microsoft Band or specialized sleep sensors like Lark, Beddit, Withings Aura, and the Hello Sense. While the raw data can help identify a problem, more actionable insights 56
April 2016
BUSINESS ENVIRONMENT
come when that feed is paired with daily activity from Nike FuelBand, fitness equipment, Withings weight/heart rate monitor, and accurate environmental sensors like NetAtmo, CubeSensors, Clime, Alima, etc. And corrective actions can only be intelligently automated when that information is connected to control of Nest thermostat to set the optimal temperature at night. Today, this is often done through ad-hoc data peering arrangements. However, these deals are time consuming to put together and therefore don't scale rapidly in scope. IFTTT has emerged as a horizontal ‘logic’ layer that can rapidly expand across new data channels, however, it is limited in the complexity of interactions and deep reasoning it can orchestrate. Privacy concerns are often cited as a barrier to this. However, history has shown time and again that concerns are pushed aside when the value proposition is compelling enough: people use cloud email and accept ads in exchange for a large and free mailbox; people share private data on Facebook because it is the best way to connect with their friends. The bottom line is that the equation is value/privacy. Typically, the winning strategy focuses on delivering high value in exchange for the data sharing rather than using privacy as a marketing tool (like Apple appears to be doing recently). However, being able to better address customers’ privacy/compliance concerns can be a key differentiator (e.g., local data centers working with existing on-prem systems and federated data). The second principle is that the winner in IoT cloud will likely be one or more of the current leading hyper scale clouds (Google, Amazon, Microsoft). This is driven by cloud economies of scale that apply across IoT and non IoT workloads and the importance of combining IoT data with non-IoT cloud data.
Connections Between Customer Segments We outlined 3 key segments. However, these segments are not independent. We saw in section 3 that competitors tend to build positions on either side: IBM, GE on the commercial side, Google and Apple on the consumer side (only Amazon is covering across segments due to its horizontal infra platform offering). Some players may be content to play only in their core segment. However, to address the adjacent segment, both commercial focused vendors and consumer focused vendors bring something to the table for winning the B2C segment. B2B competitors have established relationships with customers and can offer integration between B2B and B2C systems (e.g., helping a utility expose their backend data to their customers). Consumer competitors already own the end-customers with their first party experiences and will own a highly relevant data set (e.g. offering utility companies to integrate with Nest). If all else equal that the scale will tip towards players with consumer strength in this case. If its customers already have a Nest thermostat they are attached to, it will be easier to work with Nest and integrate with their back-end rather than replace their customer's thermostats. Internet of Things
57
Internet of Things: Technology, Segmentation, Players, Economics, and Market Forces
Low Value
High Value
Figure 6: IoT Value Spectrum
1st Party
Cl o u d
Experienc es
Platfor m
Nic he
Dev ice
laa S/ Enablers Low Scale
High Scale
Given the high level of diversity and fragmentation in endpoints as well in IoT scenarios vs. consolidating forces in cloud, ultimately the largest EV creation opportunity lies in owning the data platform that powers IoT scenarios (see Figure 6). Given the connected nature of data and favorable economics of cloud vs. local device computing, this data will be heavily centered in the cloud.
Conclusion IoT represents a huge opportunity and is critical to realizing vision of ambient intelligence. The opportunity is in fact so large, that it is unlikely to see a single winner across all parts. IoT scenarios and devices will be highly diverse. The only way to enable the explosion of (increasingly low cost) devices and the intelligent scenarios, is for a scaled cloud to play an increasingly central role for offloading device data and compute, and in analyzing/ reasoning across a highly diverse IoT dataset. Cloud will experience winner-take-most dynamics, due to both value of data aggregation and economies of scale. Therefore, the largest EV opportunity will be in owning the cloud platform/data layer, other IoT scenarios want to integrate with, as it is both broad and high value. Today’s players tend to either be strong in commercial or consumer, with B2C a largely untapped market. While a more horizontal cloud platform approach may work for B2B, owning ‘finished services’ will be a critical bootstrap (e.g. Nest, Dropcam) in consumer. In in fighting for B2C, all else equal, consumer strength will outweigh B2B strength. Given the vastness of the space and the different skill sets required, major alliances will form with complementary capabilities in order to address the opportunity across the board. Reference # 19M-2016-04-08-01 58
April 2016