The Machina Research IoT Forecasts Research Stream comprises access to the IoT Forecast Database and 58 Application Spotlight Reports, which between them provide a comprehensive guide to the global IoT market opportunity. We start with 8 major Sectors: Car, Cities, Health, Industry, Home, Business, Energy, and Consumer Electronics.
Within each of these sectors there is a diverse set of devices, applications and services. As a result we base our analysis on individual applications, of which there are 200 covering all of IoT. We then roll these applications up into application groups, of which there are 58 (see below).
The Forecast Database presents Machina Research’s quantitative view of how the IoT market will evolve.
It is almost impossible to give a definitive methodology for the Forecast Database. Each of the application groups has different dynamics requiring particular attention. As a result, the methodology for each application group will be completely different. However, there are some general things that we tend to take account of when compiling each of the application group forecasts:
- Current adoption: Our first task when forecasting an application is to find all salient data about current deployments. This involves a lengthy trawl of all available primary and secondary data.
- Regulation: One of the major drivers of IoT is government mandate. This can take the form of explicitly requiring that a device be connected (e.g. eCall in Europe). Or it might be implicit in a regulation (e.g. frequent meter reading would be impractical otherwise). Or it may be necessary to indirectly achieve government targets (e.g. reducing carbon emissions). All of these have an impact and help to drive IoT.
- Demographics: Naturally we also take account of standard demographic and socio-economic statistics. Population size and disposable income will be a key determinant of the adoption of connected consumer electronics devices. Meanwhile aging populations are strong determinants of the use of healthcare applications. There are a number of statistical indicators that might or might not be relevant in any particular application including population, number of households, penetration of cars, and many more.
- Sector-specific statistics: The addressable market for certain applications will vary tremendously depending on particularly characteristics of the market. For instance, the opportunity for the Extractive sector in a market such as the Netherlands is limited, whereas in Saudi Arabia or Canada it is substantial. Similarly, the opportunity for smart gas meters in Sweden is negligible because piped natural gas is rarely seen, although it is common in neighbouring countries. All of these market-specific idiosyncrasies need to be considered.
- Service deployment plans: We speak to a lot of companies in the IoT industry and as a result are party to lots of information about who is launching what service and where.
- Value chain positioning: The value chain for each IoT application is very different and we need to understand who does what in order to identify the revenue opportunity for all of the different participants (see Revenue, above). We spend a huge amount of time with all participants in the IoT value chain identifying who will be the winners and losers.
- Technology availability: Clearly the technology splits in our forecasts depend on the availability of those technologies. Therefore we factor in technology deployment plans into our forecasts.
- Evolving bill of material costs: One of the constituent parts of our revenue forecast relates to device. Therefore we need to be aware of the cost implicit in connecting devices. Furthermore the cost of connected modules will have a substantial impact on connections. As a result we need to closely monitor prices.
Once we have built this bottom-up forecast we then undertake a process of tallying this with the figures from operators and regulators around the world. This is specifically relevant to wireless wide-area network connections. This way we can ensure that our bottom-up picture of the current market tallies with our understanding of the total size of the market globally and on a country-by-country basis. And finally, we're constantly testing our assumptions with the toughest critics there are: our clients.
If you would like to discuss our methodology, please contact us.
For each of the application groups included in the Forecast Database we present the same sets of data. Broadly speaking this splits into three categories: connections, traffic and revenue. Further detail on these three elements is provided below. The forecasts cover every country in the world, as well as six regions.
Connections and RGUs
The Forecast Database counts IoT connections regardless of technology. Connections is split between eight categories. Traditional wireless wide area network is split between 2G, 3G and 4G. A further category is Low Power Wide Area (LPWA) which includes a range of new technologies such as Weightless and Sigfox. The remaining four are short range (including WiFi, Zigbee, Z-Wave etc), MAN (including diverse technologies such as wide-area powerline and mesh), wide-area fixed (including DSL, cable and fibre) and satellite.
It is important to note that we distinguish between connections and revenue generating units (RGUs). In many cases there are multiple connected devices per RGU (e.g. in the case of a connected alarm system). In some there are multiple RGUs per connection (e.g. in the case of a vehicle platform supporting navigation, tracking and in-vehicle entertainment applications through a single connection).
Traffic tends to closely follow connections. We apply assumptions of average traffic usage to each of the connection or RGU types as appropriate. Traffic is solely a measure of wide area network traffic.
The revenue forecasts within the Forecast Database are very detailed, including all income from the sale of IoT devices and related services. IoT Application revenue includes device costs, together with all monthly subscription, connectivity and traffic fees. IoT Services revenue includes wider ecosystem revenues related to application development and hosting, data monetisation and IoT project work.
The forecasts include granular analysis of the revenue opportunity. IoT Application revenue comprises two main elements: Upfront and Recurring revenue.
- Up-Front revenue includes the entire device cost where connectivity is integral to the device and modem costs where devices can optionally have connectivity enabled. The forecasts include figures for both Module and Non-Module device revenues.
- Recurring revenues include both Connectivity and Service Wrap, which we define as revenues for the services that are ‘wrapped around’ a connectivity link to support an IoT application. Connectivity revenues includes revenues for Cellular, LPWA, Metropolitan Area Network, Satellite, and Wide Area Fixed connectivity types.
- Cellular and Satellite connectivity revenues include analyses of Network Traffic revenues, Connectivity Support Platform revenues and Value Added Connectivity
Total IoT Service revenue comprises revenues for Applications, Data Monetisation, Platforms & Middleware and Project Work. IoT Service Revenue Data should not be used in isolation, see Research Note “Forecasting the totality of the IoT revenue opportunity” for further details.
Each of the elements presented in the chart below is available in the Forecast Database for each application group and each country.
Full definitions and explanation of forecast metrics are included in the document “Machina Research Advisory Service: Guide to the IoT Forecast Database”.
Accessing the Forecast Database
The full Forecast Database is accessible online via our Forecast Database console. This can be found via the Forecasts tab on the website.
The console (below) allows subscribers to select data by region, country and year. They can then select from 58 Application Groups across our eight segments. Finally they choose which of our comprehensive set of metrics they want. Then hit "Search!" and the data will be delivered in excel format.
If you have any further questions about the Forecast Database, or if you would like to arrange a demo, please don't hesitate to contact us.
The lead analyst for this Research Stream is Margaret Ranken.