Who is your customer, what problem do you solve that the customer is willing to pay for? These kind of questions are well known to all businesses and build the foundation to define your business cases. Maybe you use the well-known business canvas to figure out what you need in terms of technical solution, data, services, cost (capital and operational) and the personas at your customer that interact with your product. There are many good resources that explain all the details on good product management and business development. I do not want to repeat this here.

What is now specific to industrial IOT? It is about the differentiation factors in your business model (that turns data into actionable insights) and the awareness of what is going to be commodity soon and what will keep you differentiated for an extended period of time. Also, you need to be clear on the relation of cost and end customer benefit for each use case. But first things first.

In IOT I see three major kinds of business models. Which one is yours?

1. Solution business: In this model the business is based on consulting and development around custom solutions that make use of a specific data set that is emitted by devices or software systems and should be turned into a specific actionable insight. In this kind of solution business, or project business, the end-to-end use cases are clearly specified and directly implemented. Companies may use internal platforms and 3rd party technology to reuse functionality or even build a multi-tenancy system to handle their customers, but this would be only a cost-reduction measure since the inner platform would not be visible to the customer or the public. Beside consulting and solution development, service agreements and support contracts typically build the third commercial pillar here. Competition is mainly cost driven, and of course around the solution-development time (faster and cheaper, please), so companies in this business model build internal platforms and reusable components mainly for cost and speed optimization in all 3 pillars. They are typically also customers of companies that implement the other 2 business models:
2. Technology Services business: In this model the business is around providing pieces of technology that solve a particular problem in the allover problem space of industrial IOT. For instance, provide a solution for handling device certificates in the best way, or allow to easily build virtual, software defined network overlays to secure operations. There are many technology providers. AWS provides services that may serve as building blocks as well: Amazon IOT with features like Device Management, AWS Greengrass, Amazon IOT Analytics for real time analytics on incoming device data streams and many more. Here the business aims at solution builders and integrators, not at the end-user specifically. In this business model it is all about mass adoption of the provided technology, the more users, the more business. Price competition is less clear here due to the fact that functionality is not easily comparable. All vendors have different approaches and other feature sets, so integrators and solutions builders have to consider a large variety of parameters before making decisions. Some of those are: Feature set, cost, operational qualities like scalability & resilience & security, contract conditions and licensing terms, vendor stability (support for +10 years required). So to be successful here, differentiation needs to be tailored to serve solution builders and integrators well. Highly integrated and pre-configured products may serve a specific customer group, e.g. Edge devices with a pre-filled software marketplace or engineering integration with production and automation engineering tools is a strong competitive factor, because it reduces friction and time-to-market for system integrators. On the other end, highly generic services may serve a broad customer base and so provide very cost effective offerings.
3. Platform and ecosystem business: In this business, the primary business model is around enablement and linking of providers and consumers (of data, services or solutions). The platform business is an emerging trend that has many prominent examples: AirBnB, Uber, eBay, and so on. All these companies provide platforms that link providers (apartment owners, car owners, goods sellers) to potential consumers (travelers, buyers). Some of those companies do not produce or own any of the required assets to serve the consumer directly (AirBnB does not own hotels) but act as intermediate. The strength of this approach is in its massive scalability and a potential high volume business once the platform is established. To get the platform established you need a working system that implements all you need to link your ProSumers and execute deals in an automated way (security, client management, contracting, billing, integration, …) and you need both Producers and Consumers in a critical mass to be attractive for both sides. In IOT this could mean that you make it easy for owners of field devices and data to make this data available to a group of analytical companies that can generate insights and money is charged for each data flow or insight delivered. Data Marts are examples for such approaches in IOT. These kinds of platforms have a massive business potential because once established, they can be operated with relative small effort while adoption is basically not limited which can result in a positive loop effect (more consumers attract more producers and vice versa) that ultimately may result in a winner-takes-it-all market position. Social media platforms like Facebook and WhatsApp are prominent examples for such an effect. At the same time this business model is the hardest to achieve because it needs substantial effort to reach a critical mass of eco-system participants and to build a platform that serves enough differentiating features that make usage super easy. The technical platforms that enable this business model must be highly scalable, highly resilient, security around tenant isolation must be highly effective and a lot of different use case scenarios might be required to be supported. Specifically, in industrial IOT, there is a large variety of use cases you could support (make usage easy), so scope is essential to not get into a situation where the platform supports everything a little bit, but not really some one thing great. The availability of solutions makes the use of a platform more appealing to the ecosystem and the more solutions are available on the platform which cater to actual use cases and market demand, the more momentum the platform gains. This mitigates the risk of falling into the trap of unclear focus, plus the increasing momentum attracts more participants to the ecosystem who eventually help further shape the focus of the platform. And remember, making it easy for a user or developer always means high level of automation in your system.

As we can see already, depending on your business model, the focus areas and challenges to tackle are different. There is a common set of factors we will focus on, but to be clear – you need to focus on who your customer is and what your competition/success factors are. Based on that there are no finite answers to the question how you build your uniqueness best.

I would like to briefly discuss a few anti-patterns at this point that will certainly lead to problems.

• Too much focus on data collection. We often encounter approaches where the data collection is a primary goal. Let’s collect all the data, there must be something valuable in it. As a result, companies build data lakes and ingest data from different sources – sometimes with large effort – but without a clear understanding of what this data is good for. In the end, the data lake generates cost, the data lake turns into a data graveyard and becomes a burden. We draw an analogy to the problem of establishing security logging systems which end up as graveyard for log data. Merely centralizing data collection should not be a purpose of its own. To avoid this, analyze the data before you collect it on larger scale, be clear what insights you want to realize and which business outcome you want to achieve – then start implementing the entire chain from data collection to actions based on the insights. Data collection alone has no business value and so builds no USP. Of course, sometimes you do not know the value of data that was never collected or analyzed before, but starting on a smaller scale or proof of concept may help you focus the purpose of data collection to where it adds most value. It is equally important to balance the velocity and granularity of data against the actual use case. Does your use case really require nanosecond collection of raw data? Deciding on the velocity and granularity of your data collection should always be part of your business reasoning because it is a major driver for cost. Not every use case requires real-time data processing and sometimes data transmission is not possible e.g. when there is no internet connection. In such cases, would you want to push data from yesterday to your real-time streaming pipeline? Such cases can become expensive, without adding business value.

• Underestimating the integration and operations efforts. It may occur that you look at a technology provider and you see that many problems are solved. The promise is that technology is easy to integrate and you have quick solutions. However, the integration of systems, be it on connectivity side and around the data (models, semantics), can easily become an issue requiring a lot of thought and resources to solve. You need people who are able to integrate technology and operate it afterwards. At the same time, your people will need to make the right technological decisions for different use cases that balance business value along with integration and operational cost. Common best practices may not be applicable when integrating or operating at mass scale. Some things are simple at small scale but become a challenge at mass scale. You also need to figure out early on how to keep your operations efforts as low as possible if you want to avoid that this becomes a bottleneck and hinder you as you grow. To be successful in the IIOT space requires not only that you have contextual knowledge about the business use cases in order to make the right technological decisions, but also that you can integrate and operate those technologies in the most distributed, stable, secure and efficient manner possible. Providing services around this is a business model in itself and can form a USP. Underestimating the value and expertise required behind integration and operations will become one of your biggest challenges.

• Focus on functionality that is easy to commoditize – unless you are a mass scale business: When we walked around the major industrial automation fairs over the last years, there was a clear trend to observe. Three years back, only a handful of companies showcased products that allowed to collect data at the shop floor (industrial grade) and send it to the cloud. Today, almost everybody has products around this area of the IOT problem space. Collecting data and sending it to a cloud is now a functionality which became commoditized. Today you can pick from a variety of technology providers and for those vendors it becomes harder to differentiate in this area. Technology vendors and platform businesses still can differentiate by specializing but the trend is clear. General data collection, data transport into a cloud system, data storage and analytics infrastructure eventually become solved problems. If you are a new start-up, make sure you solve a more specialized or sophisticated problem. Unless of course you have a clear concept how to compete on price against incumbents.