I recently did a brief blog post for Izenda on IoT and business intelligence, and this part of my IoT series expands on some of the themes there.
The Internet of Things is a new phenomenon; that said, a simple search for ‘Internet of Things IoT’ brings back over 60 million search results in Bing. What is the Internet of Things? The Internet of Things Global Standard gives us the following definition: ‘The Internet of Things (IoT) is defined as Recommendation ITU-T Y.2060 (06/2012) as a global infrastructure for the information society, enabling advanced services by interconnecting (physical and virtual) things based on existing and evolving interoperable information and communication technologies.
Now, this definition is fine but it focuses on the ‘shiny’ aspect of IoT and, most importantly, it does not mention the data aspect of IoT. It emphasises the connectedness of the various gadgets and their interoperability. I prefer Peter Hinssen’s discussion, where he recommends that we talk about the value of the network of things. The connected devices, on their own, will fulfil their purpose. However, if you want real insights from these sources, then you need to splice the data together with other sources in order to get insights from it.
The thing is, the Internet of Things is really the Internet of You.
We are now heading towards the Zettabyte generation thanks to the Millennial generation. For example, the World Data Bank projects that 50% will have smartphone by 2016, and 80% by 2020. We sent 7.5 trillion SMS messages in 2014. In fact, one app, WhatsApp, sent 7.2 trillion messages. And that’s just data from one app. In 2000, Kodak processed 80 billion photos processed from camera film. In 2014, 800 billion photos from smartphones were shared on social networks. And that’s just the photos that were shared.
We are the Internet of Things.
From the business perspective, how do you make use of that IoT data? The consumerization of IT means that business users are often asked to manage and cleanse data, regardless of its size and nature. Research suggests that data is growing at a rate of 40% of each year into the next decade, driven by increased online activity and usage of smart devices. (ABI, 2013). (The New York Times, 2014 ). The consumerization of data means that business users should be able to access and analyze the data comfortably. When we introduce data that comes under the umbrella o the Internet of Things, business users will need to be able to access IoT data from devices as well as other data sources, to get insights from the data.
How can we harness the IoT phenomenon to understand and serve our customers better?
The addition of data from a variety of sources, including data from devices, means that IoT has a very wide scope and opportunity. IoT can focus on the devices themselves, or the network infrastructure connecting devices, or the analytics derived from the data which comes from the network and the devices. In order to get true insights, the IoT data would be deployed in tandem with other relevant data so that the business user obtains the context. The IoT would also introduce real time data, which would be mixed with historical data.
Customer expectations are rising; and customer-focused businesses will need to put analytics at the heart of their customer services. For example, customers do not distinguish between out of date data, and inaccurate data; for them, they are the same thing. The customer landscape is changing, and it includes the ‘millennials’ who expect technology to offer an unfailing, personal experience whilst being easy to use. This expectation extrapolates to data, and customers expect organizations to have their data correct, timely and personal.
For organizations who put customers front-and-center of their roadmap, management should encourage self-reliance in business users by ensuring that they have the right tools to provide customer-centered service. Unfortunately, business users can suffer from a split between business intelligence reporting, and the operation systems, as a result of decoupled processes and technology at the point at which they are trying to gain insights. Often, users have to move from one reporting technology to another operational system, and then back again, in order to get the information that they need. This issue can be disruptive in terms of the workflow, and it is an obstacle to insights. In terms of IoT data, business users may have to go and get data from yet another system, and that can be even more confusing.
What does IoT mean for BI? Business Intelligence has matured from the earlier centralized technology emphasis, to a more decentralized business-focused perspective which democratizes the data for the business users. With the advent of IoT technologies, issues on collecting, refining and understanding the data are exacerbated due to the introduction of a variety of structured and unstructured data sources. In the meantime, there is an increased interest in businesses to find insights in raw data. However, this remains a challenge with the introduction of IoT data from devices, creating a mélange of data that can be difficult for business users to assimilate and understand. Companies risk obtaining lower ROI in IoT projects, by focusing only on the data, rather than the insights.
How did the industry get to this point, with disjointed technology and processes, and disconnected users? How can we move forward from here, to including IoT data whilst resolving the issues of previous business intelligence iterations? To understand this unquenchable thirst for data by business users and what it means for the future, let’s start by taking stock of the history of Business Intelligence. What are users’ expectations about data and technology in general? Until recently, these expectations have been largely shaped by the technology. Let’s start with the history lesson. What are the historical stages of Business Intelligence?
The First Generation of Business Intelligence – change in the truth
First generation Business Intelligence is the world of corporate Business Intelligence, embodied by the phrase ‘the single source of truth’. This is a very technical discipline, which focused on the extract-transform-load processing of data into a data warehouse, and focused less on business user intervention. The net result is that the business seemed to be removed from Business Intelligence. In response, the users pushed for decentralization of the data, so that they could drive their own decision making using the data flexibly, and then confirm it independently in the context in which they are operating. In terms of technology, business users reverted to the default position of using Excel as a tool to work with Excel exports, and subverting the IT departments by storing data in email.
The Second Generation of Business Intelligence – change in the users
Second Generation Business Intelligence was the change was wrought by the business users, who demanded clean data sources on a strong technical apparatus that they could mash and merge together, and they were empowered to connect to these sources. In this stage, the industry pivoted to bring the Business back into Business Intelligence, and it is typified by the phrase self-service business intelligence. The intended end result is that the business has structured data sources that the users understand, and the technical teams have a robust technical structure for the architecture in place. As before, Excel remained the default position for working with data, but the business users had more power to mash data together. Self-service business intelligence was not faultless, however. Business users were still dissatisfied with the highly-centralized IT approach as they still relied on other people to give them the data that they need. This issue introduced a delay, which increased the ‘time to answer’ metric whilst simultaneously not recognizing that this feeds into the ‘time to question’ metric. It does not recognize that analytics is a journey, and users expect to ‘surf’ through data in the same way that they ‘surf’ through the Internet.
What problems does IoT introduce for businesses, and how can we resolve them?
Given that there are inefficiencies in the process of business intelligence in organizations at the moment, how is this exacerbated by the introduction of data from devices, otherwise known as the Internet of Things? IoT data introduces new issues for business users for a number of reasons. IoT devices will transmit large amounts of data at a velocity which cannot be simply reconciled with other time-based data. The velocity of the data will add in an additional complexity as business users need to understand ‘what happened when’, and how that marries with existing data sources which may even be legacy in nature. Business users will need that work to be presented to them simply. Further, IoT devices will transmit data in different formats, and this will need to be reconciled so that the actual meaningful data is married to the existing, familiar data. If the business users are moving around disparate technology in order to match the data together, then the disconnected technology presents an obstacle to understanding the data, and thereby obtaining the promised holy grail of insights.
IoT means that we can obtain a granular level of customer data which provides unique insights into customer behavior. However, it does not immediately follow that the data will bring insights on its own; interpretation and analysis will be vital in order to obtain insights. Businesses can interpret IoT as equivalent to data from devices, and it is easy to distracted by shiny gadgetry. The ‘shiny’ approach to IoT can mean that business users are ignored in the journey, thereby shearing their insights from the solution as a whole.
Helping Business Users along the IoT Journey
As internal and external expectations on data increase, the pressure on business users will increase accordingly. Business users will need help to travel along the user journey, to adapt these changes in the data landscape that include IoT data. One solution is to add a new application that will help the business users to work with the IoT data. However, adding a new application will exacerbate the existing issues that business users experience. This might be an easy option for IT, but it will add in a new complexity for the business user. The introduction of IoT data does not necessitate the introduction of new technology to analyze the data.
IoT data resides mainly in the cloud, which means that organization’s infrastructure is changing rapidly. It will need to be reconciled and understood, regardless of where it resides. Organizations can end up with a hybrid architecture of cloud and on premise solutions, and the midst of these complex and fast-moving architectures, business users are easily forgotten. The business users will need to have a seamless environment for resolving cloud and on premise systems to enable them to product the anticipated analysis and results. Business users will find it difficult to navigate the terrain between cloud and on premise data, which will aggravate existing issues in the putting together existing data sources.
Business users have a need for data to carry out a fundamental analytical activity: comparison. How does this data compare to that data? How did last year compare to this year? How did that customer compare with this customer? Answering these simple questions may mean that traditional analytical tools may not be able to cope with the new types of data that are generated by IoT technologies, because the data will be disconnected in terms of technology and process. Excel is excellent for rectangular data sources, but it is not designed for data sources where the data travels at such velocity, and in non-rectangular shapes. So, what’s next?
The Third Generation of Business Intelligence – change in the data
The Third Generation of Business Intelligence is where users work in the era of real change in data, and it is this change is wrought by changes in the data itself. The data has changed fundamentally; it now travels faster, has more shapes, and is bigger in size than ever before. Users are not finding it easy to compare data simply by farming data into Excel; they need to be empowered to tackle seemingly simple business questions, like comparison, in a way that fits with a fluid way of working whilst being empowered with data from the IoT sphere as well as more traditional sources. By tapping into new volumes and particularly new varieties of data, organizations can ask questions about their customers and their business in a way that they have never been able to do, previously. Further, when we add IoT into the mix, there is a promise of insights from customers and their environments, which can be incredibly valuable to companies. It is not all one way, however: In this era of tech-savvy consumers, customer relationships require planning, nurturing, and constant attention.
There should be an acceptance that business users will want access to IoT sources in the same way as any other source, but these can be exasperating and non-intuitive. Simplicity is vital in the race for the business analyst and user, and their goal is to reduce time and effort in getting the insights that they want, whilst increasing efficiency and value.
So, what gets forgotten in the IoT torrent of attention, and what can we do about it?
Simply put, the business users get lost. They are already getting lost frequently with BI projects, and this will only make matters worse for IoT projects.The ones who mash data together, clean it, make decisions on it, and put it next to other data sources in order to make sense of the data – these are the ones who should be using the data.
Given all of these issues, how do we bring the users back into an IoT architecture? I was faced with this issue recently, when designing an IoT architecture which had a focus on machine learning. IoT work involved a great deal of complexity, which is neatly hidden behind the buzzwords.
The changes in data now mean that there is a clear extension of where the industry has come from, and where it is headed. So what comes next? The third generation of business intelligence: ready to go analytics using data regardless of its shape and size.
Organisations will need to focus on the third generation of Business Intelligence if they are to be successful in facilitating users to have the access to data that they need. Users will want to try and analyse the data themselv es. Fast questions need fast answers, and businesses need to move from their initial business question through to the resulting insight quickly and accurately, in a way that they are comfortable. They also need results at the velocity of the business; answers when they need them. Remembering the users is a deceptively simple requirement that presents a number of challenges.
The dislocation between IT and the business is at its apex when we look at the opposing approaches to data. IT is still seen as a bottleneck rather than an enabler. Business users perceive IT departments as a lag in a process that needs to get from question to insight quickly, and they will look for ‘good enough’ data rather than ‘right data’ in order to get it. The way forward is to make the business users’ activities simpler whilst providing a solution that the IT department are closely involved and find the solution easier to support, so that both parties feel that they own the solution.
The solution should put the focus back on the business users who not on the humans who actually deliver service, create insights, and ultimately add business value. To do this, they need to be able to search for meaning in the data, via aggregation, broadcasting and consuming information in order to add the value that is expected of them.
To summarise, these issues were at the forefront of my mind, when I was architecting an IoT solution recently. In my next post, I will explain my technical choices, based on these considerations. On my survey, it was clear that IoT needs to be taken to a further stage so that it is usable, actionable and sensible; not just data about sensors, but data that is relevant and provides insights.
If you want to talk more about the IoT issues here, or you’re interested in having me come along and speak at your event or workplace, please email me at Jen.email@example.com