Why do IoT Architects need to think about value, not just data?

Several years ago I was pitching what would now be called an Industrial Internet of Things (IIoT) solution to the Production Manager of a large manufacturing plant. After describing all the data we could collect, and the metrics we could turn it into, I thought I had done pretty well. What Production Manager wouldn't want our system to get his finger on the pulse of his operation?

Instead, his next question floored me:

"If I don't do anything with the data your system collects, then it doesn't create any value for me, does it?"

I had never imagined that someone presented with real-time, detailed information wouldn't immediately grab it and use it to improve their business. I was so taken aback I could not think of an intelligent response, and needless to say, we didn't win that deal.

I'm not going to suggest that he was right, since "doing something with the data" was implicit in his job description, but there is a germ of wisdom for the IoT community in what he said:

Merely delivering data does not deliver value.

Even lots of accurate data, even in real time. Many IoT systems -- still -- have clearly been designed under the assumption that their responsibility ends at collecting, storing and presenting data: systems where data is collected and put in a data repository or historian; systems where data is collected an put on on-line graphs.

A real-world ACTION that benefits a group of stakeholders is still the only way that any IT system delivers value. For an IoT system to deliver that value, it must construct a chain from data to action. I suggest we call this chain:

The Information Value Chain.

The Information Value Chain is only just starting when you collect the data. Turning that data into information and ultimately into ACTION is harder, and if anything your "data only" Internet of Things (IoT) system has made the problem worse, not better: understanding a small amount of data to turn it into action is extremely taxing, and takes many different skills. Doing that with a torrent of data is overwhelming.

What is the Information Value Chain?

Very simply, the Information Value Chain is the insight that data only creates value if it goes through a series of steps, steps which eventually result in action back in the real world.

Like so:

If we focus primarily on collecting data, then we will create Data Lakes, which are impressive Information Technology constructs, but on their own are passive entities that deliver no inherent value to the organisation.

If we focus primarily on action, then we will make decisions based on inaccurate information and misleading data, resulting in the wrong action, wasted money and lost opportunity. A great example is this Case Study.

How to solve this conundrum? Before we get into the mechanics of building a robust Information Value Chain, the starting point is human, not technological.

To succeed you must start with the right goal

The starting point is this: What is the motivation for your project?

If it is to build an "IoT System," then I suggest that you are heading down the road to failure. An IoT System is a means, not an end, and has as many different embodiments as the word vehicle - Ferrari; Ford Focus; Mack truck; oil tanker.

Here is what you should be setting as your goal:

"To build a system that creates value in [this] way; by enabling [these] actions; using the best methods; with the minimal required human intervention; based on the best possible information; in as close to real time as possible."

There is a lot in this statement. Let's unpack it.

The central message of the Information Value Chain is to see our information systems as part of a sequence who's end result is action that delivers value.

1. When I approach systems analysis for a Customer, the first thing I write on the right hand side of the whiteboard is a "$" sign.
2. To the left I have the Customer help me develop an ROI model:
   • Before: X1 action by X2 participant creates X3 value at X4 cost;
   • After: Y1 action by Y2 participant creates Y3 value at Y4 cost.
3. Then we step left again to describe the decisions that lead to those actions. Now we can write:
   • Who (or what!) will make those decisions; on
   • What timescale;
   • Based on what algorithm.
4. Now we can ask what information they will need to make these decision and
   • How to extract this information from the data available.
5. Then, and only then, do we know what data to collect; how to process it, how -- or whether -- to present it; and how much of it and how to store it.

We have found this approach moves IoT from a vague concept of something the Client thinks "maybe" they should do, but are not clear on how it will impact their business, to a compelling business tool with clear purpose and value. That what this is all about!

What do the links in The Information Value Chain mean?

The terms data, information and decision, as well as knowledge and intelligence get thrown around a lot, often interchangeably, yet these are distinct concepts. It is important to understand what we are talking about so that we can define and deliver each link in the chain successfully. Let's start from right to left, as we have just described in our systems analysis process so that we always keep our end goal in mind:

• Action: something that results in a change in the real-world which has a $ measurable value to a key stakeholder;
  
• Decision: a choice between possible Actions made according to a set of rules that maximize the value of the action taken;
• Information: Data interpreted in a specific context to best support the Decisions the User needs to be able to make;
• Data: individual facts collected from the Real World environment, as accurately and as timely as possible, not all of which will be relevant to the Decisions to be made;
• Real World: The totality of systems, machines, people and environmental factors that can affect the right Action to take in any given circumstance.

How do we turn The Information Value Chain into practice?

The Information Value Chain is a great conceptual framework to think about how to get from Data to Value, but as IoT system architects, we are concerned with the practical question of how to deliver Value from Data. This is the purpose of the 5D IoT Architecture, which maps the links in The Information Value Chain to 4 specific architectural components, suggests core requirements for each of those components, and adds a 5th component to continuously improve the solution itself.

This paper is the development of a series on concepts in Big Data, IoT and systems architecture originally published on Fraysen Systems.