Vijay (who describes himself as “just an average Joe, making a living doing SAP consulting for IBM” and is not affiliated with Norada) makes a strong case for why it’s important to understand the context of information streaming (re-posted with permission). - Steve (founder)
Seeing only some of the information relating to a client or project creates a distorted view of what’s really happening, causing missteps and passive engagements. When your team has all related information in front of them the context becomes obvious and will alter how they react to it. Simple, eh? Read on…
This happened in the year 2000. I am fairly new to USA, and sitting in my client’s IT offices in Colorado Springs, CO. It is close to lunch break – and my mobile rings. I pick it up – and it is my dear friend from India. A minute later I am telling him – in an exasperated voice - “Do not let that bitch on your bed, she won’t let you sleep one wink all night”. Next thing I know, 20 heads are staring at me with horror written large on their faces. Little did they know that I was actually talking about a dog – a female German Shepherd that I bought and sent to my friend in India to show there in dog shows. Despite me explaining to the best of my abilities, I am sure not all 20 believed me then, or ever after.
That is the thing about context – data without context means nothing. And this is especially true in the world of analytics. A given set of data can mean many things to many people. Consider this example. If you are at the physician’s office to check blood work results. You and the doctor are both staring at the same numbers. Yet – doctor and you have two levels of understanding about what those numbers mean. But why does that happen?
It happens because we try to abstract everything into some common model for all users of the information. Since you and the doctor have a difference in your level of knowledge on the subject, the idea is to make it useful to you – who has minimum knowledge. It can also work in another way – like the stock market. There, the information is skewed towards the more knowledgeable users – and the layman investor cannot make use of it easily.
But why does this happen? We have raw data – so this should not be hard to represent it in ways that a given user can figure out. The reason is that if we were to create a report per type of user, it is a nightmare to develop and maintain. But what if there was a tool and a framework that could take raw data and present it differently to different users without a developer writing report after report?
Here is another scenario. Lets say an order entry clerk is entering a sales order. The context has all the information on customer, product, location and so on. And there are probably plenty of BI reports that analyze the product and customer in a hundred different ways. But the clerk probably does not know that or care about that. Wouldn’t it be nice if system had the ability to ask the Clerk “hey did you know this customer always pays on time, and hence you should check if he can be given a loyalty discount?” or “we have this other product that is similar, but no one is buying it. Why don’t you ask the customer if he would take that instead for a reduced price?”. That is actionable information – presented in a way that a user can understand. And depending on context, system can figure out what the user can be presented with.
The best part is – there are plenty of technologies around us that can do parts of these already. Missing aspect is the integration of all of them. So – will all the smart product developers recognize this and do something about it? I am counting on them ! As Intel says in their campaign – “it is not what we make, it is what we make possible”.