Thursday, November 21, 2013

Product Management - Value of Data Visualization


Recently I was asked to talk about product visualization, and this blog is a summary of that.

Human brains are designed to think graphically - i.e, we see  things in pictures and try to understand the various relationships between them. Once the images are understood, human mind is capable to taking swift action. But if the brain fails to understand the image then there will no action. To illustrate this consider the game of Cricket (for those who don't know about cricket - think baseball)

In the game, the batsman has to see the ball, judge the speed, the angle, the height, the turn/swing and then makes a decision on how to play the ball. If he gets it correctly - batsman can score runs else if he get it wrong he is out.

Now imagine that you could document all that visual information into text and present it to the batsman and ask him to make a decision in a fraction of a second! I bet there is no human who can process that quantity of text and play the shot in time.

While the idea of presenting all the bowling data in text sounds ridiculous, but we tend to do that every day. Computer systems today are capturing so much data - often in text or in binary (which is non-readable by humans) and then asking people to make decisions on that!

Today, we are dealing with the challengs of big data. One of the main challenge with big data is visualization and product managers have to deal with it. Data visualization is not an exact science, meeting this challenge is called as the art of product management.

As a product manager of software, Service Assurance Suite - populalrly known as Smarts that analyzes big data - millions of events coming from thousands of devices. The data is collected continiously in real time - so that it can do an effective data-driven analysis. The software provides up-to-the second updates, so that operations team can immedieatly access the imact of any application or infrastructure changes - even before the end users can notice the effects.

The software will have to process this information in real time and present it to users in meaningful ways.

Let me explain this with an example.

In Infrastructure management, Smarts product identifies a physical server has failed. A statement "Server A1234 is down" is useful - but it does not help the data center operator in terms of solving the problem. So to make his job easier, we added a geographic map and location map information to the server, which tells where exactly the server is located, in which rack, in which data center etc. Now the data center operator knows where the device is located and can quickly remedy the problem. Similarly, proving a heat map of the data center along with high temperature alerts is more useful.


While it may not be obvious to non-users abour the importance of the above example because Smarts as a product is something you may not use, but think about it in your life example: Your friend/spouse says he/she has a car broken down on the road - and is trying to tell you the location, and compare it with getting the same information - with a map and photos of the car. Now its is easy for to understand and decide on the action you need to take.

This is the key to the value proposition of data visualization!

You can convey a whole lot of information with visualization. But then there is a challenge for product management - in terms of covering all possible use cases,  which is almost unviable. Instead we need to open up the data visualization aspects & allow customers to create their own visualization.

Often, customers have multiple data streams which relate to the same event. So having an ability to mashup multiple data steams and create highly valuable interactivity with the underlying data. In other words the visuals are not static - customer can add or delete different data sources, understand relationships, patterns within the complex data set. Customers now have the ability to click on various parts of a visual to drill into different views of the same data on the fly. Customer can intergrate other Business intelligence tools to create drill down performance reports, compliance reports etc. Such a visual mashup will be very powerful and useful for customers.

As part of complete makeover of the software, we added a mashup capability in the latest release and one of the best features of this modern visualization tool is that they permit interactivity with a wide range of data streams. Customers now have the ability to click on various parts of a visual to drill into different views of the same data on the fly. Customer can intergrate other Business intelligence tools to create drill down performance reports, compliance reports etc. Customers should no longer be constrained by a limited data visualization options. Data visualization tools of today must not apply limits on what users can do, instead visualization tools must open up a lot more options for analyzing data.

Never underestimate the values of an appropriate visual when it comes to helping customers understand the situation and help get your point across. Having a high-impact visualization is a must in order to fully grasp the huge volumes of data.

The good news is that modern visualization tools can really help users at any skill level do a better job of analyzing, comprehending, and presenting information. There is a reason for the saying "A picture is worth a thousand words."