How to Get Actionable Business INTELLIGENCE using BI & Analytics.

This is one of my favorite topics of all time when it comes to business intelligence. If you look at any traditional BI solution whether it is a dashboard with charts or reports you get in your email, they all give information about what, where & when some things happened. Essentially providing you only with information instead of Intelligence. Let's look at some business questions to see what I mean:

• What are our sales in Philadelphia by Sales Rep & Product?
• How many targeted ad campaigns did we run this year on various social platforms?
• What are the Orders shipped vs. fulfillment rates for Women's Shoes broken down by warehouse and sales channel?

One common theme among these questions is that you are asking for something that you think may have a problem. Once you get the results from your favorite BI tool (dashboard, report, SQL queries, excel dumps &, etc.), you then look for anomalies in the results hoping to find a problem then hopefully fix it. Essentially what you get in return is part of the problem when it comes to using Business Intelligence tools to get intelligence.

What you get back is exactly what you asked for and nothing else. Essentially you are getting back Information NOT Intelligence.

Still not sure what I mean? Here is an example. Let's assume I am the director of sales for Pennsylvania & I notice I am not going to hit my sales targets for this year. As a manager, I know there is a problem that I need to address in somewhere within the products, locations and the sales rep that I am responsible for. So I start by asking the following question using our corporate BI tool of choice.

What are our current sales by products, reps and locations in PA?

As a result, I get back exactly what I asked for. After all, this is what database SQL queries are designed to do. They return results that you asked for and remove anything else that you didn't. Once I get the results as either a report, dashboard, excel file &, etc., I can then go through the information and try to spot some products, reps or locations that may have less than ideal results so I start working on improving them.

So far what I got back is information, not Intelligence. I knew we sold a bunch of products in a bunch of locations by a bunch of sales reps and simply got a list of that information with matching metrics for each combo. However, my real business problem was to grow sales in PA not to figure out much stuff we are selling in PA. Unfortunately, this is how most people think when it comes to solving business problems when it comes to BI & Analytics and it is how most BI products are designed to work. Ask a question, get the answer. Find an anomaly or something interesting in the answer and ask another question.

Question is, are you asking the right questions? What if the anomaly was not in the initial answer? What if you never thought to ask the right question that will show you an anomaly? This means you never get to ask the follow-up questions to find the problems that exist in your business. Using traditional BI, your analysis is pretty much limited to what you know & ask for. This is where you can benefit from getting actual intelligence from your BI platform so you can ask the right questions to lead you into the problems so you can fix it. In short;

You can't analyse what you don't know.

What I needed was to find out what I did NOT know about the sales in PA. This is the real INTELLIGENCE that can lead me to solve my business problem. What I should be focusing on is what is NOT happening in PA. To do this, I need a special breed of BI tool that can help me figure out all the things that I DON'T know about sales in PA such as:

• What products are NOT selling in PA so I can add new products to my market?
• Which reps are NOT selling our current most popular products in PA. We got products that sell well in PA but not all reps are selling them. Notify these reps to start pushing these existing profitable products.
• What locations(cities, counties, zips &, etc.) in PA we are NOT selling anything? Identify new locations in PA to market our products. (Add sales reps, expand marketing &, etc.)
• Which PA customers purchased something from us this year but have NOT purchased any of our top selling products in PA? (Create a list of customers to up or cross-sell)

As you can see, when you start asking business questions with the word NOT in them, you start getting actionable business INTELLIGENCE instead of just business INFORMATION that you need to dig in for some possible intelligence.

There are some issues with this approach. The first one is the human problem. It requires users to change their standard way of asking questions. Also, how can users possibly think of all these combinations of "NOT" questions for every business problem that they face? Second is the technical problem. None of the typical BI tools (reports & dashboards) are equipped to provide these kinds of answers. It may require redoing most of our BI resources but we]you still have to think of all these combinations of questions and create more reports & dashboards to give you the NOT results. Since the problem is both human and technical related, you need a new breed of BI tool that allows users to ask the standard questions they ask as they always have and use the best in class technology to provide them with both answers at the same time. Showing what they asked for along with highlighting all the things they haven't thought about asking.

This is where the QlikSense Enterprise BI platform comes into the picture with its unique ability to identify actionable insights within your data. Qlik's BI platform which runs on its unique analytical engine is already being used by many Fortune 500 and thousands of small & mid-sized companies around the globe and has been in the Gartner BI top quadrant spot for many years in a row. Even though Qlik is well known within Enterprise size companies for its extremely high-speed analytics engine that can crunch through millions upon millions of rows of data within milliseconds to allows users to get answers as fast as they can ask questions, it has one game-changing feature that no other product has which in my opinion have not been socialized enough.

This awesome feature has been providing some of the biggest organizations with invaluable new insights that they could not otherwise recognize even though they have been working with the same data sets for many years.

Marketing Lingo for this feature is called "Power of Grey" or "Green, White & Grey". It is essentially a very simple way of displaying associations among data values within your analytical data model that may contain numerous tables & columns form a variety of data sources using specific colors. This is what 3 specific colors indicate:

• Green values indicate values that you choose to filter your data. (In SQL terms, anything that would be in your WHERE clause)
• White values are the results that are directly associated with your (green)filter values. (These are same values you would get back from a traditional SQL query in your reports or dashboards)
• Grey values are the potential hidden insights in your data and exclusive to Qlik. Grey values are the hidden gems that saved millions of $$$ for its users. Grey values are the data values that are NOT associated with your filters. These are the values that would normally be removed by any traditional Query base BI solution and not be exposed to users mainly for performance reasons associated with relational SQL databases. These limitations simply do not exist when you are using a world-class in-memory engine.

Imagine being able to run a corresponding NOT EQUAL query with every single user request and be able to display both results sets to users at all times. (Matching ones in White, NOT matching ones in Grey). In SQL geek language, So grey values are like running two simulations queries every time user applies a filter to their data and displaying both results sets in a very user-friendly way using two different colors (white for matches, grey for nonmatching values):

Typical BI query for Products Sold in PA:

SELECT Products FROM Orders WHERE CustomerLocation = 'PA'

Qlik Engine for Products Sold in PA (includes products NOT sold in PA by default):

SELECT Products FROM Orders WHERE CustomerLocation = 'PA' (White)

+

SELECT Products FROM Orders WHERE CustomerLocation <> 'PA' (Grey)

If you think automatically being able to see what is not sold in PA is awesome, Qlik's engine takes it to the next level by not only giving you intelligence around products that are not sold in PA but about every other value in every other column in every other table from every other data source that exists in your analytical data model. Now you have instant visibility into Customers not associated with PA in customers table, Product web pages that PA customers are looking at but also products that they are not looking at from Web Analytics systems such as Google, Warehouses, Sales Reps, Freight Companies, Sales Channels, Shipment Methods, Suppliers, Contractors and everything else that may exist in your sales analytical data model that may be coming from various sources like CRM, Sales Ops, Customer Service, ERP, WebAnalyics & etc. but are NOT associated with customers in PA.

The best part is all this intelligence is delivered to users the moment they click on PA anywhere within a Qlik Dashboard whether it is located in a drop-down list, table, bar or a pie chart. It doesn't end there either. Start adding more filters by selecting values from other fields like choosing a specific product, shipping company or a sales rep to see more specific results around what is not matching. With a few simple clicks, you instantly start seeing and discovering things that you did not think to ask in the first place. Here are some of the things that our customers discovered withing days of using Qlik by simply filtering on the same data values they had been using for years:

• Why USPS is Grey in the shipment field when I select Product X or Warehouse Y along with Ground Shipment? USPS is generally cheaper than most others, Why are we are not using them for these specific products or from these Warehouses?
• Why does this commercial mortgage Loan code turn grey when I select all the loan numbers for this year? I just discovered a bunch of un-serviced loans worth over 10 million dollars due to using an invalid loan code to these loans.
• Why did the state of "MARYLAND" turn grey when I selected "Zillow" in my target ad campaign. We always assumed we were marketing to all states with Zillow. Who knew we haven't been advertising in his state for months due to an expired contract?.
• Why are these OrderID numbers grey when I select all of my orders in the header table? I always taught all line items should have a matching order header. just identified a software bug in our systems causing faulty reporting.
• Why is this surgical glue brand turn grey when we select Orthopedic Surgery? This glue is half the price of others and not being used in any of the surgeries in any of our hospitals. Who knew the glue name in the database had a non-ASCII character which prevented it from being displayed in the webpage drop-down list that doctors use to order supplies for their operations. Dozens of hospitals * Lots of Orthopedic Surgeries = Huge amount of $$$ saved by using a cheaper glue.

At the end of the day, all of these users assumed things were supposed to be one way and never think to question otherwise. Once using Qlik, these insights were highlighted to them as grey values while they were doing what they would normally do with any other BI tool. Just making simple selections and filtering data!

But this time Qlik was the new BI platform and this simple feature alone was able to give them invaluable actionable business insights instead of the traditional Business Information dump by showing things that they never thought to ask resulting in millions of dollars in additional revenue or savings.

Just to illustrate the concept, here is what I quickly discovered using a popular sample dataset which is around sales. All I did was, drop some dimensional columns on the canvas as drop-down lists and just started clicking on values. In less than a minute, I was able to spot two important insights within this data set that you would most likely miss using a traditional BI tool.

PS: I found a bunch more but 2 screenshots are enough to make this point.

Did you know they are not selling any Furniture products in North Dakota?

And why none of these products are being shipped using Standard Shipping.

As a result, 'BI' in BI & Analytics should stand for Business Intelligence and not for Business Information. Today's businesses need actionable intelligence from their data to stay competitive. In my opinion, is best achieved by presenting them with things that they do NOT already know with every single selection they make instead of delivering information about only the known things using common BI tools.