Business Intelligence and Intelligent Business Decisions

Background and History of Business Intelligence:

According to Wikipedia, the term ‘business intelligence’ was first used by IBM researcher Hans Luhn in 1958. He defined it as ‘the ability to apprehend the interrelationships of presented facts in such a way as to guide action towards a desired goal.’

Since that time, business intelligence (also known as ‘BI’) has developed and grown, and nowadays when the term ‘BI’ is used it generally refers to the gathering, storing and analyzing of data for the purpose of making intelligent business decisions.

Purpose of BI Technology:

Most (if not all) businesses deal with huge amounts of data. Analyzing this raw data in a quick and accurate manner is extremely difficult due to key trends that are masked by the sheer amount of information to digest.

This is where business intelligence software and technologies can help. BI provides a proven, well-defined methodology to process and analyze business-related information quickly and accurately. This is accomplished via the BI ‘stack.’

Business Intelligence Stack:

The BI stack has been traditionally defined as follows:

1. Data Layer:

a. Consists of the raw data that needs to be analyzed.
b. Data can originate from multiple sources, such as: MySQL, MS SQL, Oracle and Access databases; OLAP (online analytical processing) sources; various spreadsheets like MS Excel; CSV files; and even data sources that are not structured.

2. Analytics Layer:

a. This layer is responsible for transforming the raw data into meaningful information.
b. Components that constitute this analytics layer can be:

i. Data mining: refers to the process of extracting patterns out of raw data.
ii. Predictive analysis: involves the analyzing of data and then predicting future events or patterns.

iii. KPI formulation: the formulation of key performance indicators (KPIs) which are meaningful to a business.

iv. And any other business-specific methods of transforming and massaging data.

3. Presentation Layer:

a. The presentation layer is responsible for visually representing the data provided by the analytics layer.
b. Data visualization can be accomplished via digital dashboards, performance scorecards, graphs, reports, gauges, indicators and any other visualization components.

To summarize how the BI stack functions:

1. Data is collected from a variety of sources.
2. The data is then transformed into meaningful information.
3. The massaged information is displayed to end users using data visualization methods.

The Future of Business Intelligence:

BI technology is constantly evolving and this is reflected by changes to the BI stack.

It is important to note that the stack is not ‘set in stone.’ The nature of business varies to a great extent and how a company chooses to implement business intelligence in their decision-making processes will affect their implementation of the BI stack.

Recent modifications to the stack include things such as:

• The mass adoption of mobile BI, which is reflected in the presentation layer.
• Major advancements in predictive analytics, a component of the analytics layer.
• Environmentally friendly data storage.

Business Intelligence Software Review – BI Solutions For the Midsize Market

In September 2009, IBM published the results of their survey with 2,600 CIOs from 78 different countries and 19 different industries. 83% of these personally interviewed CIOs selected Business Intelligence and Analytics as their top priority for 2009 and 2010.

In the last 6 months, we tested several business intelligence (BI) solutions like Cognos, iQ4bis, Business Objects, Qlikview, Micro Strategy and BDA. They varied a lot in functionality, pricing, required resources (external and internal), target groups, integration with multiple data sources and more.

In addition to our own research, we interviewed more than 50 companies to get their input about the experience with these BI solutions, their TCO (total cost of ownership) and their future plans with Analytics.

We found that larger enterprises were less price-sensitive and were still investing in external consultancy and adding additional user to their BI system. However the vast majority of the smaller and mid-size companies, (most of them with $50 million to $500 million annual turnover, so called SME or Small and Mid-Size Enterprises) want to reduce their cost for analytics applications and take more ownership by using internal resources to build or expand their data warehouse and analysis applications.

Therefore, we concentrated our efforts in finding and researching BI applications which were inexpensive, feature rich and easy to implement with in-house resources. We found one clear winner, a relatively new business intelligence software, called BDA which stands for Business-Data-Analysis. BDA offers a data warehouse based on Microsoft SQL-Server technology (but works with any kind of data sources like Oracle, IBM DB2, SQL-Server, Access and more), a reporting and analytics front-end and many pre-packaged business solutions like Sales Analysis for JD Edwards, Sage, BPCS, SAP and many more. Two features we liked most was that BDA offers an unlimited amount of front-end users for their reporting and analytics system and the fact that IT departments can take full ownership of BDA to build or expand the application.

BDA also offers a free POC (Proof of Concept) with your own data and the implementation time for a typical business analytics application like sales-analysis is less than one week.

Using Heat Maps to Improve Sales – Making Business Intelligence Work For the Bottom Line

A problem looking for a solution or a solution looking for a problem?
Someone said that heat maps are the candy of business intelligence: taken with pointed moderation, heat maps are great for you, but abuse them and they’ll produce unwanted fat in your budget and time. What do I mean? Let’s see if you can answer this question: what is the main reason why business intelligence projects fail? If you’ve said “technical problems,” you’re stone-cold. If you’ve said “for lack of a meaningful business-intelligence strategy,” you’ve hit the bull’s-eye.

According to analyst firm Gartner, only five to ten percent of firms deploying business intelligence (BI) have a clear strategy for implementing their new resource. And guess what? These are the companies for which BI makes a tangible difference–yes, I’m talking ROI, as cliche’ as that sounds. In other words, these companies begin with outlining the direction in which they want to go, then select the appropriate tool to take them there. And in most cases, they reach their destination. For the others, instead, BI is–at best–a tool looking for a problem to fix.

Let’s use a practical example of how a common business situation can benefit from finding the right BI tool–in this case, heat maps.

You are the VP of sales. Your goal: to aggressively increase your outside salesforce’s effectiveness. Your resources: your existing salesforce, say, 100 strong, and a few tens of thousands of dollars in budget.

You break down your tactics into three major pieces. First, seeing the effectiveness of your current salesforce: who’s performing well and who isn’t. Secondly, identifying the reasons behind well- and poor-performing salesmen. And thirdly, taking corrective action to maximize the effectiveness of your resources, and testing the results.

The first thing you need is hit your data, and find a way to ask it the right questions.

Your data is in an old-fashioned table. The rows represent the individual salesmen, or their territories. The columns show you different things–from orders taken year to date (YTD) in units and dollars, to numbers of visits, to percentage achieved of sales-goal, to percentage achieved in each product line. How do you make immediate tactical sense of all this? This is when business intelligence–and meaningful business intelligence tools like heat maps–will come in handy. See? Problem first, tool later.

I know what a heat map looks like; can you tell me what exactly it is?
There are different ways to define heat maps. I like mine. Heat maps are a way to display data from a table visually through cell-size and color in a way that a) makes immediate sense and b) helps you quickly answer your “why” questions. “Immediate” and “quickly” being the operative words. Each cell of a heat map represents a row of data in the table; cell-size and color represent two columns–whichever you want them to be. A color slider under the heat map allows you to spot outliers on either end of the values of the column associated with color.

And unlike a traditional graph, heat maps are optimal to display multiple rows of data–up to hundreds of rows–which would result in a visual mess on a traditional table.

Using our scenario as an example, you now have 100 cells on your heat map–your 100 rows in the table, or your 100 salesmen. You choose to assign cell size to (say) YTD sales in dollars, and cell color to number of orders taken.

Also unlike a traditional graph, a heat map will automatically sort things for you–by cell size. So you’ll immediately see that Johnson’s big cell is at the top-left of the map, telling you that he’s got the highest YTD sales in dollars, while poor old small Flaherty’s on the bottom right has the fewest. Cell color, though, tells you that Johnson has taken fewer orders than Flaherty–all this without you doing nothing but staring at a colored gizmo for a few seconds. And now, if you move your slider left, you see that Joyce has taken the least amount of orders, while Hernandez has taken the highest.

So, you’re now convinced that a heat map can give you a useful way to implement your sales strategy–time to spend the budget on BI and roll forward.

Using the heat map to drive up your sales
Let’s think back to the three tactical pieces that the heat map will help you fulfill.

First: seeing the effectiveness of your current sales. As a sales manager, you know that YTD revenue is not necessarily directly proportional to effectiveness. A more useful yardstick would be to compare the individual salesman’s YTD sales quota percentage achieved to the number of sales visits: if you can spot a positive relation between them, you can easily identify the outliers as the problem situations that need the most immediate fixing. And a heat map is the perfect tool to do this quickly.

Assign the “sales quota % achieved” column to cell size, and the “number of sales visits” column to cell color. Immediately, the map will tell you if there’s a (generally) positive correlation between the two values. If the color tends to go to one hue as the cells get bigger and to the other as they get smaller, you have got your answer. Yes, the relation is positive: the more visits, the higher the percentage of the sales goal each salesman achieves. In this case, spotting areas of inefficiency is as easy as moving your color slider to the left: see which salesmen have paid the least amount of visits, and urge them to make more calls.

If the relation is not positive, spotting outliers will also tell you about potential problem areas. Move your slider to the right, and see which salesman, in spite of a high number of visits, has underachieved. Time to retrain? To realign the customer-base? To revisit the sales targets? Regardless, the heat map has alerted you of the situation.

Second: identifying the reasons behind your sales performance. Let’s say that the relation between visits and YTD goal met is positive. You already know one potential reason for the performance of your salesmen–frequency of visits. To confirm this, quickly see if there’s also a direct relation between visits and orders taken, which is as simple as assigning a new column (number of orders taken) to either color or cell size.

But how do you explain the outliers? Play around with your heat map. For instance, have you asked yourself if the product mix is right for all territories? To asses this, simply reassign your heat-map cell size to percentage achieved in each product line. This will tell you more pointedly if there are some product lines that are impervious to more visits, suggesting that there may be an opportunity to optimize the mix for that territory.

A handy feature along these lines would be drill-down. If you could drill down on each individual cell to see–for instance–what YTD goal each salesman has achieved in each individual line, you would be giving yourself a lot of flexibility and would be able to obtain even more answers quickly from your heat map.

Third: taking corrective action and testing the results. Depending on the answers you obtained in steps 1 and 2, you now have a strong grasp of your current situation as well as the possible reasons behind them. You can now start outlining some tactics for improving the situation–starting, for instance, with urging salesmen to pay more visits to their clients or to readjust the product mix for certain territories or regions.

The usefulness of a heat map–with its giving you an immediate (but accurate) snapshot of a situation–is even higher as you test your new solutions. After a week, for instance, you can already see if there’s a positive correlation between the increased number of visits by your under-performing salesmen and a higher number of orders.

All this analysis and testing can be performed in minutes, while without the heat map, it would take days if not weeks.

In summary: make heat maps your ally to drive up sales
As long as you ensure that you have a clear strategy, and that you break down this strategy into manageable tactical steps, modern interactive business intelligence tools like heat maps are invaluable. In the hands of an intelligent and creative manager, a heat map is nothing less than an oracle that will reveal–in an intuitive way–the situation, the possible causes behind it and likely ways to overcome your problems and propel your company towards more revenue.