»Me, myself and BI«

Bissantz ponders


|

The fundamentals of management information (2)

Onwards with our compact list of terms which we feel are important if management information and business intelligence should become effective. This time: from D as dashboard to F as frequency distribution.

Dashboard

A concise collection of KPIs on one page (see also → gauges).


One topic that we frequently discuss at our forums is: What factors determine if the time that you invest in management information is worth the effort?

Data collection

An underestimated means of Business Intelligence. The analysis of existing data suffers, in part, because data was collected for another purpose as the one that interests you today. DeltaMaster supports → instant BI, which means that business users can quickly build an application based on Excel or text files and then analyze it using built-in business, statistical, or data mining methods.

Data density

Criteria for reporting (see also → resolution). Good reports are filled with data yet easy to read. If they aren’t, that is probably due to the design and not the amount of information. Strangely, the market has promoted → dashboards which have only 5 to 10 % of the data density of a simple table showing the results from last week­end’s soccer matches. In DeltaMaster, sparklines, microcharts, and graphical tables ensure high levels of both → readability and data density.

Data Mining

Our favorite topic and the roots of our business. Data mining stands for the desire to automate routines, generate moments of suspicion, and recognize new corre­lations. This topic is experiencing its renaissance through everything that has to do with → Big Data. DeltaMaster is a ground-breaking product for business data mining and has won the Innovation Prize from the German Informatics Society (GI).

Data scientist

A new job title for people who already know everything that is written here. You typically find them among business information systems experts who are very up to date in their field. The main challenge for data scientists is to find the right ba­lance between business and IT. On the one side, business departments have the hopes of making the impulsive behavior of markets and customers more predic­table. On the other hand, IT cannot and does not want to change their data-delivery processes so quickly because the technical questions involved are changing as well. And that is another reason why people want → self-service BI.

Dichotomy

A property of attributes. Some common attributes that only have two forms are: on or off, yes or no, and man or woman. In management information, dichotomy interests us as a postulate for designing warning signs and dealing with colors. → Traffic lights have three states, which makes them too vague in our opinion. You can, however, designate KPIs dichotomously with the help of → color. They either have a positive effect on profits or a negative one, in which case, we color them blue or red respectively in DeltaMaster. We feel if reporting cannot differentiate what is good or bad for the business, it is dodging around its actual purpose.

Differentiation

A criterion for quality charts and reports. Differentiation can be measured as a percent value difference per pixel. It differentiates how quickly you realize what has stayed the same and what is new and different. Differentiation depends on the → scale and → resolution. To obtain the necessary differentiation, you need to display the variances to the average or a target value instead of the absolute values. All templates in DeltaMaster are designed according to a simple rule: Devote every pixel to the difference.

Dot bars, dot columns chart

A new chart type based on a dotted line and a large point that designates the value it is representing. Dot bars/columns were first suggested by William Cleveland. What makes them so sophisticated is that their → perspective priority applies to the gaps between the dots and not the gap to the base line. As a result, you can use them as you would use bars or columns but can apply a → logarithmic scale. This, in turn, solves the problem of → size breaks. Sadly, this chart type is not widely used and, as far as we know, is only offered in DeltaMaster.

Early warning

An important aspect of solid reporting. The minimal requirement is a forecast showing the expected result for a given period. That is not even difficult – as far as the underlying algorithm is concerned. The significance of cumulated variances – between budget values and actuals as well as the current and previous year – are often overlooked. Yet, they say what things will look like in the end if the current level of development stays as it is. DeltaMaster offers a spectrum of early warning methods. Standard reports deliver cumulations and projections. Sparklines support forecasting functionality. → Trumpet curves visualize variance corridors and what-if simulations.

“Easter egg” paradox → Attributes

Excel

The “bad boy” of management information – feared by some, loved by others. Excel is the basis for most spreadsheet data marts, also known as spreadmarts. Users take extracted data from their ERP systems, add structure, make correc­tions, add variances, and then present it using Excel charts that they export to PowerPoint. Accordingly, spreadmarts are the natural enemy of all BI vendors. Bella mainly snarls over a few shortcomings in the visualization. DeltaMaster supports Excel both as a data source and export target, yet also provides an alternative to spreadsheets.

Exception Reporting

Management reports that are only sent when a previously defined threshold has been exceeded. Exception reporting is supposed to help reduce the flood of information. The German information systems expert Norbert Szyperski already questioned this concept back in 1978 and warned that the combination of prede­fined thresholds and the fiction of management by exception are dangerous. In his opinion, management should be curious and search for new correlations among information instead of attentively dozing like a control person at a switchboard. (Source: Integrierte Informationsverarbeitung 2, Wiesbaden 2002)

We also think that way – partially, because managers like to hear that everything is going well and see how that looks like right now. We, too, cannot warm up to the concept of fixed thresholds that trigger alerts. How would you even implement that? How much work does that entail in complex corporate environments? How general or specific should these thresholds be – on country, regional divisional, product, industry, and/or customer levels? Should they be relative or absolute values? Sometimes, people turn to → traffic light charts. Yet, because these signals are cautious in both directions, most things are displayed in yellow. You can set thresholds in DeltaMaster for the rare cases where contractual, qualitative, or technical limits provide clarity . We, however, are big fans of sorting everything in descending order based on the value that is important to the company. That allows you to view everything at a glance – without needing thresholds.

Eye path

An important criterion of report quality. Seeing and thinking are so closely intertwined that seemingly small differences in eye movements encourage or hinder thoughts. Vertical eye movements are less strenuous than horizontal ones. Big jumps are highly unpleasant and ping-pong movements are a clear sign of a failed report design. The so-called → saccade movements of the eye are particularly sensitive to jerking or counterintuitive → animation.

Eyespan

A term coined by Edward Tufte. Eyespan means that the eye-brain systems of observant readers will prioritize and easily process information that it can review with just a few eye movements and without scrolling. This is also a reason why a smaller size is a criterion for quality charts. But beware: Something that is on one page isn’t automatically at a glance. Good design should provide an overview and details within a small space and, therefore, is a challenge. In DeltaMaster, graphic tables with microcharts and sparklines ensure high data density and good → readability for concise yet highly informative reports.

Eyetracking

The process of measuring eye positions and movements. Technical devices measure the eye movements of a test person in order to calculate his or her → eye path, for example, in order to assess the usability of websites. We think that you often don’t have to wait that long. A little self-observation often suffices to see if something is good or really bad.

Frequency distribution

A statistical method used to describe the distribution of values. Frequency distributions are often contained in management reports dealing with marketing or market research. We believe that managers should view data as it is – in other words, organized but not summarized – and not just as → histograms. This can be clearer and more enlightening also due to the → outliers.

Comments are closed.