Why is ‘pervasive business intelligence’ important?

[guestpost]This post is by Alexandra Carvalho, SAP Mentor and Head of Analytics at BI Group Australia. She has 18 years of experience in the implementation of Business Intelligence Solutions in the United Kingdom, Australia, New Zealand and Singapore. You can find Alexandra on Twitter at @AlexTwittAU.[/guestpost]

If you have never heard of the term before, ‘pervasive business intelligence’ means: “a data warehouse that combines historic and real-time data, available to the majority of the organisation’s employees 365/24/7”. (Please refer to this article for an introduction).

Now back to the question: it is important because it democratises Business Intelligence. It gives everyone in the organisation the ability to make informed decisions based on facts. Everyone, at all levels, from managers to operations. It makes analytics solutions relevant to everyone, at all times. As a consequence, it gives the organisation ability to perform work that could not be done before.

Many of us have experienced this frustration – not long ago, one of my colleagues worked on a very technically challenging project to extract and report Employee Full-Time Equivalent information, based on SAP HR, in an organisation that has several employees allocated to temporary functions.

The solution was quite neat: it used all the information available in SAP HR and reported past, present and forecasted future Full-Time Equivalent for each area of the organisation – with full slicing and dicing capabilities – as any good analytics solution must have (ah, and he tells me proud: it reduced the extraction-transformation-loading (ETL) processing time from five hours to 15 minutes).

However, as he spoke to the business, he realised that the information was only relevant at the very beginning of the day (after the overnight load). As transactions occurred during the day (and employees got allocated to temporary positions, were hired, left the company, or retired, and so on), the information in SAP BW became out of date and, for lack of a better word, useless. In a word: HR analysts could not rely on those reports to make decisions during the day.

The information was still relevant to the management: they could see historically which areas are over- or under-allocated, they could see which areas are likely to become over or under-allocated in the future and they could follow one person’s full history in the company. All that is great for management, but not for people making decisions on the go. It is a bit like trying to balance your cheque-book without all the last transactions included.

If we made this solution an ‘Active Data Warehouse’ (one that is continuously updated with live data), then it would have been different: as transactions occurred, they would migrate immediately to SAP BW and the reports would remain relevant at all times. This could be done with their current technology, but the latency of 15 minutes may still not be acceptable.

This is where tools such as in-memory computing become critical. Had this client have their BW on HANA (and maybe the ECC system too), the load times would have been significantly lower. This would make the report relevant at all times and meaningful for those making tactical decisions during the day.

The concept can be extrapolated to other areas: a classical example of the use of Pervasive BI is the system that credit card companies use to match our expenditure to our past pattern. The problem with the current solutions is that they define “normal” very narrowly. And as soon as they detect abnormal expenditure, they block the credit card and it becomes the client’s responsibility to phone the card company and tell them that the expenditure is genuine or not.

I have seen this one too many times: I was in a supermarket and saw a man on the phone, trying to reach his credit card company to unblock his only credit card. He had gone outside his home state of California and happened to use his card in Las Vegas (a known risky place). As soon as he used it, it was blocked.

Now, let us imagine this same man had in his smartphone an app that informed him whenever his expenditure went outside the normal – and normal would be defined more broadly, taking into account the fact that he normally travels a few kilometres from his hometown – or that he occasionally uses it on the internet. Now, as soon as something goes out of the norm, this app alerts him and gives him the option to accept that abnormality. This accepted abnormality now becomes part of his profile. Let us say that now the credit card company uses something like SAP’s Predictive Analytics to develop the model for normal expenditure and to highlight outliers.

This would significantly reduce the risk of charge-backs (one of the major concerns of the credit card companies) and prevent the inefficiency, embarrassment and annoyance that unnecessarily blocked credit cards cause.

On that example, two technologies work together to make this possible: in-memory computing makes it easy to merge real-time data and expenditure pattern data as transactions happen, and mobile computing let the client interact with the credit card company.

There are many other situations where pervasive BI can significantly contribute: condition monitoring (in plant maintenance) alerting for situations where a piece of equipment requires urgent attention, preventing breakdowns; or in health care, alerting for abnormalities in vital signs, preserving life.

Another use is in crew scheduling: our app ‘EquipeMobile’ combines live crew feedback, traffic and positioning information to propose (with several updates during the day), the best schedule for a crew to service their clients on time, respecting priorities, preventing breakdowns and, in some cases, preserving life.

Why is pervasive business intelligence important? Because it opens a plethora of new possibilities for all of us. All thanks to faster processing of in-memory computing and the omnipresence of mobile devices.

 alexandra carvalhoAlexandra Carvalho is an SAP Mentor and Head of Analytics at BI Group Australia.