Assessing approach towards BI in a challenging environment

eLearningCurve’s Dave Wells says initiatives need to be taken to re-establish credibility and belief in the value of analytics-driven business management

Are managers trying to use analytics suited for the past to manage into a very different future?

Dave Wells, director of education , eLearningCurve, certainly believes so.

Wells, who is scheduled to chair the forthcoming Business Analytics Summit to be held in San Jose (November 12-13), spoke to Business Analytics News about the current trends in business analytics. Excerpts:

Corporations are capitalising on business analytics to achieve new breakthroughs in process performance. What according to you have been critical breakthroughs in data mining and text mining over the past few months?

Dave Wells: I believe that the most significant breakthroughs have been in the area of sentiment analysis which goes beyond mining of facts to develop inferences from those facts. Customer sentiment is probably the most wide-spread application today and it is not as widely adopted as it might be. I'd say that the maturity of the technology is ahead of the mainstream corporate world today -- not to say that the technology doesn't have room to grow, but that most companies aren't ready to fully embrace what is available today.

Do you think the emergence of sophisticated analytical tools have automated the processes of pattern recognition and prediction? How do you assess the role of statistical regression and time series methods, and machine learning techniques in pushing this sector forward?

Dave Wells: There is no doubt that the field has advanced, and that the integration of statistical methods and time-series analysis are significant contributors. I think machine-learning lags a step or two behind statistical and time-series methods but has an important role to play.

But I think the big inhibitor - the bridge yet to be crossed - is making this level of analytics available to the masses. Most of the people who do business analysis aren't trained statistical modelers. They're business managers who perform analysis out of necessity. They need easy, intuitive desktop tools with the models embedded, the statistical methods hidden, and connectivity to enterprise data from the desktop as a prominent feature.

How do you assess the advancements in tools used for discovering and extracting knowledge from text documents and overall converting unstructured content to structured content?

Dave Wells: I think there has been a lot of progress in the direction of extracting meaning from unstructured data -- really extracting the hidden structure from what is coming to be known as semi-structured data. The implication there is that the data has a combination of structure and noise, and that separating the structure from the noise is the key to finding meaning. The basic capabilities of the tools have come a long way.

The next big step is in contextual frameworks for finding structure and meaning -- taxonomic, semantic, and lexicological frameworks that vary both geographically, by industry, by profession, and so on.

Earlier this year you had mentioned that many of today’s business managers rely on the same analytics – the same metrics, scorecards and dashboards – as used in the past. Considering that relevance, usefulness and value of yesterday’s analytics have diminished severely, why do you think this is happening? Does it highlight the fact that business analytics is yet to gain acceptance in organisations across various industries in a big way?

Dave Wells: BI had gained pretty wide acceptance when the economy was healthy and managing business was fun and rewarding. Now it is stressful and a lot of hard work only to be followed by new problems to be solved.

At the time when analytics are most needed, the resources - both people and dollars - to pursue them are limited. So many managers are trying to use analytics suited for the past to manage into a very different future. That's not going well, and it is certain to cause some disillusionment with BI. I think there is some work to be done to re-establish credibility and belief in the value of analytics-driven business management.

It is highlighted that the most common issue relative to analytical information and technology is having too much of it scattered about. Companies have multiple BI software packages installed in different functions and they have multiple data warehouses. Also, firms lack overall technology architecture both complete and flexible enough to organise and manage information. How do you assess the situation?

Dave Wells: I think it will always be a reality that we struggle to get at the right information, in the right ways, and to do it quickly. We'll never succeed at fully integrating all of the data and all of the technology in any enterprise. We continue to push the leading edge of technology (and we must do so to keep up) but we rarely take time to pull forward the trailing edge. As the gap between leading and trailing continues to get wider, the challenges of technical integration become more complex. Now add in the dimension of mergers and acquisitions and you further compound the problem. Maybe integration isn't the Holy Grail. Maybe its time to embrace a combination of mashup, federation, and agile methods instead.

Considering the current economic environment, how do you think the metrics need to be benchmarked in a different manner? Can you elaborate on this please?

Dave Wells: The basic difference between old-economy and new-economy metrics has to do with the questions that they answer. The old metrics are designed to answer questions of: What happened? How much happened? Why did it happen? Those are all retrospective questions, and the new economy calls for prospective questions.

Instead of concentrating on the past we need to look to the future, answering questions like: What is the best that can happen? What is the worst that can happen? What is probable? The challenge of new metrics is that they still have to be based upon the data, and the data is all about the past. Therefore, the change is not in what data we use to produce metrics, but in how we interpret that data. With a seismic shift such as deep recession the data of the past is no longer a reliable predictor of the future. To make the shift from looking backward to looking ahead calls for a few fundamental changes in how we approach metrics -- less focus on monitoring, and more focus on simulation; less reliance on data alone, and more blending of data, human knowledge, and feedback systems; less concern about insight, and more attention to foresight.


 

Business Analytics Summit

Business Analytics News is scheduled to conduct the two-day Business Analytics Summit at San Jose in November (12-13) this year. The conference will feature  leading Business Analytics executives including ones from Monster Worldwide, JetBlue Airways, New York Times Company, Boire Filler Group and Data-Miners Inc.

For more information, click here: http://www.businessanalyticsnews.com/usa/agenda.shtml

Or contact: Ben Satchwell by email ben@businessanalyticsnews.com