http://www.b-eye-network.com/blogs/devlin/ As one of the founders of data warehousing back in the mid-1980s, a question I increasingly ask myself over 25 years later is: Are our prior architectural and design decisions still relevant in the light of today's business needs and technological advances? I'll pose this and related questions in this blog as I see industry announcements and changes in way businesses make decisions. I'd love to hear your answers and, indeed, questions in the same vein. Copyright 2012 Wed, 02 May 2012 06:56:42 -0700 http://www.movabletype.org/?v=4.261 http://blogs.law.harvard.edu/tech/rss Donald Farmer, now of Qliktech, offered to the Boulder BI Brain Trust (BBBT) last week that what we in "BI" do is better described as decision support rather than business intelligence.  The comment was greeted by a flurry of Tweets and Grunts of agreement.  It's an observation I've also made, and for similar reasons.  In essence, BI tools support decision making; to attribute intelligence--business or otherwise--to software seems somewhat presumptuous.  And yet, there is a further problem with the term business intelligence.  It implies a level of rationality in decision making that is beyond the reality most of us encounter.  This implication is carried even further as various analysts and vendors begin to talk about business analytics as if it will be the ultimate solution to all business decision-making needs.

There are facts, we are told.  And if we have all the facts and we apply comprehensive analytics, we will discover the past, understand the present and predict the future.  We are told this is the scientific method; the truth is in the numbers.  Is this a valid way of interpreting the way the world works?  I would argue that it is so far from reality that we are in danger of creating a fantasy world worthy of Tolkien.

History, it is said, is named thus because it is "his story".  History, they say, is written by the victors.  The implications are far reaching.  Yes, there are indeed facts, but it's the stories we weave around the facts that are what really matter.  To quote Liz Greene(1): "Mehmet the Conqueror invaded Constantinople in 1453.  That is an historical fact.  But depending on which history book we read, Mehmet was either a redeemer or a cruel tyrant, a warrior for the True Faith or a vile heretic."  In terms of the story we tell ourselves about this incident and its value as guide for the future, which part of the quote is more relevant - the historical fact or its interpretation?  And if you haven't yet looked at the endnote for the source of this quote, do so now.  And be brutally honest with yourself.  Do your beliefs about astrology affect the weight you attach to the quote?  And when I tell you that Dr. Liz Greene is also a fully trained and qualified Jungian psychoanalyst, how does that cause you to re-evaluate your judgement.

Our current obsession with analytics is dangerous.  It's based a number of simplifications, misconceptions and downright errors.  It is a simplification that business is an entirely rational, fact-driven process.  It is a misconception that given sufficient data you can predict the future.  It is a downright error to assume that in the future, business can be entirely (or even largely) driven by business analytics.

Does that mean we should abandon analytics?  Of course not.  There are facts to gather that have so far remained undetected.  These facts can influence our interpretations.  If they are indeed relevant to the story at hand.  And if we allow them to do so.  And if our business users can avoid statistical errors such as confusing correspondence with causality.  There are many examples already of significant benefits to be gained for businesses who adopt analytics.

The questions of relevance and abuse of statistics are ones of good analytic practice and education of users.  I have no doubt that, as we move beyond the hype phase, these issues will be addressed.  The issue of interpretation is much more difficult to tackle.  Because it is at the heart of how we imagine our decision makers behave.  Our focus on intelligence--rational and logical--obscures two other keys aspects of decision making:  intent and intuition.  Intent we ignore and intuition we dismiss.  All decision making includes the intent of the decision maker.  That intention drives everything from what data is gathered, through how it is evaluated, all the way to the final choice of action.  How many decisions are post-justified by careful data selection and evaluation?  If a decision maker is motivated by personal gain (and they do exist, you know), won't analytics be enlisted to support that goal?  And regarding intuition, it is evident that not all decision contexts are wholly driven by measurable or predictable metrics.  Low prices may be important, but so too are ambience, history, ethos and personal relationships when customers choose where to shop.  Data measures for the latter are hard to define and capture.  The intuition of an experienced manager is needed in such circumstances.  Target's decision to focus marketing of maternity products to women in the early stages of pregnancy was based on sound analytics according to a story in the New York Times, but the reaction of prospective customers was intuitively obvious.

The bottom line is that we focus exclusively on big data and analytics at our peril.  We need to move beyond traditional concepts of business intelligence and decision support.  I see our goal as supporting full-spectrum business insight.

(1) Greene, L., "Apollo's Chariot - The Meaning of the Astrological Sun", CPA Press, (2001)

]]> http://www.b-eye-network.com/blogs/devlin/archives/2012/05/death_by_a_thou_1.php http://www.b-eye-network.com/blogs/devlin/archives/2012/05/death_by_a_thou_1.php Wed, 02 May 2012 06:56:42 -0700 Donald Farmer, now of Qliktech, offered to the Boulder BI Brain Trust (BBBT) last week that what we in "BI" do is better described as decision support rather than business intelligence.  The comment was greeted by a flurry of Tweets and Grunts of agreement.  It's an observation I've also made, and for similar reasons.  In essence, BI tools support decision making; to attribute intelligence--business or otherwise--to software seems somewhat presumptuous.  And yet, there is a further problem with the term business intelligence.  It implies a level of rationality in decision making that is beyond the reality most of us encounter.  This implication is carried even further as various analysts and vendors begin to talk about business analytics as if it will be the ultimate solution to all business decision-making needs.

There are facts, we are told.  And if we have all the facts and we apply comprehensive analytics, we will discover the past, understand the present and predict the future.  We are told this is the scientific method; the truth is in the numbers.  Is this a valid way of interpreting the way the world works?  I would argue that it is so far from reality that we are in danger of creating a fantasy world worthy of Tolkien.

History, it is said, is named thus because it is "his story".  History, they say, is written by the victors.  The implications are far reaching.  Yes, there are indeed facts, but it's the stories we weave around the facts that are what really matter.  To quote Liz Greene(1): "Mehmet the Conqueror invaded Constantinople in 1453.  That is an historical fact.  But depending on which history book we read, Mehmet was either a redeemer or a cruel tyrant, a warrior for the True Faith or a vile heretic."  In terms of the story we tell ourselves about this incident and its value as guide for the future, which part of the quote is more relevant - the historical fact or its interpretation?  And if you haven't yet looked at the endnote for the source of this quote, do so now.  And be brutally honest with yourself.  Do your beliefs about astrology affect the weight you attach to the quote?  And when I tell you that Dr. Liz Greene is also a fully trained and qualified Jungian psychoanalyst, how does that cause you to re-evaluate your judgement.

Our current obsession with analytics is dangerous.  It's based a number of simplifications, misconceptions and downright errors.  It is a simplification that business is an entirely rational, fact-driven process.  It is a misconception that given sufficient data you can predict the future.  It is a downright error to assume that in the future, business can be entirely (or even largely) driven by business analytics.

Does that mean we should abandon analytics?  Of course not.  There are facts to gather that have so far remained undetected.  These facts can influence our interpretations.  If they are indeed relevant to the story at hand.  And if we allow them to do so.  And if our business users can avoid statistical errors such as confusing correspondence with causality.  There are many examples already of significant benefits to be gained for businesses who adopt analytics.

The questions of relevance and abuse of statistics are ones of good analytic practice and education of users.  I have no doubt that, as we move beyond the hype phase, these issues will be addressed.  The issue of interpretation is much more difficult to tackle.  Because it is at the heart of how we imagine our decision makers behave.  Our focus on intelligence--rational and logical--obscures two other keys aspects of decision making:  intent and intuition.  Intent we ignore and intuition we dismiss.  All decision making includes the intent of the decision maker.  That intention drives everything from what data is gathered, through how it is evaluated, all the way to the final choice of action.  How many decisions are post-justified by careful data selection and evaluation?  If a decision maker is motivated by personal gain (and they do exist, you know), won't analytics be enlisted to support that goal?  And regarding intuition, it is evident that not all decision contexts are wholly driven by measurable or predictable metrics.  Low prices may be important, but so too are ambience, history, ethos and personal relationships when customers choose where to shop.  Data measures for the latter are hard to define and capture.  The intuition of an experienced manager is needed in such circumstances.  Target's decision to focus marketing of maternity products to women in the early stages of pregnancy was based on sound analytics according to a story in the New York Times, but the reaction of prospective customers was intuitively obvious.

The bottom line is that we focus exclusively on big data and analytics at our peril.  We need to move beyond traditional concepts of business intelligence and decision support.  I see our goal as supporting full-spectrum business insight.

(1) Greene, L., "Apollo's Chariot - The Meaning of the Astrological Sun", CPA Press, (2001)

]]> http://www.b-eye-network.com/blogs/devlin/archives/2012/05/death_by_a_thou.php http://www.b-eye-network.com/blogs/devlin/archives/2012/05/death_by_a_thou.php Big data Wed, 02 May 2012 06:56:42 -0700 Attending the Teradata Universe 2012 in Dublin, an impressive line-up of speakers from Tim Berners-Lee to customers doing real data warehouse implementations got me thinking beyond the normal boundaries about our assumptions about the real role and value of data - both traditional and big.  A few observations follow, but first...

As an ex-pat Irishman, I have to say that the new Convention Centre Dublin is a wonderful venue for events with up to a couple of thousand attendees.  The main auditorium is a superb space and there's lots of room for expo and breakouts.  And the facilities and staff are first rate.  Well done!  My only regret is that the area around the Centre, especially towards the Port, remains blighted by vacant sites and unfinished blocks - the legacy of Ireland's boom and bust - but not much can be done about that for now.

Much of the main tent focus at this year's event was on the future of information, with big data featuring... well... large in the presentations of speakers such as Erik Brynjolfsson, Professor and Director of the MIT Center for Digital Business and Sir Tim Berners-Lee, inventor of the World Wide Web.  Michio Kaku, Professor of Theoretical Physics at City College of New York, also addressed the theme of the central role of data in every aspect of our future.  The tone of these presentations is best described as expansive and optimistic - given better and more data and technology, the future of business and humankind in general is rosy.  This is an expectation that I, personally, believe to be of somewhat low probability.

While I am a long-time supporter of the need for and value of good and extensive information in business, my experience of the purposes for which such information is used and the extent to which decision making benefits is less sanguine.  In general business, business intelligence is used almost exclusively in support of a narrowly-focused drive for bottom-line profit.  At the risk of being labeled a Communist, I remain unconvinced that this is always a good thing.

This niggling doubt is best expressed through an example - the use of data warehousing in retail, something that has been going on for over 25 years.  BI can be very effective in optimizing the supply chain from manufacturer all the way to customer, supporting the intent of the business to reduce cost.  When that focus is pursued as a sole strategy, it can have highly undesirable effects, through driving local suppliers out of business, reducing a community's disposable income and creating an unbreakable downward economic spiral.  As a BI community we can say that BI is not responsible, and on the level of cause and effect, that's true.  But, at a deeper level, we cannot ignore the side effects of the tools and techniques we invent and promote, any more than cigarette manufacturers can avoid responsibility for the impact of passive smoking.

Getting back to big data, the problem is that as we focus on, and get excited about, a technique such as statistical analysis of social behavior to predict marketing trends for a brand, for example, we simultaneously narrow our focus on potentially interesting or important information that is external to that data.  Big data encourages us to somewhat obsessively analyze in ever greater depth the minutiae of life.  Why?  Often to drive profit for some business.  The optimistic view I mentioned earlier imagines that we will use this data to solve medical issues, world hunger, climate change, and more.  I don't have data to confirm this, but I guess that the proportion of profit-driven big data analytics vs. altruistic is greater than 10 to 1.  And which of these two categories of information have the highest impact on the medium- and long-term survival of humanity?  The last speaker, Deb Roy, CEO of Bluefin Labs, showed us just how much analysis can be done to link social network activity to TV shows and advertising.  All to decide where to spend millions of advertizing dollars.  There two ways of looking at this: (1) everybody needs to do this type of processing in order to compete, or (2) we need to examine our underlying model of doing business that drives such net-non-productive activity.  I would invite you to share your views on this.

At a more mundane and practical level, speakers from current Teradata customers focused in a very different area - creating consistent and integrated enterprise data warehouses for very traditional transaction business data.  Unsurprisingly, the majority of enterprises are still struggling with the old issues that drove data warehouse development for the past 30 years.  I have no doubt that this will continue for most businesses for many more years.  

But, while this continues, we need to start thinking about the more philosophical issues that the conference brought up for me.

]]> http://www.b-eye-network.com/blogs/devlin/archives/2012/04/not_only_sql_no_1.php http://www.b-eye-network.com/blogs/devlin/archives/2012/04/not_only_sql_no_1.php Wed, 25 Apr 2012 04:18:03 -0700 Attending the Teradata Universe 2012 in Dublin, an impressive line-up of speakers from Tim Berners-Lee to customers doing real data warehouse implementations got me thinking beyond the normal boundaries about our assumptions about the real role and value of data - both traditional and big.  A few observations follow, but first...

As an ex-pat Irishman, I have to say that the new Convention Centre Dublin is a wonderful venue for events with up to a couple of thousand attendees.  The main auditorium is a superb space and there's lots of room for expo and breakouts.  And the facilities and staff are first rate.  Well done!  My only regret is that the area around the Centre, especially towards the Port, remains blighted by vacant sites and unfinished blocks - the legacy of Ireland's boom and bust - but not much can be done about that for now.

Much of the main tent focus at this year's event was on the future of information, with big data featuring... well... large in the presentations of speakers such as Erik Brynjolfsson, Professor and Director of the MIT Center for Digital Business and Sir Tim Berners-Lee, inventor of the World Wide Web.  Michio Kaku, Professor of Theoretical Physics at City College of New York, also addressed the theme of the central role of data in every aspect of our future.  The tone of these presentations is best described as expansive and optimistic - given better and more data and technology, the future of business and humankind in general is rosy.  This is an expectation that I, personally, believe to be of somewhat low probability.

While I am a long-time supporter of the need for and value of good and extensive information in business, my experience of the purposes for which such information is used and the extent to which decision making benefits is less sanguine.  In general business, business intelligence is used almost exclusively in support of a narrowly-focused drive for bottom-line profit.  At the risk of being labeled a Communist, I remain unconvinced that this is always a good thing.

This niggling doubt is best expressed through an example - the use of data warehousing in retail, something that has been going on for over 25 years.  BI can be very effective in optimizing the supply chain from manufacturer all the way to customer, supporting the intent of the business to reduce cost.  When that focus is pursued as a sole strategy, it can have highly undesirable effects, through driving local suppliers out of business, reducing a community's disposable income and creating an unbreakable downward economic spiral.  As a BI community we can say that BI is not responsible, and on the level of cause and effect, that's true.  But, at a deeper level, we cannot ignore the side effects of the tools and techniques we invent and promote, any more than cigarette manufacturers can avoid responsibility for the impact of passive smoking.

Getting back to big data, the problem is that as we focus on, and get excited about, a technique such as statistical analysis of social behavior to predict marketing trends for a brand, for example, we simultaneously narrow our focus on potentially interesting or important information that is external to that data.  Big data encourages us to somewhat obsessively analyze in ever greater depth the minutiae of life.  Why?  Often to drive profit for some business.  The optimistic view I mentioned earlier imagines that we will use this data to solve medical issues, world hunger, climate change, and more.  I don't have data to confirm this, but I guess that the proportion of profit-driven big data analytics vs. altruistic is greater than 10 to 1.  And which of these two categories of information have the highest impact on the medium- and long-term survival of humanity?  The last speaker, Deb Roy, CEO of Bluefin Labs, showed us just how much analysis can be done to link social network activity to TV shows and advertising.  All to decide where to spend millions of advertizing dollars.  There two ways of looking at this: (1) everybody needs to do this type of processing in order to compete, or (2) we need to examine our underlying model of doing business that drives such net-non-productive activity.  I would invite you to share your views on this.

At a more mundane and practical level, speakers from current Teradata customers focused in a very different area - creating consistent and integrated enterprise data warehouses for very traditional transaction business data.  Unsurprisingly, the majority of enterprises are still struggling with the old issues that drove data warehouse development for the past 30 years.  I have no doubt that this will continue for most businesses for many more years.  

But, while this continues, we need to start thinking about the more philosophical issues that the conference brought up for me.

]]> http://www.b-eye-network.com/blogs/devlin/archives/2012/04/not_only_sql_no.php http://www.b-eye-network.com/blogs/devlin/archives/2012/04/not_only_sql_no.php Big data Wed, 25 Apr 2012 04:18:03 -0700 Many BI vendors now offer collaborative support as an additional feature of their tools.  Unfortunately, that's exactly what it often looks like--an add-on feature to an existing environment.  What I describe in my new White Paper, "iSight for innovation", is how collaborative decision making could (and maybe should) be addressed in a new context.  Not as a bolt-on afterthought to existing business intelligence, but as a new environment of which existing BI tools and methods are but a part.

Fifty years after the advent of decision support systems, and despite two decades of BI, enterprise decision making and, in particular, highly innovative decision making, remain a hit-or-miss affair.  This is because we have bought into a myth that great decisions and innovations spring from the mind of a lone genius.  But if we examine innovative decisions in most organizations, particularly large enterprises, we see most breakthroughs coming from teams--not from some whiz kid.

Business intelligence tools for most of their twenty year history have focused their efforts on the individual decision maker.  Where does he get the data needed and in what form?  What techniques should be provided for exploration and querying?  How will she visualize the results?  All of this is certainly valid, but it homes in on one single facet of decision making, which I call investigation. Investigation begins with a challenge, followed by the gathering and integration of formal information in the shape of documents and databases.  Then she works on it.  And generates further formal information, such as spreadsheets and presentations, in the process.  Her intention determines what information is gathered, the path followed and results produced.  It's a largely solitary process, with peer or managerial review at predetermined points in time--often only at the end.

If we accept the premise that most innovative decisions in business emerge from teams working together--and there is much research that suggests this is so--we see immediately that BI tools, as currently structured, don't fit the bill.  Furthermore, bolting collaborative tools onto them cannot change the underlying process from individualized to team oriented.  The iSight model of collaborative decision making and innovation starts from interaction.  This is the process that goes on between team members.  Having explored that, we can combine it with the investigation process that each individual performs as part of the team.  When we look at interaction, we see that there exists another category of information, labelled informal, that is continuously exchanged between team members and is the source of most, if not all, of the innovation of the team.

Thus, iSight brings together formal and informal information, the worlds of Business Intelligence and Enterprise 2.0, in a framework that drives novelty in decision making.  It presents a high-level architecture that maps to specific tools and methods required to create a systematic process within the enterprise that delivers real, implementable innovations.  This model's strategic power comes not only from assisting in making today's decision well, but also by capturing informal information of the group interactions in each decision-making event so as to make useful recommendations for subsequent decisions.

The goal of this paper, developed in close collaboration with Scott Davis, the visionary CEO of Lyzasoft, is to introduce a new way of looking at decision making in a team context. The iSight model is really just a foundation for the extensive new thinking I believe is needed to define how to support collaborative decision making.  I mean, really support it and facilitate it.  I encourage you to take a look and welcome your comments or questions... ]]> http://www.b-eye-network.com/blogs/devlin/archives/2012/03/collaborative_b_1.php http://www.b-eye-network.com/blogs/devlin/archives/2012/03/collaborative_b_1.php Mon, 26 Mar 2012 09:11:46 -0700 Many BI vendors now offer collaborative support as an additional feature of their tools.  Unfortunately, that's exactly what it often looks like--an add-on feature to an existing environment.  What I describe in my new White Paper, "iSight for innovation", is how collaborative decision making could (and maybe should) be addressed in a new context.  Not as a bolt-on afterthought to existing business intelligence, but as a new environment of which existing BI tools and methods are but a part.

Fifty years after the advent of decision support systems, and despite two decades of BI, enterprise decision making and, in particular, highly innovative decision making, remain a hit-or-miss affair.  This is because we have bought into a myth that great decisions and innovations spring from the mind of a lone genius.  But if we examine innovative decisions in most organizations, particularly large enterprises, we see most breakthroughs coming from teams--not from some whiz kid.

Business intelligence tools for most of their twenty year history have focused their efforts on the individual decision maker.  Where does he get the data needed and in what form?  What techniques should be provided for exploration and querying?  How will she visualize the results?  All of this is certainly valid, but it homes in on one single facet of decision making, which I call investigation. Investigation begins with a challenge, followed by the gathering and integration of formal information in the shape of documents and databases.  Then she works on it.  And generates further formal information, such as spreadsheets and presentations, in the process.  Her intention determines what information is gathered, the path followed and results produced.  It's a largely solitary process, with peer or managerial review at predetermined points in time--often only at the end.

If we accept the premise that most innovative decisions in business emerge from teams working together--and there is much research that suggests this is so--we see immediately that BI tools, as currently structured, don't fit the bill.  Furthermore, bolting collaborative tools onto them cannot change the underlying process from individualized to team oriented.  The iSight model of collaborative decision making and innovation starts from interaction.  This is the process that goes on between team members.  Having explored that, we can combine it with the investigation process that each individual performs as part of the team.  When we look at interaction, we see that there exists another category of information, labelled informal, that is continuously exchanged between team members and is the source of most, if not all, of the innovation of the team.

Thus, iSight brings together formal and informal information, the worlds of Business Intelligence and Enterprise 2.0, in a framework that drives novelty in decision making.  It presents a high-level architecture that maps to specific tools and methods required to create a systematic process within the enterprise that delivers real, implementable innovations.  This model's strategic power comes not only from assisting in making today's decision well, but also by capturing informal information of the group interactions in each decision-making event so as to make useful recommendations for subsequent decisions.

The goal of this paper, developed in close collaboration with Scott Davis, the visionary CEO of Lyzasoft, is to introduce a new way of looking at decision making in a team context. The iSight model is really just a foundation for the extensive new thinking I believe is needed to define how to support collaborative decision making.  I mean, really support it and facilitate it.  I encourage you to take a look and welcome your comments or questions... ]]> http://www.b-eye-network.com/blogs/devlin/archives/2012/03/collaborative_b.php http://www.b-eye-network.com/blogs/devlin/archives/2012/03/collaborative_b.php Collaborative BI Mon, 26 Mar 2012 09:11:46 -0700 "Social networks already know who you know", "recommendation engines get much smarter", "early detection mitigates catastrophes".  Three of ten ways big data is creating the science fiction future.  These types of headlines appeal to the geek optimists in many of us.  We think that mitigating a catastrophe is certainly a good thing.  That smarter recommendations to whom we should connect and what we might be interested in buying could probably save us time, that most precious of commodities.  Most of us have grown up with a belief system that science and, by extension, technology and computers, are a sine qua non in today's world.  In truth, the world we live in today could not exist without them.

But, at what cost?

]]> http://www.b-eye-network.com/blogs/devlin/archives/2012/02/big_brother_or_1.php http://www.b-eye-network.com/blogs/devlin/archives/2012/02/big_brother_or_1.php Fri, 17 Feb 2012 02:47:37 -0700 "Social networks already know who you know", "recommendation engines get much smarter", "early detection mitigates catastrophes".  Three of ten ways big data is creating the science fiction future.  These types of headlines appeal to the geek optimists in many of us.  We think that mitigating a catastrophe is certainly a good thing.  That smarter recommendations to whom we should connect and what we might be interested in buying could probably save us time, that most precious of commodities.  Most of us have grown up with a belief system that science and, by extension, technology and computers, are a sine qua non in today's world.  In truth, the world we live in today could not exist without them.

But, at what cost?

]]> http://www.b-eye-network.com/blogs/devlin/archives/2012/02/big_brother_or.php http://www.b-eye-network.com/blogs/devlin/archives/2012/02/big_brother_or.php Big data Fri, 17 Feb 2012 02:47:37 -0700 Now, I may be accused of getting up on my soap box in this first post of 2012, but... a few recent articles on the topic of big data / predictive analytics have really got me thinking.  Well, worrying, to be more precise.  My worry is that there seems to be a growing belief in the somehow magical properties of big data and a corresponding deification of those on the leading edge of working with big data and predictive analytics.  What's going on?

The first article I came across was "So, What's Your Algorithm?" by Dennis Berman in the Wall Street Journal.  He wrote on January 4th, "We are ruined by our own biases. When making decisions, we see what we want, ignore probabilities, and minimize risks that uproot our hopes.  What's worse, 'we are often confident even when we are wrong,' writes Daniel Kahneman, in his masterful new book on psychology and economics called 'Thinking, Fast and Slow.'  An objective observer, he writes, 'is more likely to detect our errors than we are.'"

I've read no more than the first couple of chapters of Kahneman's book (courtesy of Amazon Kindle samples), so I don't know what he concludes as a solution to the problem posed above--that we are deceived by our own inner brain processes.  However, my intuitive reaction to Berman's solution was visceral: how can he possibly suggest that the objective observer advocated by Kahneman could be provided by analytics over big data sets?  In truth, the error Berman makes is blatantly obvious in the title of the article... it always is somebody's algorithm.

]]> http://www.b-eye-network.com/blogs/devlin/archives/2012/01/big_data_big_mi_1.php http://www.b-eye-network.com/blogs/devlin/archives/2012/01/big_data_big_mi_1.php Mon, 16 Jan 2012 08:28:55 -0700 Now, I may be accused of getting up on my soap box in this first post of 2012, but... a few recent articles on the topic of big data / predictive analytics have really got me thinking.  Well, worrying, to be more precise.  My worry is that there seems to be a growing belief in the somehow magical properties of big data and a corresponding deification of those on the leading edge of working with big data and predictive analytics.  What's going on?

The first article I came across was "So, What's Your Algorithm?" by Dennis Berman in the Wall Street Journal.  He wrote on January 4th, "We are ruined by our own biases. When making decisions, we see what we want, ignore probabilities, and minimize risks that uproot our hopes.  What's worse, 'we are often confident even when we are wrong,' writes Daniel Kahneman, in his masterful new book on psychology and economics called 'Thinking, Fast and Slow.'  An objective observer, he writes, 'is more likely to detect our errors than we are.'"

I've read no more than the first couple of chapters of Kahneman's book (courtesy of Amazon Kindle samples), so I don't know what he concludes as a solution to the problem posed above--that we are deceived by our own inner brain processes.  However, my intuitive reaction to Berman's solution was visceral: how can he possibly suggest that the objective observer advocated by Kahneman could be provided by analytics over big data sets?  In truth, the error Berman makes is blatantly obvious in the title of the article... it always is somebody's algorithm.

]]> http://www.b-eye-network.com/blogs/devlin/archives/2012/01/big_data_big_mi.php http://www.b-eye-network.com/blogs/devlin/archives/2012/01/big_data_big_mi.php Big data Mon, 16 Jan 2012 08:28:55 -0700 It's that time of year when every analyst worth his or her salt is making predictions for the coming year.  Acquisitions.  Big data.  Mobile BI.  Cloud.  Social media.  Predictive analytics... hey! Wait a minute!

My question is: how many of these predictions about BI 2012 are based on the use of predictive analytics?  My hunch is... none.  Perhaps I'm being unfair?  Is it predictive analytics to use all those surveys of buying intentions as input?  What about using trend numbers for market share over the past few years? 

So, here is the Wikipedia definition: "Predictive analytics is an area of statistical analysis that deals with extracting information from data and using it to predict future trends and behavior patterns. The core of predictive analytics relies on capturing relationships between explanatory variables and the predicted variables from past occurrences, and exploiting it to predict future outcomes."  What do you think?  How is the fit?

]]> http://www.b-eye-network.com/blogs/devlin/archives/2011/12/bi_2012_predict.php http://www.b-eye-network.com/blogs/devlin/archives/2011/12/bi_2012_predict.php Mon, 19 Dec 2011 10:15:28 -0700 It's that time of year when every analyst worth his or her salt is making predictions for the coming year.  Acquisitions.  Big data.  Mobile BI.  Cloud.  Social media.  Predictive analytics... hey! Wait a minute!

My question is: how many of these predictions about BI 2012 are based on the use of predictive analytics?  My hunch is... none.  Perhaps I'm being unfair?  Is it predictive analytics to use all those surveys of buying intentions as input?  What about using trend numbers for market share over the past few years? 

So, here is the Wikipedia definition: "Predictive analytics is an area of statistical analysis that deals with extracting information from data and using it to predict future trends and behavior patterns. The core of predictive analytics relies on capturing relationships between explanatory variables and the predicted variables from past occurrences, and exploiting it to predict future outcomes."  What do you think?  How is the fit?

]]> http://www.b-eye-network.com/blogs/devlin/archives/2011/12/bi_2012_predictions_-_no_way.php http://www.b-eye-network.com/blogs/devlin/archives/2011/12/bi_2012_predictions_-_no_way.php Business intelligence Mon, 19 Dec 2011 10:15:28 -0700 I wrote a White Paper for an interesting, UK-based start-up, NeutrinoBI, back in October on the topic of freeform search in BI.  So, a new paper by Marti A. Hearst, "'Natural' Search User Interfaces", in the November issue of "Communications of the ACM" caught my attention.  I was particularly interested because Hearst has been one of the main proponents of faceted search, an approach that is relatively unsuccessful in BI.  I wondered if I had missed some new developments in the field.

]]> http://www.b-eye-network.com/blogs/devlin/archives/2011/11/freeform_search_1.php http://www.b-eye-network.com/blogs/devlin/archives/2011/11/freeform_search_1.php Wed, 30 Nov 2011 05:10:44 -0700 I wrote a White Paper for an interesting, UK-based start-up, NeutrinoBI, back in October on the topic of freeform search in BI.  So, a new paper by Marti A. Hearst, "'Natural' Search User Interfaces", in the November issue of "Communications of the ACM" caught my attention.  I was particularly interested because Hearst has been one of the main proponents of faceted search, an approach that is relatively unsuccessful in BI.  I wondered if I had missed some new developments in the field.

]]> http://www.b-eye-network.com/blogs/devlin/archives/2011/11/freeform_search.php http://www.b-eye-network.com/blogs/devlin/archives/2011/11/freeform_search.php Business intelligence Wed, 30 Nov 2011 05:10:44 -0700 BI Excellence and Innovation Awards (closing date for entries is 30 November), I'd like to share with you a novel approach to making your entry stand out.  For those of you not eligible to enter, the approach also works when trying to raise funding from the business for a new BI project.  However, you may need to be very brave, even foolhardy; sometimes, you need to tell the as-is horror story to prove just how much better the to-be situation will be.

This story that I'm about to summarize comes courtesy of Teradata's Bill Franks, who posted it about 10 days ago on the Smart Data Collective blog, where it's attracting lots of attention.  Entitled "The Dire Consequences of Analytics Gone Wrong: Ruining Kids' Futures", Bill recounts the story of a local school who invested for the first time in an analytics package designed (allegedly) to detect cheating in English essays.  The package was run for the first time against a set of essays submitted at the start of term by a class of highly motivated and high performing students.  The package promptly reported that each and every student was a cheater, with pervasive copying and plagiarism throughout the group.  The school failed all of the students on the assignment and was about to note the offense on the students' records, an action that would have had severe consequences for their future educational chances, until the parents stepped in...

I leave you to check the rest of the story on Bill's post.  But the bottom line was that the school and the teachers trusted the results of the package more than their own prior experience of the students.  A result of stunning implausibility from the software was accepted without question, without any resort to reason or common sense.  Apparently, the school backed down on noting the offense on the students' records; but it stood by the decision to fail every student in the class on the essay.

As BI practitioners, I trust you get the message.  Analytics in the hands of naive users can be more dangerous than a Kalashnikov.  Self-service BI may speed up delivery, but how reliable are the conclusions?  There's a good reason why assault rifles are not available on open supermarket shelves (in most countries!).  I'm not against self-service BI, but I do have a problem with willful and uneducated business users.  BI must complement common sense, not to substitute for it.  It's up to IT to ensure the users are informed of the strengths and weaknesses of new tooling.

To return to your entries for the BI Awards, to be presented at the BI Summit in Johannesburg on 28 February next, I hope you don't have a horror story of such magnitude, and if you did, I imagine your CEO would be reluctant to have it told in public.  But do remember that the difference between the before and after pictures is a strong indication to the judges of the value of the project to the business.  BI Excellence, beyond technical architecture, also includes data governance and user education.  BI Innovation is about changing the way users make business decisions.  Good decisions based on reliable information. ]]> http://www.b-eye-network.com/blogs/devlin/archives/2011/11/bi_excellence_o_1.php http://www.b-eye-network.com/blogs/devlin/archives/2011/11/bi_excellence_o_1.php Mon, 21 Nov 2011 05:32:35 -0700