Peter Rakers De Verstoring van Big Data Na globalisatie en digitalisatie komt dataficatie, het continue genereren van real time gegevens. Voor veel bedrijven werkt dit fenomeen eerder verstorend. Peter Rakers geeft voorbeelden over de opportuniteiten en valstrikken van big data. meer weten over Peter Rakers: www.cropland.be
Disruptive in Belgium Seminar @ Gent, 26 september 2014 Peter Rakers
Data has become raw material that can be used in an innovative way in order to get new products, services or insights with significant added value. 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 3
Cropland will look at your (big) data business case from 4 different angles Customers Products Team Channels 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 4
and on 3 different levels LEVEL 1 INTERNAL DATA ANALYSIS Although obvious at first sight, connecting system A with system B or exchanging marketing information with financial (sales) data or consumer data appear not to be easy in practice. During the start up of a Cropland project, we always analyze the internal data at first in order to take out maximum information from structured sources. 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 5
and on 3 different levels LEVEL 2 THE COMBINATION WITH EXTERNAL DATA After mapping the internal data, we look for associations and correlations with external sources. Free data, but also paid information from other stakeholders or data providers are combined. Unexpected events such as weather, demographic patterns or others could have an important influence on the company s objectives. 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 6
and on 3 different levels LEVEL 3 AN INTERCONNECTED NETWORK OF COMPLEMENTARY BUSINESSES Without exposing company-owned information, Cropland will look at innovative segmentation models based on a concept called the DATA POOL. 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 7
What is foresight? Level of abstrac ction Short term Qualitative Long term Qualitative Short term Quantitative Long term Quantitative TODAY Time into the future 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 8
For today, we can call it Level of abstrac ction War gaming g Scenarios Thinking Forecasting Predictive Analytics TODAY Time into the future 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 9
What is surprising is not the magnitude of our forecast errors, but our absence of awareness of it Nassim Nicholas Taleb Author of 'The Black Swan 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 10
Human beings and organizations do not act in response to reality but to an internally constructed version of reality Kees van der Heijden Strategic Planning Royal Dutch / Shell 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 11
And that s why I believe, That seemingly gyrandom events have potential to become disruptive for others Mohamed Bouazizi, a Tunisian vegetable seller, set himself alight on December 17, 2010 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 12
Topic of the day Scenarios Thinking Predictive Analytics 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 13
Big data definition Big Data usually includes data sets with sizes beyond the ability of commonly used software tools to capture, organize, manage, and process the data within a tolerable elapsed time. Big data sizes are a constantly moving target, as of 2012 ranging from a few dozen terabytes to many petabytes of data in a single data set. kb (kilobyte) = 1000 1 =10 3 MB (megabyte) = 1000 2 =10 6 GB (gigabyte) g = 1000 3 =10 9 TB (terabyte) = 1000 4 =10 12 PB (petabyte) = 1000 5 =10 15 EB (exabyte) = 1000 6 = 10 18 ZB (zettabyte) = 1000 7 =10 21 YB (yottabyte) = 1000 8 =10 24 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 14
Big data definition Amount from big to bigger Range from structured to unstructured Speed from batch to streaming Process from Excel to Hadoop from IT/BI to data science 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 15
Big data definition Business Intelligence: Descriptive statistics, trends, linear (structured) data Big Data Inductive statistics, laws, nonlinear (unstructured) systems 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 16
Big data mind shift The ability to work directly with vast amounts of data (N =all) instead of extrapolating smaller sets The willingness to embrace real-world messiness. Only 5% of all available data is structured, meaning in a form that fits neatly into a traditional data base The fact that by machine learning ability, correlations become more powerful over time. N=all time 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 17
Post hoc, ergo propter hoc After this, therefore because of this or correlation is not causality More trumps better The loss in accuracy at micro level is largely compensated by the insights we get at macro level New data scientists and strategists are needed to cope with this paradigm shift Most of our institutions were established under the presumption that human decisions are based on information that is small, exact and causal in nature. More trumps better. The obsession with exactness is an artifact of the informationdeprived analog era. -Viktor Mayer Schönberger, Big Data 2013-2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 18
Big data predictive analytics Data always speaks. It always has a story to tell and there's always something to learn from it. Data scientists t see thisover and over again across PA (predictive analytics) projects. Pull some data together and, although you can never be certain what you'll find, you can be sure you'll discover valuable connections by decoding the language g itspeaks and listening. That's the Data Effect - Eric Siegel, Predictive Analytics 2013-2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 19
Predictive analytics Most data is not accumulated for the purpose of prediction but PA can and will learn from this massive recording of events the same way as human beings learn from the accumulation of life experience. 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 20
Consequences for the public There ain t no such thing as a free lunch Robert Heinlein: The Moon is a Harsh Mistress, SF novel, 1966 according for free for free for free for free for free for free for free for free 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 21
Today Recent mood analysis of 2013 by Twitter (Newsmonkey, March 13, 2014) 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 22
Today Prof. Veronique Hoste of UGent investigated the content of Twitter messages on the risk of suicide and found several key words with significant predictive signaling. The Bringham Young University connected Twitter messages with HIV sexual risk behaviour in well specified areas of the US 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 23
Today DANIEL KATZ, MICHIGAN STATE UNIVERSITY 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 24
Big data adoption position 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 25
Big Data adoption position Core business BLINDSIGHTVICTIM You are attacked on your core activities by players who don t follow the established rules of the game Late BIG DATA B I G D A T A DOPTION FORESIGHT DISRUPTOR As market leader or innovative challenger, you adopted big data opportunities early in your strategy to stay in or enter the game Early I HINDSIGHT SURVIVOR M P INSIGHT EXPERT A You are at the end of your marketing era. C You keep your core activities up to speed Your expertise declines in valued as the T and in parallel, you develop new services game is changing gquickly based on big data insights Parallel business
New kids on the block Core business B I G late D A T A I M P A C T Parallel business early
Conclusion for now Big data is here and it is here to stay It will become an (new) strategic tool to look at the future Access to big data (in real time) offers an important knowledge and decision taking advance It deals with larger sense of group behavior but business derivates focus on an individual approach Companies with datafication of human behavior as core business (e.g. Google, Amazon, Facebook) will only increase their dominance on the global market The next step is that companies can not only predict but also prescibe our behavior without us being aware And so, although we experience mainly the benefits of big data today, there s an important risk (or dark side) in the further evolution of datafication for individuals 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 28
The future for now 2014 Cropland: Disruptive in Belgium - 10 jaar SC, Gent, Sept 26, 2014 29