A Presenta on from Big Data


[PDF]A Presenta on from Big Data - Rackcdn.comb219de03d4fa586d9adb-0e04f025bcde650ce6d504251ebbc914.r60.cf3.rackcdn.co...

0 downloads 130 Views 2MB Size

A  Presenta*on  from   Big  Data   22  February  2013  

Does Big Data = Big Insight?  

Colin  Strong   GfK  

All  copyright  owned  by  The  Future  Place  and  the  presenters  of  the  material   For  more  informa:on  about  NewMR  events  visit  NewMR.org  

Does Big Data = Big Insight? Colin Strong MD Technology GfK UK

Colin Strong, GfK, UK Big Data, 22 February 2013

Big Data’s meteoric rise •  Internet of things growing availability of Big Data •  Expectations of Big Data high –  twice as likely to substantially outperform their industry peers –  big data expected to improve organisational performance by 41% over the next three years

•  Should we be sceptical? •  Does Big Data always ‘work’?

Colin Strong, GfK, UK Big Data, 22 February 2013

Big Data Assumptions •  Big Data reflects our inner lives •  We only need to observe to understand •  Big Data remains inside the firewall •  The individual is the only valid unit of measurement

Colin Strong, GfK, UK Big Data, 22 February 2013

The pitfalls •  John Ioannidis: ‘Most published research findings are false’ •  HBR: ‘There’s an odds-on chance that someone in your organization is making a poor decision on the basis of information that was enormously expensive to collect’ •  Nate Silver: ‘Our predictions may be more prone to failure in the era of Big Data’

Colin Strong, GfK, UK Big Data, 22 February 2013

Think like a Bayesian •  Take into account the context through prior probability •  Consider for level of false positives •  Probabilistic outcomes •  Integrate multiple data sources

Colin Strong, GfK, UK Big Data, 22 February 2013

How to do the numbers: Prior  probability   Ini:al  es:mate  of  service  being  successful  

X  

4%  

Consumer  survey  indicates  50%  consider  service  will  be   success  

Y  

50%  

Probability  of  survey  incorrectly  indica:ng  success  

Z  

5%  

New  informa*on  is  delivered:  

Posterior  probability   Revised  es:mate  of  likelihood  of  success  

Colin Strong, GfK, UK Big Data, 22 February 2013

xy                     29%   Xy+z(1-­‐x)  

A manifesto for Smart Data

Colin Strong, GfK, UK Big Data, 22 February 2013

1: ‘Individual’ case study   Use digital data to collect information concerning online behavior   Integrate this with survey data to understand the ‘why’   Used successfully in purchase journey analysis

Colin Strong, GfK, UK Big Data, 22 February 2013

2: ‘Segment’ case study   Conducted segmentation of BT prospects database   Overlaid this with database of new customers   Integrated with the tracking studies to model the segmentation onto survey data   Created a propensity model to facilitate targeting

Colin Strong, GfK, UK Big Data, 22 February 2013

3: ‘Social’ case study   Big Data allows us to start exploring relationships between individuals   Allows us to start exploring different types of real-life social networks that have proved elusive   Big Data provide us with tools that help us to understand the way in which these work   More generally allows us to build our understanding of complex systems Colin Strong, GfK, UK Big Data, 22 February 2013

4: Cultural analytics   Analysis of massive cultural data sets and flows   Publically accessible data sets increasingly digitised allowing much more access and analysis   Ngram a leading tool, groups experiementing with images   Exploring ways in which we can access flows of images   GfK Visual Survey of Domestic Space 12 country, relational database of 18,000 photos

Colin Strong, GfK, UK Big Data, 22 February 2013

Thank you

Colin Strong Twitter: @colinstrong Personal blog: www.colinstrong.net Colin Strong, GfK, UK Big Data, 22 February 2013

Q&A

Colin  Strong   GfK  

Ray  Poynter   Vision  Cri:cal  University  

Colin Strong, GfK, UK Big Data, 22 February 2013