The Human Side of Data

If you like this presentation – show it...

Slide 0

The human side of data Contact: Colin Strong c.strong@addverve.com @colinstrong +44 (0) 7494 529311 October 2015

Slide 1

Is the big data enterprise struggling? Through 2017, 60% of data projects will fail to go beyond piloting and experimentation and will be abandoned Gartner 72% of business and analytics leaders aren’t satisfied with how long it takes to retrieve the insights they need from data Alteryx 90% of digital start-ups fail Mashable Only 27% the executives surveyed described their data initiatives as successful Capgemini 65% of CEOS think their organisation is able to interpret only a small proportion of the information to which they have access The Economist 2

Slide 2

Is the answer technical? Better integration of legacy systems More data scientists Better tools Necessary but not sufficient conditions? 3

Slide 3

Or is it something more fundamental? I have lost count of the times I have been presented with some amazing fact that data has told us through the use of some incredible new technology, to be left thinking “so what?” or “isn’t that obvious”? Caroline Morris Sky IQ 4

Slide 4

Three aspects of humans in data: Data does not speak for itself The human hand in the big data machine Finding the human behind the data 5

Slide 5

Data does not speak for itself

Slide 6

Does big data = big insights? •  More data does not automatically equal more insight •  Statistical value does necessarily reflect real relationships. •  Increase in number of false positives 7

Slide 7

What is significance? •  In 1967 a professor spent much of the year flipping a coin 300,000 times •  Found it came up heads 50.2% of the time •  Deemed this to be ‘statistically significant’ •  John Ioannidis: “Most published research findings are false” “…more likely to happen in fields that chase subtle, complex phenomena” 8

Slide 8

So what do we do? •  Undertaking due diligence such as: •  using ranges rather than directional predictions •  establishing multiple corroboration •  comparing models •  But ultimately it is contextual expertise: “An understanding of consumer behaviour and category expertise is often the litmus test to determine whether a statistically significant result has any real validity.” 9

Slide 9

Otherwise risk of ‘cargo cult analytics’: •  The illusion that something is scientific when it has no basis in fact •  Leads to the use of ‘vanity metrics’ •  Calls for: •  Much closer relationship between data analytics and consumer insights •  A structured approach to framing the questions to ask 10

Slide 10

The human hand in the big data machine

Slide 11

The analytics industry fails to question its assumptions of how humans ‘work’: The big debates VS Are our thoughts dictated by neurons or vice-versa? VS Are our behaviours shaped at an individual level or by our social circumstances? 12

Slide 12

But these assumptions shape our analytics: Our implicit thinking abut how humans operate will influence the way in which we: •  Chose which questions to ask •  Select the data we look at •  Determine how we clean the data •  Shape our choice of analytics •  Influence the way we interpret patterns 13

Slide 13

And a simplistic understanding of peoples’ lives generally does not work: Fitness devices: Owned < 6 months: 88% use ≥ once a week Owned > 6 months: 62% use ≥ once a week 14

Slide 14

The human side to the analytics process: 29 different teams of analysts asked to determine whether soccer refs more likely to give red cards to players with darker skin tones. •  Each team was given an identical dataset. •  21 different sets of variables chosen for analysis. •  Different teams used different statistical models. No surprise that teams came to fundamentally different conclusions 15

Slide 15

And we are all susceptible to biases: Calls for: •  Greater awareness of implicit assumptions about human behaviour •  Use of behavioural science to help frame questions / catch potential biases •  Greater reflexivity in the analytics process 16

Slide 16

Finding the human behind the data

Slide 17

The data analytics environment is fundamentally changing: From: To: Who you are What you are like Where you are now Where you are going What you are doing Why you are doing it Outer lives Inner lives 18

Slide 18

Customer interactions moving from physical to digital: §  More of our lives played out online – revealing more about ourselves §  Facebook Likes: Highly predictive of wide range of human attributes where Like did not necessarily explicitly referenced the outcomes “The present time is a very special time in the history of social science because we are witnessing a dramatic transformation in our ability to observe and understand human behaviour.” Duncan Watts, Microsoft §  Opportunity lies in exploring the human characteristics that can be derived from corporate data 19

Slide 19

Offers a real opportunity to optimise the customer experience: •  Identify the key psychological attributes that could make a difference •  Develop appropriate measurement tools (psychometrics) •  Test impact •  Predict attribute from data to roll-out 20

Slide 20

Analytics is a human endeavour: Companies need to invest in understanding this to really generate value from their data assets

Slide 21

Thank you: Drop me a line @colinstrong c.strong@addverve.com www.addverve.com www.colinstrong.net 22

Slide 22