The self(less) publisher what can game theory teach us about open data
Download as PPTX, PDF•0 likes•168 views
This document outlines a training course on developing an effective open data strategy. The course aims to explain the importance of data strategy and common pitfalls organizations face without one. It discusses how to set a vision for data use, clearly state the value of data, and create simple actions through a data strategy. The document provides examples of data strategy traps like being siloed, resistant to change, or only focusing on limited data types. It also presents frameworks for developing data literacy at different organizational levels and integrating a strategy. Finally, the document discusses using incentives to encourage adoption of an open data strategy and has participants play a contribution game to experience different incentive structures.
![literacydevelopment oversightintegration
The keys to putting [open] into data strategy](https://image.slidesharecdn.com/theselflesspublisher-whatcangametheoryteachusaboutopendata-161104101453/85/The-self-less-publisher-what-can-game-theory-teach-us-about-open-data-31-320.jpg)
![Exercise
Use the spectrum to map 5 datasets from your own
organisation [1 for each category]
15 minutes in groups](https://image.slidesharecdn.com/theselflesspublisher-whatcangametheoryteachusaboutopendata-161104101453/85/The-self-less-publisher-what-can-game-theory-teach-us-about-open-data-51-320.jpg)
The self(less) publisher what can game theory teach us about open data
- 1. Benjamin Cave – Trainer @cave_ben4 November, 2016 The Self(less) Publisher
- 2. Enable participants to understand the the importance of open data strategy Course aim
- 3. Explain the key pitfalls of organisations without a data strategy List the key characteristics of an effective open data strategy Understand the four keys open data strategy on the organisation Solve a simple open data strategy challenge by applying game theory principles Outcomes
- 4. Data strategy as roads
- 5. We build roads to connect places we value We build data to get to decisions we value
- 6. https://www.flickr.com/photos/ronsaunders47/7459822850 Which is sometimes all you need…Without strategy, roads all look like this…
- 7. https://www.flickr.com/photos/highwaysagency/6194409693 But most times we need strategy to ensure our roads make sense together…
- 8. When data strategy is optimised friction is reduced
- 11. What is data strategy? 5 minutes in groups
- 12. “A single, unified, organisation-wide plan for the use of corporate data” Credit: Capstone
- 13. “A roadmap and plan to identify what to do with a company’s data and to support accessing, sharing and managing the content,” Credit: SAP
- 14. “A statement of the value the organisation sees in their data and a set of actions to capture that value,” Credit: Benjamin Cave, ODI
- 15. Why is data strategy important? Call-out answers
- 16. Bad data costs the U.S. alone $3.1 trillion annually Credit: IBM
- 17. Workers in knowledge & data roles spend 50% of their time correcting and testing data Credit: Thomas Redman
- 18. Costs roughly 10 times as much to work from incorrect data as from perfect data Credit: Rule of 10, Thomas Redman
- 19. Without data strategy the organisation can fall into one of these five traps…
- 20. Trap 1: Siloed
- 21. “We don’t know who holds what” Siloes companies don’t know where data is stored and, in some cases, how it is used. Different parts of the business use their own data for their own goals with very little sharing
- 23. “That’s not the way I work” Resistant companies have staff who refuse, deliberately or otherwise, to tow the line when it comes to data management & use. Different people pursue their own goals regardless of the strategy.
- 25. “This time lets do it this way” Piecemeal companies lack a central data strategy. They have many data strategies for different projects or data types and each new area is set up without reference to the others.
- 26. Trap 4: Legacy
- 27. “The system means we have to do it this way” Legacy companies have strategies designed around the constraints of their systems or workforce skills. Even where the goals of the business are different, strategy is led by resources rather than objectives.
- 28. Trap 5: Focussed
- 29. “That data isn't relevant to us” The focus company only cares about data that serves a single or limited set of objectives. They will discard or neglect any data not directly relevant to their focus area(s).
- 30. So strategy is important… How does it get made?
- 31. literacydevelopment oversightintegration The keys to putting [open] into data strategy
- 32. development
- 33. 1. Set the vision for data 2. State the value clearly 3. Create simple actions
- 34. 1. Set the vision for data
- 35. Visions don’t have to be complex… just clear
- 36. Exercise With a partner, write a 1-sentence vision for how your company will ideally use data 5 minutes in pairs
- 37. 2. State the value clearly
- 38. Group Discussion Data strategy is about a clear statement of the value you expect data to bring to the business Generate a list of types of value data can bring to your business 15 minutes in groups
- 39. 3. Create simple actions
- 40. Exercise What are the main elements of a data strategy ? 5 minutes in groups
- 41. Map Describe Maintain Share Govern Find out who holds what and where it is Ensure all data is well described and can be searched Schedule the correct support to ensure consistent access Create access arrangements that maximise value & respect privacy Integrate data into reporting structures & management priorities
- 42. Data literacy
- 43. 1. Levels Everyone What is data and why it matters to the organisation Specialists How to analyse, store and work with data effectively Managers How to manage work with data and achieve strategic goals Visionaries How to deliver impact and innovation
- 44. 2. Domains Arrangement Visualise how skills can be built up to develop deep understanding & confident application Domains Break down levels into priority skillsets
- 45. 3. Topics
- 46. Using the framework 1. Classify level of pre-existing data literacy 2. Determine target level based on role 3. Select relevant domains based on function 4. Pick topics based on interests For best results, topics should be adjacent.
- 47. Exercise Use the framework to classify the best skills for our four personas 20 minutes in groups
- 48. Brian - administrative assistant Brian doesn't work with data. He doesn’t understand why the organisation has a data strategy. Brian also has a career interest in business management Nita – customer insights manager Nita looks after the customer insight team. Her staff analyse large volumes of data. Nita wants to increase productivity and monitor results. Elise- market analyst Elise produces market models for company strategy. Her job involves frequent data interaction. Elise is interested in improving her sources of data and range of analytic techniques Terrance – head of partnerships Terrance heads up strategic partnerships. His role involves regular discussion of data sharing arrangements. Terrance is also interested in the business potential of open data Personas for diagnosis
- 49. integration
- 50. Integrating strategy 1. Take a phased approach based on quick wins 2. Start with low resistance areas 3. Ensure buy-in at team level 4. Prioritise comfort over scale 5. Capture evidence for iterative improvements
- 51. Exercise Use the spectrum to map 5 datasets from your own organisation [1 for each category] 15 minutes in groups
- 53. oversight
- 54. Why create incentives? Open data benefits the organisation but some people will lose out Those who resist perceive the current state as ‘pareto optimal’; any benefit to one person bringing at least equal harm to others. Incentives allow rational actors to pursue their interests while contributing to the goals of your open data strategy
- 55. Incentives to implement A key driver of positive strategy adoption is strong, well-aligned incentives Different functions within the organisation respond to different incentives Finding the right incentives to overcome differences, harness competition and encourage collaboration is a priority for strategy architects
- 56. Exercise Play the contribution game 20 minutes
- 58. Rules Every player starts with 100 pounds Each round the player decides how much to contribute to a ‘public pot’ between 0 and 100 pounds The public pot pays everyone double the average contribution of all players Every player ends the round with their share of the public pot plus anything they did not contribute
- 59. Example 3 players Player 1 gives 60 pounds, player 2 gives 40 pounds and player 3 gives 20 pounds The average contribution to the public pot is therefore 40 pounds Each player receives 2*40 pounds from the public pot to add to their leftover money
- 60. Round 1 – No incentive Budget: 100 pounds Reward: 2 * average contribution http://goo.gl/forms/Nm5B2aAwd7xJzb753
- 61. Round 2 – Individual punishment Budget: 100 pounds Return: 2 * average contribution Punishment: Lowest balance eliminated http://goo.gl/forms/vANFi1GKsXUNgfHy2
- 62. Round 3 – Collective reward Budget: 100 pounds Return: 2 * average contribution Reward: If the average end balance > 150 pounds, everyone gets a sweet http://goo.gl/forms/RHtk4xX56B0Oi3203
- 63. Round 4 – Race to the middle Budget: £100 Reward: 2 * average contribution Present: Closest contribution(s) to average get 10 sweets
- 64. Round 5 – Observation principle Budget: 100 pounds Reward: 2 * average contribution Visibility: All contributions made public http://goo.gl/forms/6Iu2Eu6vIntkTdBG3
- 65. Strategic Incentives Incentive Metric Pros Cons Overall Quantity Numeric Target Collective action, scale of result Free-rider problem, less useful data Selected Quantity List of Targets Useful data, equal contributions Time-intensive, unit resistance Competitive Reward Number from Unit Competitive drive, broad engagement Less useful data, penalise success Competitive Sanction Number from Unit Overcome apathy, broad participation Race to the middle, Hostile environment Impact Reward Qualitative Maximise results, focus resources Free-rider problem, inconsistent
- 66. Benjamin Cave – Trainer @cave_ben4 November, 2016 Thank-you