Giving Organisations new Capabilities to ask the Right Business Questions
2 likes1,277 views
The document discusses various approaches organizations can take to gain insights from data. It begins by noting that making data work is difficult and that value is captured through outputs and outcomes. It then describes three common approaches: the "all in" approach of fully committing resources, the experimental approach of running small trials, and the "wait and see" approach. The document advocates for an experimental approach using agile experimentation. It provides examples of areas where organizations need to improve such as asking the right questions, choosing technologies, and interpreting results. Finally, it discusses various analytic methods and structured techniques that can be used, including decomposition and visualization, hypothesis generation and testing, and challenge analysis.
Giving Organisations new Capabilities to ask the Right Business Questions
- 1. Giving Organisations new Capabilities to ask the Right Business Questions Stephen Simpson CTO Office [email protected] @sharplyunclear
- 2. Making Data Work is Hard Value Captured Outputs Outcomes Sales Growth Profit Growth Sales Growth to Existing Customers Product Performance Technology Leadership New Customers Process Improvement Effective Project Execution Processes Inputs Balanced Innovation Portfolio Supportive Strategy, Structure, & Systems Partners’ Value-add Employee Commitment to Innovation Quality of Innovation Pipeline Access to Talent
- 3. The “All In” Approach Ron Johnson, CEO Myron Ullman, CEO
- 4. You start out thinking you have a sales problem but might find it is not really sales but marketing or customer retention... …you could spent a lot of time on analysis that doesn’t lead to solving the right problem.”
- 5. The Experimental Approach "We did a Hadoop trial last year, it didn't go very far because we weren't getting the intelligence out of it that we thought we would. So we are looking at some other initiatives with different vendors this year. "We tried to put three different data sets together, and then tried to see if we could find some causality between the data sets that would gives us intelligence that would allow us to manage our operations better… "Whether that was how we set the trial up or the software I don't know, so we are going to try some different things.”
- 6. The “Wait and See” Approach Incumbents are rarely disrupted by new technologies they can't catch up to, but instead by new business models they can't match. Institutions will try to preserve the problem to which they are the solution.
- 7. Satisficing Can rarely evaluate all outcomes with sufficient precision Usually don’t know relevant probabilities of outcomes Possess limited memory
- 8. Results are often Modest
- 9. Obtaining new insights Business Strategy We need to make sure that we’re asking people to research the right questions Domain Expertise Company Systems & Data And then we iterate to improve the insight gained, or address the next business question… Data Mining Agile Experimentation We need to choose the right storage technologies, integration services & architecture Sourcing We need to look in many more places to find data… Extraction …and it will take a lot of different skills and approaches to bring it together We need to perform analysis quickly inside small projects, with a specific business goal. Some of these will fail. We need to be careful to curb our enthusiasm and separate out the signal from the noise Interpretation Implementation We need simple, easy to use production tools to act upon the new insights. Authority needs to be delegated to where the information is captured Visualisation We need new techniques to interpret and manipulate vast numbers of data points on a single surface
- 11. Candidate Sources CRISP-DM Richards J. Heuer & Randolph H. Pherson
- 12. Analytic Methods Decomposition & Visualisation Idea Generation Expert Judgment Scenarios & Indicators Quantitative Methods using Expert-Generated Data Quantitative Methods using Empirical Data Hypothesis Generation & Testing Assessment of Cause & Effect Challenge Analysis Conflict Management Structured Analysis Decision Support
- 13. 13
- 14. Structured Analysis a step by step process for analyzing the kind of incomplete, ambiguous and sometimes deceptive information that analysts must deal with.
- 15. Structured Analytic Techniques contain Diagnostic + Contrarian + Imagination elements 16
- 16. The Techniques
- 17. Choosing what you want to do 1. Define the project? 2. Get started? Decomposition & Visualisation Decomposition & Visualisation 3. Examine & make sense of the data? Figure out what is going on? Idea Generation Scenarios & Indicators 4. Assess the most likely outcome of an evolving situation? 5. Monitor a situation to avoid surprise? 6. Generate and test hypotheses? Hypothesis Generation & Testing Assessment of Cause & Effect 7. Assess the possibility of deception? 8. Foresee the future? 9. Challenge your own mental model? Challenge Analysis Conflict Management 10. See events from the perspective of other players? 11. Managing conflicting mental models or opinions? 12. Support a manager in deciding course of action? Decision Support
- 19. Template Structure Overview When to Use It Value Added The Method Relationship to other Techniques Origins of this Technique 20
- 20. Long term personal healthcare Branded Currency Personalised Interactions 21
- 21. 1. Decomposition & Visualisation When forced to work within a strict framework the imagination is taxed to its utmost – and will produce its richest ideas. Given freedom the work is likely to sprawl.
- 23. Value Proposition Understand your clients’ needs at the finest level of detail Client micro-segmentation using multiple sources of data Description FOR marketing operations WHO want to understand the growth potential for each identified customer subdivision THE understand your clients’ needs at the finest level of detail solution PROVIDES understanding of the root causes for your current share of each identified slice THAT lets you act on the information quickly with targeted retail product placement & location selling UNLIKE your existing solution WHICH is coarse-grained and retrospective Scenarios • • • • • Retail product placement & location selling Counteracting effectiveness of competitors Understanding local reputation via ”voice of the customer” Real-time decision making such as mobile-based coupon positioning to particular segments Partner organisations’ service effectiveness
- 24. 2. Idea Generation The best way to have a good idea is to have a lot of ideas
- 26. Creativity Value Creation Out-of-box thinking In-the-box thinking Raw & refined ideas Experimentation Engineering/process improvement Ambiguity/uncertainty Precision Research Well-calculated trade-offs Intuition Buying/selling of ideas Surprise Do things right Courage Answer questions & verify solutions Find the right things Ask questions & explore unknown innovation Seize opportunities Visualize future & consider all options Include incremental & radical ideas Avoid major risks Get product into the marketplace Bias for incremental
- 27. Cross Impact Matrix For when “Everything is connected to everything else” Business is in flux Context for discussion of interactions System is stable Discover variables once thought to be simple - Need to identify and monitor all factors that might upset this A significant event has occurred - Need to understand implications & independent are actually interrelated Focus on - Interactions that may have been overlooked - Variables that might reinforce each other
- 28. Cross Impact Matrix A B C C. Existing core banking solutions D. Apps & Cloud Service interaction E F ++ ++ A. Personalised Interactions B. Existing mobile solutions D --- ++ + - E. Offers ++ F. Analytics ++ - + + ++ ++ ++
- 29. 3. Scenarios & Indicators Scenarios are plausible & provocative stories about how the future might unfold
- 31. Indicators Observable Phenomena that can periodically be reviewed to help track events Make humans recognize early signs significant change Spot emerging trends Quality indicators are critical - If narrowly defined or out of date - Reinforce bias - Warn unanticipated changes - Discard new evidence - Avoid surprise - Lull people inappropriately Forward looking, predictive Objective baseline for tracking - Dashboards… Indicators Validator Instil rigour into analytic process - Quality and strength of indicator Enhance credibility of what delivered - Whether appears in all scenarios Exchange knowledge between experts from different domains
- 32. Indicators 2013 Q4 Q3 Mobile Offers Reaching right segment People engaged Volunteering information Infrastructure Holding initiative back Cloud Security, regulatory, compliance Service Take-up standard services 3rd party composing new apps Industry Trends Personalised CRM Branded Currency Device as Bank Ecosystem Retailers using your backbone Competitive launches Q1 □ ● ▫ □ ▫ ▫ □ ○ ○ ○ Q2 ● ▪ □ ▫ ▫ Q1 ● ▪ □ ● ▫ Q4 2015 Q3 □ ● □ ● □ □ Q2 2014 Q3 Neglible concern Low concern Moderate Substantial Strong ▫ ▪ □ ○ ● Q4
- 33. 4. Hypothesis Generation & Testing A possible explanation of the past or a judgment about the future is a hypothesis that needs to be tested by collecting and presenting evidence
- 35. 5. Assessment of Cause & Effect We are slow to accept the reality of simple mistakes, accidents, unintended consequences, coincidences, or small causes leading to large effects
- 37. Personalised Interactions will increase: Key Assumptions check Legal and privacy – Caveated. Components available across entire chain – Caveated. Customers want seamless, personally relevant services – Solid Devices will progress sufficiently – Solid Analytics techniques are sufficiently refined, accurate and timely – Caveated Back-end systems will support workload – Solid Systems will be cost effective – Caveated. What’s the ROI of something you don’t know? Employees trained and authority delegated to act – Unsupported
- 38. 6. Challenge Analysis It is the mark of an educated mind to be able to entertain a thought without accepting it.
- 40. Pre-mortem analysis Imagine the future where your plan has been implemented, but has failed Advantages: Take people out of perspective of defending their plan & shielding themselves from its flaws Increase level of candour Can be used to show decision makers that are typically over-confident that their decisions and plans will work Questions re-framed, to elicit different responses to original ones Legitimises dissent – asked to make a positive contribution by identifying weaknesses in previous analysis Examples - Internal inertia or uneven execution - Competitors’ actions - Law of unintended consequences - Economic changes
- 42. 7. Conflict Management Disagreements sparked by differences in perspective, competencies, & access to information… actually generate much of the value that can come from collaboration across organisational boundaries. http://hbr.org/2005/03/want-collaboration-accept-and-actively-manage-conflict
- 44. 8. Decision Support …without overstepping the limits of their role…; just structures all the relevant information in a format that makes it easier for the decision maker to make a choice.
- 46. A word on Dashboards It is also unfortunate to see how many business intelligence and enterprise data warehousing projects get waylaid by the singular pursuit of pretty dashboards…
- 48. Time is Key
- 49. Self-Inflicted Complexity When we sacrifice dealing with detail complexity to focus on dynamic complexity, the solutions don’t produce the outcomes that we really want. http://blogs.hbr.org/2013/09/our-self-inflicted-complexity/
- 50. In Summary Does provide new capabilities to ask right questions - Offers path to clearer business goals - Discourages “wait and see” approaches Encourages cross-organisational linkages Validates or challenges experts’ “hunches” More limited use in monitoring subsequent change 12