Download as PDF, PPTX
Uploaded byLars Trieloff
Business Reasons for Predictive Applications
Automated decision-making in business is faster, cheaper, and more effective, addressing the complexities and inefficiencies of human decision-making. With 99.9% of business decisions potentially automatable, leveraging data through machine learning can enhance predictive capabilities and overall decision quality. The adoption of predictive applications is essential for optimizing business operations and financial outcomes.
More Related Content
Similar to Business Reasons for Predictive Applications
Business Reasons for Predictive Applications
- 2. Business Reasons for AutomatedDecisions Lars Trieloff | @trieloff
- 3. • They arefaster • They are cheaper • They are better • That’s it. Thank you for your attention. Why Automated Decisions make Business Sense
- 4. 4%Worldwide average profit marginin retail: 4%
- 5. 4‰German average profitmargin in retail: 4‰
- 6. Your Customer givesyou this
- 7. All you gotto keep is that
- 8. — –Libby Rittenberg “Economicprofits in a system of perfectly competitive markets will, in the long run, be driven to zero in all industries.”
- 9.
- 10. — Abraham Maslov– probably never said this. It’s true anyway. “Data has Human Needs, too”
- 11.
- 12.
- 14. — W. EdwardDeming “In God we trust, all others bring data”
- 15. How Data-Driven Decisions shouldwork Computer Collects Computer Stores Human Analyzes Human Predicts Human Decides
- 16. — Daniel Kahneman “Prejudiceagainst algorithms is magnified when the decisions are consequential.”
- 17. How Data-Driven Decisions REALLYwork Computer Collects Computer Stores Human Analyzes C O M M U N I C AT I O N B R E A K D O W N Human Decides
- 18. — Led Zeppelin CommunicationBreakdown, It's always the same, I'm having a nervous breakdown, Drive me insane!
- 19. • Drill-down analysis… misunderstood or distorted • Metrics dashboards … contradictory and confusing • Monthly reports … ignored after two iterations • In-house analyst teams … overworked and powerless How Data-Driven Decisions REALLY work CO M M U N I C AT I O N B R E A K D O W N
- 20. How Decisions REALLYshould work Computer Collects Computer Stores Computer Analyzes Computer Predicts CO M P U T E R D E C I D E S
- 21. — Everyone atBlue Yonder, all the time 99.9% of all business decisions can be automated
- 22. How Decisions areBeing Made
- 23. 90% No Decisionis made
- 24. — Robin Sharma “Makingno decision is a decision. To do nothing. And nothing always brings you nowhere..”
- 25. Business Rules forBeginners Not doing anything is the simplest business rule in the world – and also the most popular
- 26. 90% No Decisionis made
- 27.
- 28. Advanced Business Rules Computersare machines following rules. This means business rules are programs.
- 29. • Business rulesare like programs – written by non-programmers • Business rules can be contradictory, incomplete, and complex beyond comprehension • Business rules have no built-in feedback mechanism:“It is the rule, because it is the rule” Business rules are Programs, just not very good ones.
- 30. 1% Human Decisionmaking
- 31. Human Decision Makinghas two systems – and only one is rational.
- 32. Not quite Almostthere That’s it.
- 33. Quick: What doyou see here?
- 35. — Steven Pinker,describing Moravec’s Paradox “The hard problems are easy and the easy problems are hard.”
- 36. Quick: Add alleven numbers 65 7 1 0 60 63 18 80 547039100
- 37. 94 39 3731 92 70 100 67 4956080 69 20 26 73
- 38. 51 60 2322 5 48 43 14 9525669 23 67 1 43
- 39.
- 40.
- 42. — Daniel Kahneman “Allof us would be better investors if we just made fewer decisions.”
- 44. How we aremaking decisions (Like the big apes we are) Anchoring effect IKEA effect Confirmation bias Bandwagon effect Substitution Availability heuristic Texas Sharpshooter Fallacy Rhyme as reason effect Over-justification effect Zero-risk bias Framing effect Illusory correlation Sunk cost fallacy Overconfidence Outcome bias Inattentional Blindness Benjamin Franklin effect Hindsight bias Gambler’s fallacy Anecdotal evidence Negativity bias Loss aversion Backfire effect
- 46. K-Means Clustering Naive Bayes SupportVector Machines Affinity Propagation Least Angle Regression Nearest Neighbors Decision Trees Markov Chain Monte Carlo Spectral clustering Restricted Bolzmann Machines Logistic Regression Computers making decisions (cold, fast, cheap, rational)
- 47. • A machinelearning algorithm is a system that derives a set of rules based on a set of data • It is based on systematic observation, double- checking and cross-validation • There is no magic, just data – and without data there is no magic either Machine Learning means Programs that write Programs
- 48. Better Decisions through PredictiveApplications
- 49. How Predictive Applications Work Collect& Store Analyze Correlations Build Decision Model Decide & Test Optimize
- 50. — Warren Buffett “Ichecked the actuarial tables, and the lowest death rate is among six-year-olds, so I decided to eat like a six-year- old.”
- 51. More than halfof the apps on a typical iPhone home screen are predictive applications.
- 52. Building Predictive Applications MachineLearning ModelPredictive Application Enterprise Integration
- 53. Story Time (Not safefor vegetarians)
- 54. The Ground BeefDilemma
- 55. How much ground beefare we going to sell on Friday?
- 56. How much groundbeef are we going to sell on Friday? And how much on Saturday?
- 57. Challenge #1 Accuratelypredict demand
- 58. Great. But howmuch do we need to order each day?
- 59. Great. But howmuch do we need to order each day? Let’s reduce the risk of running out of stock to 20%
- 60. Sales Forecasts forFriday SalesProbability 0 0,01 0,02 0,03 0,04 0 4 8 12 16 Friday Sales Amount
- 61. Sales Forecasts forSaturday SalesProbability 0 0,01 0,02 0,03 0,04 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Saturday Sales Amount
- 62. Great. But howmuch do we need to order each day? Let’s reduce the risk of running out of stock to 20% So it’s 3 on Friday and 5,5 on Saturday.
- 63. Sales Forecasts forBoth Days SalesProbability 0 0,01 0,02 0,03 0,04 0 4 8 12 16 Friday Sales Amount Saturday Sales Amount
- 64.
- 65. Bad news… We needto skip the Saturday delivery.
- 66. Bad news… We needto skip the Saturday delivery. How big should we make the Friday delivery instead?
- 67. If you need3 on Friday and 5,5 on Saturday to fulfill 80% of the demand, how much do you need to fulfill 80% of the combined demand?
- 68. 3 + 5,5= 8,5 Common Sense isn’t it?
- 69. — Albert Einstein Commonsense is what tells us the world is flat.
- 70. Combined Sales Forecasts SalesProbability 0 0,01 0,02 0,03 0,04 04 8 12 16 Combined Sales Amount
- 71. If you ordered8,5 cases, you would waste a lot of meat, the ideal order amount is 8 cases.
- 72. Predictive Apps ina Nutshell Batch and streaming data ingestion, batch and streaming delivery (with real-time option) Reduce risk and cost » increase revenue and profit Trend Estimation Classification Event Prediction Optimize Returns Collect Data Predict Results Drive Decisions
- 73. — John MaynardKeynes “When my information changes, I alter my conclusions. What do you do, sir?”
- 74. One Common Platformfor Predictive Applications Multi-Tenant Runtime Environment Link Store Build Run View Link your own and third-party data, easily integrated via API Store your data in high-performance database as a service Build machine learning and application code Automatically run and scale ML models and applications Monitor and inspect resource usage and model quality Secure Micro Cloud Infrastructure Domain Model Predictive Model Application Code
- 75. — Kevin Kelly “Thebusiness plans of the next 10,000 startups are easy to forecast: Take X and add AI”
- 76. How Enterprises adopt PredictiveApplications Learn about ADDD Define Target Process Build Predictive App Go Live Make Lots of Money
- 77. — Daniel Kahneman “Prejudiceagainst algorithms is magnified when the decisions are consequential.”
- 78. How Enterprises REALLYadopt Predictive Applications Learn about ADDD Define Target Process Build Predictive App Make Lots of Money D O U BT S CO N C E R N S O B J E C T I O N S
- 79. Decision Quality Status QuoPredictive Prescriptive Automation
- 80. — The Economist,May 2015 “The best chess players in the world are‘centaurs’, amalgamated teams of humans and algorithms.”
- 81. Decision Quality Status QuoPredictive Prescriptive Automation 'Centaurs'
- 82. Ready, set, gofor ADDD?
- 83.
- 84.
- 85. Financial Impact ofPredictive Apps -120 -90 -60 -30 0 30 60 90 March April May June July August September Break-Even after 3 months
- 86. • They arefaster • They are cheaper • They are better • That’s it. Thank you for your attention. Why Automated Decisions make Business Sense
- 87.