Please enable JavaScript.
Coggle requires JavaScript to display documents.
Human Obsession with Prediction - Coggle Diagram
Human Obsession with Prediction
Predictive models are the bread-and-butter of analytics
Good prediction needs :
Sound Models
Business apps transform data information knowledge
Apps are “models” translated into software code
– Predictions only as good as managerial insight in underlying models
– Function-specific e.g., marketing, operations, accounting, or finance
•Non-IT managers are the only sources of:
– Things competitively worth predicting (the DV in stats-speak)
– Models in their functional doma
Sound Data
Data: Raw facts
– Ex: 24, 32, and 28 are data
• Information: Sorted, condensed, and contextualized data
– e.g., temps in your ZIP code
• Insight: Actionable information
– e.g., need a coat tomorrow
– Unreliable data -> untrustworthy insight
• Collecting and managing data is the IT unit’s job
– But not all data is worth managing
– Not all information it produces is competitively valuable
• What’s competitively valuable requires non-IT managers’ judgments
– Without it, IT excels at collecting more without a clear business purpose
– Troves of irrelevant data bog down business decisions
– Proverbial needle in the wrong haystack
The used of good prediction
Predicting a baseball superstar, rain, a box-office hit, pricing insurance, catching
spam
Assumption : Past behavior reasonably predicts the future
– Rely on historical data
Firms got better at collecting more data
– Spreadsheetsdata to warehouses data then to centers
A 2,000 year history
Business analytics: Using data to predict outcomes your firm cares about
8BC: Ancient Greeks thronged to hear predictions from the Oracle of Delphi
• Since then: Commonsense, nonsense, and wishful thinking
• Prediction turned scientific with statistics in England around 1700AD
• Hyper-progress around World War II (1939-1945)
Led to a flurry of innovative practices in all functional area
For example :
Segment markets
Tier customers on lifetime value
Forecast demand for a product any hour