Dr. Eric Siegel's Key Insight

Predictive Analytics Is the Money Maker

Among the four types of analytics, prediction drives the most revenue. Not because it's fancier — because it's actionable at the individual level.

Explore
The Four Types

Not all analytics are created equal

Every company uses analytics, but the financial impact varies wildly depending on which type you're using. Click each card to see the details — and the limitations.

📊

Descriptive

"What happened?"
Your rearview mirror. Dashboards, reports, KPIs. Last quarter's sales were down 12%. Website traffic dropped. Customer complaints rose.
Example

A hospital reports that 18% of heart failure patients were readmitted within 30 days last year. Important to know — but knowing that number alone doesn't prevent a single readmission.

💡 Limitation: It's information, not action. Every company does this. It's table stakes. You're summarizing the past, and the past is already spent.
🔍

Diagnostic

"Why did it happen?"
Root cause analysis. Why did churn spike? Because you raised prices on the basic tier and didn't grandfather existing customers.
Example

A retailer discovers that customers who experienced a shipping delay on their first order had 3x higher churn. Great insight — but it doesn't tell you which of today's customers are about to leave.

💡 Limitation: You're explaining yesterday's losses, not preventing tomorrow's. It's forensic accounting for decisions already made.
⚡

Predictive

"What will happen next?"
The money maker. Assigns a probability to each individual — each customer, patient, transaction, or machine — in time to act.
Example

This customer has a 73% chance of churning in 60 days. This transaction has a 94% probability of being fraud. This patient has a 41% chance of readmission within 30 days.

🏆 Advantage: It tells you what a specific person or thing will do next — at scale, in time to do something about it. That's where the money is.
🧭

Prescriptive

"What should we do?"
Optimization and simulation. If we adjust prices by 5% and shift ad spend, here's the projected outcome.
Example

An airline's prescriptive model says "offer dynamic pricing on Route 47." But the real value comes from the underlying prediction: this specific traveler is price-sensitive and will book if we drop the fare by $30.

💡 Limitation: Prescriptive analytics is only as good as the predictions feeding it. Without knowing which customers are at risk, what are you optimizing?
The Core Argument

Three reasons prediction wins financially

Siegel's argument is elegantly simple. Predictive analytics has three properties that no other type combines.

Individual-Level Action

You don't just know "churn is up." You know who is about to churn, so you can intervene with exactly those people — one at a time.

Operates at Scale

Score a million customers overnight. Flag 200 fraudulent transactions per hour. Prioritize 50 patients for follow-up. The per-unit economics are incredible.

Acts Before the Cost

Every other type is backward-looking or system-level. Prediction is forward-looking and individual-level. That combination protects and generates revenue.

Across Industries

Where the money shows up

Click any card to explore how prediction drives financial impact across different sectors.

🏦

Banking

Which loan applicants will default?
Avoid millions in bad debt by declining or repricing risky loans before they're approved.
$2.1T global lending market
🛒

Retail

Which customers will respond to this coupon?
Stop wasting marketing spend on people who'd buy anyway. Target only those the coupon actually persuades.
30-50% marketing waste reduction
🏥

Healthcare

Which patients will be readmitted?
Reduce penalty payments — CMS charges hospitals roughly $500M per year for excess readmissions.
~$500M/yr CMS penalties
🏭

Manufacturing

Which equipment will fail this month?
Prevent unplanned downtime that can cost between $10K and $250K per hour depending on the line.
$10K–$250K/hr downtime cost
🛡️

Insurance

Which claims are likely fraudulent?
The FBI estimates insurance fraud costs over $40B annually. Prediction catches it early, before payout.
$40B+ annual fraud cost
📱

Telecom

Which subscribers will cancel?
Retaining a customer costs 5–7x less than acquiring a new one. Prediction tells you who to save.
5–7x retention vs acquisition
The Punchline

The hospital analogy

"

Prediction is the multiplier that makes all other analytics more valuable. Prescriptive analytics without prediction is just theoretical optimization. Descriptive analytics without prediction is just a report nobody acts on.

🌡️
Descriptive

Tells you the patient's temperature

🔬
Diagnostic

Tells you why they have a fever

⚡
Predictive

Tells you which patients walking in right now are about to get sick

💊
Prescriptive

Recommends the treatment plan

Test Yourself

Knowledge check

See if you can match the scenario to the right type of analytics.

Quick Quiz