Imddey

When people make choices, they don’t always act in a purely rational way. Sometimes, they consider price and time, while other times, brand loyalty, convenience, or mood might influence their decisions. The scale parameter (μ) helps capture this randomness in choice models like the logit model. In this post, we’ll break down a simple ride-sharing decision and explore how μ impacts the probability of choosing Uber or Lyft.

The consumer buys when the utility of buying exceeds the utility of not buying.

Disclaimer : This post is for learning and data is not real. And not reflect the service quality as well

---

Step 1: Defining the Utility Function

To model a customer’s decision, we assume their choice depends on several factors like price, wait time, driver rating, and personal preference. The utility function (a numerical representation of how much a customer values each option) can be written as:

Where:

• are weights representing how much each factor contributes to the decision.
• A higher U value means a higher preference for that option.
• The scale parameter (μ) determines how deterministic or random the choice is.

Step 2: Assigning Values to Uber and Lyft

Let’s assume a customer is choosing between Uber and Lyft with the following attributes:

Attribute	Uber	Lyft	Weight (β)
Price (Rs.)	300	320	-0.02
Wait Time (min)	5	8	-0.5
Driver Rating	4.8	4.6	+2.0
Brand Preference	1	0	+1.5

Now, we compute the utility for each ride-sharing option.

Utility for Uber:

Utility for Lyft:

Since Uber has a higher utility (10) compared to Lyft (8), the customer is more likely to choose Uber. However, real-world decisions involve uncertainty, and that’s where the scale parameter μ plays a role.

---

Step 3: Calculating Choice Probabilities with Different μ Values

Step 4: Key Takeaways

μ Value	Effect on Decision	Interpretation
Small (μ = 0.1)	Rational, utility-based choices	People always pick Uber because utility differences dominate
Moderate (μ = 1)	Some randomness	Most choose Uber, but a few may still prefer Lyft
Large (μ = 5)	Highly random choices	People may pick Lyft even when its utility is lower

Final Thought: The scale parameter μ controls how much “human randomness” is involved in decisions—a small μ means purely logical choices, while a large μ allows for unpredictable consumer behavior. Marketers and businesses can adjust their models to reflect real-world behavior more accurately.