This AI trick changes everything

Guesstimate questions are a common part of Product Manager interviews and day-to-day problem-solving.

Their goal is to test your ability to break down complex, ambiguous problems into simpler components and come up with a logical estimate.

I have interviewed 1000+ candidates and I have seen that candidates can identify the key parameters, but struggle to tie these parameters together.

In this newsletter, I’ll dive into a simple yet effective framework to break down guesstimates and see how AI can help accelerate your mastery of these skills. With consistent practice and AI-powered tools, you can transform your approach to guesstimates, making the process second nature.

The Goal of Guesstimates:

The goal is to identify all parameters that impact a particular metric and articulate a clear formula to tie these factors together. For instance, consider the question: “How many pizzas are sold in a city on a given day?”

Most PMs can easily list the influencing factors (e.g., population, average daily consumption rate, average household size, etc.).

But building a formula that accurately captures all these factors is what separates a good guesstimate from a great one.

Here’s where practice comes in — and AI can be your practice partner.

Step 1: Understanding the Parameters

Take a moment to look around you and think of a simpler object — a pizza, a cup of tea, or even a pair of shoes.

Ask yourself: “How many of these would be sold in any given day?” Begin by identifying the key parameters:

1. Target Market Size: The total number of people in a specific area.

2. Average Consumption Rate: What percentage of the population orders this product daily?

3. Average Quantity Per Order: How many units are purchased in a single transaction?

4. External Factors: Weather, time of day, and special events.

Now, once you’ve listed the parameters, the next step is to build a logical formula. For example:

The more you practice, the easier it becomes to recognize patterns and relationships between variables. With enough repetition, you’ll start seeing opportunities for guesstimates everywhere!

Step 2: Introducing AI as a Guesstimate Coach

Here’s where AI can step in to turbocharge your practice. Tools like ChatGPT, can act as a virtual coach by helping you:

1. Validate Parameter Lists: You can ask, “What are some factors that impact pizza sales in a city?” and compare them with your own list.

2. Suggest Missing Variables: The AI can suggest additional parameters you might have overlooked (e.g., weekday vs. weekend consumption).

3. Help Create a Formula: After listing your variables, ask the AI to suggest formulas and see how close your logic is to its suggestion.

For instance, let’s take a simple practice problem:

Question: How many cups of coffee are sold in an office building daily?

1. Identify Parameters:

   • Number of employees

   • Percentage of employees who drink coffee

   • Average cups per employee

   • Office hours (e.g., some employees may skip morning coffee if they arrive late)

2. Create a Formula:

3. AI Feedback & Enhancement:
   Use AI to test variations and explore different scenarios, e.g., "What if the office introduces a new espresso machine? How might it change consumption?"

Step 3: Moving to Digital Product KPIs

Once you’ve mastered physical objects, transition to digital products. Consider an app and practice guesstimating its key KPIs, such as:

• Daily Active Users (DAUs)

• Monthly Install Rate

• Churn Rate

Example Problem: Estimate the daily installs for a productivity app with 50,000 total users.

1. Identify Parameters:

   • Average new installs per day

   • Uninstall rate (churn)

   • Average install duration (retention period)

2. Build a Formula:

3. AI Application: Use AI to analyze different retention periods and churn rates, or simulate scenarios like a sudden surge in installs due to a viral marketing campaign.

Step 4: Refining Estimates and Handling Outliers

I have seen that many times candidates focus too much on the outliers that they forget the key factors. In fact, after some time struggling with this, most candidates realize the struggle but are unable to come out of it.

It’s important to note the outliers. Often, there is merit in noting them early but include them in the calculation towards the end.

This ensures that you have completed the key answer before grappling with the outliers.

One common pitfall is failing to account for outlier scenarios that throw off your estimates. These include:

• New Competitors entering the market

• Macro-economic changes like recessions

• Seasonal Demand Shifts

To address this, ask the AI to provide you with potential disruptors and then adjust your formula accordingly.

For example, for the coffee example above, what if the office building introduces a new work-from-home policy three days a week? The AI can simulate how this impacts overall coffee consumption, making your estimate more robust.

I have created and refined this GPT with my insights from taking 1000+ interviews. Use this to your PM interview questions and get feedback from an interviewer's perspective.

If you are looking for support with preparing for an upcoming interview or looking to improve your first-round interview calls, book a 1:1 call here 👇

If you have any questions, reply to this email and I will be happy to help.

I will see you next week.

All the Best

Aditi