Practice interviews

How many washing machines sell in one year?

Turn an installed stock into annual replacement and first-purchase demand.

“Any information it provides should be reviewed and validated to ensure accuracy and relevance to your application.” (opens the source)

BCG, Official AI application guidance · BCG Careers, Undated

The question

Northline Research is sizing an invented urban appliance market in India. Every number is a case assumption. The target is one-year new-unit category demand and gross retail value. It is not the installed base or one retailer’s sales.

Annual demand assumptions
Value
Installed units at start900000
Mean replacement life (years)6
First-time installations this year50000
Retail price per unit (₹)20000

Assume an evenly distributed replacement age profile. Each installation uses one new unit; replacement and first-time purchases are disjoint.

Range

Test replacement lives of 5 to 10 years, with all other inputs fixed. Ignore used machines, inventory changes and price differences. No retailer market share is given.

“Explain the estimate, the assumption range and the next check.”

A strong answer, step by step

Open each step in turn. This is one strong answer, not the only one.

Step 1: Define the annual unit

What would you do here? Open the move you’d make.

  • Separate replacements from first-time demand.

    The strong move. The target is a flow over one year.

  • Use 900,000 as annual sales.

    Tempting, but it misses something. The 900,000 figure counts installed units at one point in time, not purchases in this year.

See how a strong candidate says it

“Annual new units equal installed units divided by replacement life, plus first-time installations. The division turns a stock into units per year.”

What a strong candidate could say

Why it works: The target is a flow over one year.

Step 2: Show the range

What would you do here? Open the move you’d make.

  • Vary replacement life and show both ends.

    The strong move. Longer life reduces replacement demand while first-time demand stays fixed.

  • Scale both replacements and first-time installations when life changes.

    Tempting, but it misses something. The sensitivity changes product life only; the 50,000 first-time installations are held fixed.

See how a strong candidate says it

“At six years, 900,000/6 + 50,000 = 200,000 units a year, worth ₹4 billion. Lives of 5 to 10 years give 140,000–230,000 units, worth ₹2.8–4.6 billion. This is a scenario range, not a confidence interval.”

What a strong candidate could say

Why it works: Longer life reduces replacement demand while first-time demand stays fixed.

Step 3: Revise the assumption

A repair survey supports a new estimate: replacement life is now 8 years. The installed stock, first-time installations and price stay unchanged.

What would you do here? Open the move you’d make.

  • Use the revised life in the equation.

    The strong move. New evidence changes the replacement flow.

  • Keep the original 200,000 because it was a round number.

    Tempting, but it misses something. The longer life lowers annual replacements by 37,500; rounding does not preserve the old base.

See how a strong candidate says it

“At eight years, replacements are 112,500. Add 50,000 first-time installations for 162,500 annual units, worth ₹3.25 billion. I would update the base estimate.”

What a strong candidate could say

Why it works: New evidence changes the replacement flow.

Step 4: Choose the next research

What would you do here? Open the move you’d make.

  • Check age profile and replacement evidence next.

    The strong move. The estimate depends on a steady replacement approximation.

  • Use ₹3.25 billion as our retailer’s forecast.

    Tempting, but it misses something. No retailer share, channel reach or realized selling price was supplied.

See how a strong candidate says it

“I would validate the installed age profile and replacement behaviour before using this as a forecast. I would not call category retail value our revenue without a separate share and channel model.”

What a strong candidate could say

Why it works: The estimate depends on a steady replacement approximation.

The answer, in short

Decision
Use 162,500 units as the revised base, within the stated scenario range.
Because
Replacement life changed while the other assumptions stayed fixed.
What would change it
An uneven stock age profile or different first-time demand would change the equation inputs.

Common traps

  • Confusing stock with annual purchases.
  • Turning category value into company revenue.

Check your answer

  • I named units per year.
  • I separated replacement and first purchase.
  • I tested the life range.
  • I revised the base and named the next evidence.

Last reviewed Nov 10, 2026