When we are preparing for consulting(or even pm), guesstimates are often sidelined. Often, they come to us intuitively, and it’s a good thing if that happens with you. But when that does not happen, it’s good to have a list of checkpoints and ideas in your back-pocket.
For those who are not aware of what guesstimates are, the one-line definition is that they require you to estimate a number basis the problem statement.
I will list down some insights with a simple example:
Number of fridges running/operational in India
Step 0: Realize that your final output will be a number ‘x’ to help put things in perspective.
Step 1: Begin by asking preliminary questions.
Think of them as means to get clarity and to sharpen the problem statement. Beyond basic doubts, I always thought of them to achieve the following:
a. Reduce (or limit) the size of the universe we are working in (by reducing complexity - for e.g., Do we need to include industry grade refrigerators in the calculation? Do we need to consider commercial spaces keeping refrigerators in their pantry or only domestic households?)
b. Put the number ‘x’ in between an upper limit and lower limit to help sharpen the problem. (Not applicable for this guesstimate but for e.g., - If the guesstimate is to estimate the bid cost of an IPL team; Ask your interviewer if Rs. 100cr is a fair upper limit for setting up an IPL team)
Note: Sometimes you can apply both to a problem statement and other times you will just be able to apply one of them.
Step 2: Pick a base number which will be the starting point of your calculation.
For e.g., - 1) Population of India and subsequently the number of households in India could be your base number for estimating the refrigerators;
2) For identifying number of refrigerators in commercial spaces/offices - Working population of India can be a base number to eventually assign a number to y employees/refrigerator
This will be directly proportional to your final number ‘x’. It will be a standard number like population or GDP of India.
Always align on this number with your interviewer and give a rationale behind why you are choosing something as a base number.
Note: In some cases, this might always not be needed. Always try to think of a base number, but if it does not make sense, do not force fit it.
Step 3: Add a first level filter to break down the base number into broad buckets. Apply second level filters on an individual case basis
Filters are factors and parameters that you can apply to the base number and reduce it to bring it closer to the final number ‘x’.
One way which works out in most of the cases is the following:
a. Segment the problem statement into various cases which can be thought of an individual basis
b. List the factors which will tailor the base number to the problem statement and then estimate a % number for each of those factors which will reduce it to ‘x’.
Base number -> Case 1, 2,3 …
Final number ‘x’ = Sigma Base number of Case i * % Factor 1 * % Factor 2 * % Factor 3 …
One way of thinking which worked for me was to think of each statement from a contextualised supply and demand side. I used to pick the easier one (supply/demand) and start from there.
Now you must have achieved your final number ‘x’.
For e.g., - 1) Number of households can be broken into urban, rural & income groups as broad buckets. Second level filters will include buying power basis discretionary income, electricity availability and so on.
2) Working population can be classified as organised, unorganised. Organized can further be classified as small, medium and large businesses and so on.
Step 4: Sanity check – the secret to knocking it out of the park!
In the consulting industry, checking if your estimates are correct is key. You need to think of it from another perspective and reach a similar number.
Don’t worry, you wouldn’t have to do another guesstimate. This is a way to do a quick check on whether your output is broadly correct. It also communicates your ability to doubt your answer and check it once again to the interviewer. This quality is again very important in the industry.
There are various ways to do this, I have listed down some of the obvious ones:
a. Pick another base number and quickly apply some filters
b. Use the same base number and apply different filters
With reference to the supply/demand strategy, if you solve first using the supply side, perform a sanity check from the demand side.
The sanity check answer doesn’t have to exactly match the final number ‘x’. Just try to keep it within a delta of 20%.
For e.g., - If you consider the average life of a refrigerator to be 10 years (10% of total refrigerators replaced each year) + assume a 5% growth in the market of refrigerators - Broadly, the industry would be manufacturing 15% of the refrigerators currently in use on a per annum basis (assuming no fridges are imported)
x*15%/365 will give the daily production number of refrigerators in India. Now, one can align with the interviewer if this seems reasonable.
You can also estimate the number of factories in India and their daily production to tally it with x*15%/365.
Conclusion
Preliminary questions are the most important bit to make your life simpler while solving a guesstimate. It helps make it less complex and constrains the areas you can explore. It’s not feasible to list down the types of guesstimates. Rather, listed below are some base number ideas which may be helpful in different cases:
Population: India - 140 crores(1.4 billion)
GDP: Income distribution - Below Poverty Line(Poor) - 30%
Lower Middle Class - 25%
Upper Middle Class - 25%
Rich - 20%
Area : India ~ 3.2 million km sq.
Others : Urban Rural Split - Rural - 70%, Urban - 30%
Average Family Size: Rural - 5. Urban - 4
Always try to segment the base number into various cases so that you can think of the applicable factors easily. Try to work out a sanity check to cross-verify your output.
Most importantly, always align at each step with your interviewer. Explain to them the rationale behind each decision before moving forward with it. This will help you to steer the guesstimate in the right direction and get a sense check.
This article is not an exhaustive piece. There may be various other ways and insights that would be helpful while solving guesstimates. I’d encourage all to post them in the comments to create a larger knowledge base.
Here is one guesstimate for you to solve : Estimate the number of subscribers of Amazon Prime in India. Do post your approach and numbers in the comments.
Loved your post.
My approach to guesstimate:
Population of India: 1.4 cr
Urban Rural Split = 30%-70%
Urban Population = 420 million
Urban Household ~ 100 million
Interent Penetration = 70%
Poor:Middle:Rich = 20%:50%:30:
I assume 20% middle class households to have Prime subscription and 40% Rich to have Prime subscription
So my final number boils down to 7+8 = 15 million