Top-Down vs Bottom-Up - Choosing Your Guesstimate Approach
After you clarify the question before you calculate, the next decision is the approach: should you start from a macro market and filter down, or start from a unit and scale up? This matters because the same guesstimate can be solved through either route, but one route is usually cleaner based on what is easier to estimate. In interviews, choosing well shows structure before arithmetic, and cross-checking with the other route shows judgement.
- Top-down starts from a macro figure such as population, total market size, total vehicle sales, or gross domestic product and narrows it using percentage filters.
- Bottom-up starts from a unit such as one customer, one outlet, one transaction, one MW of data-centre capacity, or one company and multiplies up.
- Use top-down when the base market is known or can be anchored, such as annual 2-wheeler sales of 1.9 crore units.
- Use bottom-up when operational units are easier to estimate, such as revenue per MW for data centres or named player capacity for EV 2-wheelers.
- The strongest answers often use one route for the main estimate and the other route as a cross-check.
- A good guesstimate ends with a plausible range, not just a point estimate, and invites the interviewer to flex assumptions.
Big Picture - Choosing the Route at a Glance
Think of top-down and bottom-up as two routes to the same credible answer. Your job is not to prove that one is universally better, but to choose the route that has stronger anchors for the specific question.
Top-down means starting from a macro figure and narrowing by percentage filters. Bottom-up means starting from a unit such as one customer, one outlet, or one transaction and multiplying up.
How the Choice Works in Practice
The choice is usually made after clarifying scope. Once you know geography, time period, unit of measurement, and whether the answer is volume or revenue, ask: which side has the more reliable anchor?
If the market base is known, top-down is often faster. If the business model has clear operating units, bottom-up may be more realistic. In many strong answers, candidates build the main case one way and then use the other way as a sanity check.
For FY28 electric 2-wheelers in India, a top-down route starts with total 2W sales: 1.9 crore × 1.05² ≈ 2.1 crore units. Applying projected EV penetration of ~18% gives about 38 lakh EV 2W units, with a range of 32-45 lakh units. The bottom-up sanity check uses named players: Ola targets 10 lakh per year by FY27, while TVS, Bajaj, and Ather together add another 15-20 lakh, giving around 30 lakh organised volume plus the tail. The strategic so what: the top-down demand pool and bottom-up supply view broadly agree, so the estimate is credible.
Top-Down Approach
The top-down approach starts with the broadest relevant market and narrows it using filters. A filter can be a penetration rate, share, conversion rate, usage rate, or eligibility rate. For example, to estimate FY28 EV 2W units, the base is total FY28 2W sales, and the filter is EV penetration.
This approach matters because it forces you to define the addressable universe first. In market sizing, the largest mistakes often come from starting with the wrong universe, such as using all vehicles when the question is only 2-wheelers, or using all users when the question is daily active users.
Top-down is especially useful when the macro base is already available from a known anchor. The source cheat card gives annual 2W sales as 1.9 crore, active 2W fleet as around 20 crore, smartphone users as 80 crore, internet users as 95 crore, and India population as 145 crore. These are the kinds of anchors that make a top-down estimate efficient.
The nuance is that every filter must be defensible. A top-down answer can look elegant but fail if the candidate casually invents a penetration rate, conversion rate, or share. To avoid that, state why the filter is reasonable and then cross-check with another route.
Bottom-Up Approach
The bottom-up approach starts from a unit and scales up. The unit could be one business, one transaction, one registered seller, one data-centre MW, one airport, one vehicle, or one daily active user. Instead of asking "what share of the market is this?", it asks "how much activity or value does one unit generate, and how many such units exist?"
This approach matters when operational behaviour is easier to estimate than market share. For data centres, the source approach anchors on operational IT load capacity of about 900-1,000 MW across Indian data centres, applies revenue per MW of ₹12-16 crore per year, and then adjusts for occupancy of around 70-75%.
Bottom-up is powerful because it mirrors how a business actually operates. For example, the after-market 2W spare-parts estimate starts with an active fleet of about 20 crore vehicles, applies annual maintenance spend of ₹3,000-5,000 per vehicle, and then uses a spare-parts share of 55-60%.
The nuance is representativeness. If the chosen unit is unusually large, small, urban, premium, or mature, multiplying it across the whole market can overstate or understate the result. A bottom-up answer needs either a blended unit or a clear segmentation.
Top-Down vs Bottom-Up Comparison
In interviews, the right approach depends on the nature of the estimate. A revenue pool, product adoption question, or operational capacity question may each need a different starting point.
For product-management guesstimates, the source explicitly warns that sizing a feature as "DAU × ₹" misses the intermediate stack. A better structure decomposes the funnel into exposure, adoption, activation, retention, and revenue impact. That is a bottom-up style of thinking because it builds from user actions rather than jumping from the whole user base to value.
Worked Example - FY28 EV 2W Market in India
This example shows how to combine both routes: use top-down for the main estimate, then bottom-up as a sanity check.
The important move is not the multiplication alone. The answer becomes interview-ready because it explains why the primary route was chosen, then checks the result against real named players and a plausible tail of smaller brands.
Cross-Check and Sanity Check
A cross-check means solving the same problem two ways, such as demand side versus supply side. A sanity check means comparing the final number to an external benchmark, a per-capita implication, a comparable market, or a known company target.
In the EV 2W example, the top-down answer gives 38 lakh units. The sanity check says Ola alone targets 10 lakh per year by FY27, while TVS, Bajaj, and Ather together add another 15-20 lakh, giving roughly 30 lakh organised volume before adding tail brands. That makes the 32-45 lakh range more believable.
The nuance is that a cross-check does not need to match perfectly. It needs to be directionally consistent. If two routes give answers within a plausible range, you can defend the estimate. If they differ sharply, explain which assumption you would revisit first.
Structuring a Top Interview Answer
"Estimate the FY28 revenue market size for electric 2-wheelers in India. Would you use a top-down or bottom-up approach?"
The number one way candidates get this wrong is choosing an approach silently. Say why your route fits the available anchors, then use the other route as a cross-check rather than treating it as a competing answer.
Conclusion
Top-down and bottom-up are complementary guesstimate routes. Start macro and filter down when the market base is known, start from units and scale up when operations are clearer, and always use a cross-check to make your answer defensible.
The most frequent error is forcing one approach even when the question gives better anchors for the other. This costs points because the interviewer sees arithmetic without judgement, and a single-route answer is harder to defend when challenged!