A market-size number without a source is worthless

A market-size claim is one of the most commonly asserted, least commonly sourced numbers in any business plan — repeated from deck to deck, sometimes without anyone along the chain having read the original research it supposedly came from, occasionally without any original research existing at all. Once a figure like 'the market is worth ₹500 crore' is written into a plan, it tends to sit there unquestioned, because a large, specific-sounding rupee number reads as credible on its own, independent of whether anything behind it has actually been checked.

That's the trap: precision looks like evidence even when it isn't one. A number with several significant figures and a rupee symbol in front of it feels more researched than a vague sentence would, regardless of whether a real, traceable source produced it or it was typed in because it sounded plausible.

A market-size number is only useful for actually deciding something — how big to build, how much to raise, whether an opportunity is worth pursuing — if the person reading it can trace it back to what it actually measured: whose research this is, from what year, covering what geography, including and excluding what. Without a publisher's name, a link, and a date it was retrieved, there is no way to tell whether '₹500 crore' describes a market from three years ago, a structurally different country's version of the same category, or a figure with no real research behind it at all that simply got repeated until it sounded established.

A business making a real decision on that number has no way to distinguish these possibilities from each other, which means the number is functioning as decoration rather than evidence — present in the plan, contributing nothing checkable to the decision it's supposedly informing.

A bottom-up estimate starts from things a business can actually know about itself, rather than a claim about the world at large. How many potential customers exist within a reach the business can actually define — a delivery radius, a city, a specific segment? How often would they realistically buy? What would they pay, based on real pricing rather than a hoped-for number? What share of that reachable group is genuinely serviceable given real constraints — capacity, category fit, delivery limits? And what share of that serviceable group can the business realistically capture given where it actually stands today, not where it hopes to be in three years?

Multiplying these five figures together produces TAM, SAM and SOM as arithmetic a reader can actually check line by line — every input is something the business itself is naming and can defend, not a total borrowed wholesale from somewhere else and hoped to be relevant.

A bottom-up estimate is only as good as the business's own inputs, and it says nothing on its own about whether an outside analyst's very different total agrees or disagrees with it — which is a genuinely useful cross-check, but only if that outside figure comes with exactly the same sourcing discipline being demanded of everything else: a named publisher, a URL, and the date it was actually read.

Two numbers built from opposite directions — one from the business's own assumptions upward, one from an industry report downward — landing far apart from each other isn't a problem to quietly paper over. It's a signal worth sitting with, because a real disagreement between a bottom-up and a top-down estimate of the same thing usually means one of the two has a bad assumption buried in it somewhere, and figuring out which one is more valuable than smoothing the disagreement away.

A blended figure — the bottom-up number and the top-down number simply averaged together — belongs to neither method. It inherits the top-down number's lack of traceable, line-by-line arithmetic and the bottom-up number's dependence on the business's own possibly-optimistic assumptions, while displaying the reasoning behind neither. It looks more authoritative than either input alone, precisely because it hides where it actually came from.

Reporting the two figures side by side, with the ratio between them stated plainly, gives a reader something they can actually interrogate — a specific, visible disagreement worth investigating, rather than a single smoothed-over number that quietly launders two very different kinds of uncertainty into one confident-looking total.

Before writing a market-size figure into a plan, name the five inputs a bottom-up number would actually need — reachable customers, purchase frequency, price, serviceable share, capture-rate share — and build the estimate from those first, as arithmetic that can be checked line by line. If an outside total is being added as a cross-check, require the same three things any other fact in the plan would need: a named publisher, a link, and the date it was retrieved. And never merge the two into a single blended figure — report them side by side and let the gap between them, if there is one, say something honest instead of hiding it.

The full mechanics of how a bottom-up estimate is built and cross-checked, including what's refused outright without proper sourcing, are documented at /docs/yukti. Once a plan has real numbers behind it, turning that plan into scheduled, owned work with actual milestones is a separate discipline, documented at /docs/guru.