Relative valuation looks easy. Choosing the right peers and the right multiple is where the judgement lies.
Comparable company analysis is often taught as the quick alternative to a DCF: find similar companies, take the median multiple, apply it. In practice, comps require as much judgement as any model, just hidden in different places.
Choosing the peers
"Same industry" is a starting point, not a definition. Two pharmaceutical companies can have completely different economics if one sells branded formulations in India and the other makes generic APIs for export. Good peers share growth, margins, risk and capital intensity, not only a sector label.
Choosing the multiple
EV/EBITDA suits capital-intensive businesses with different debt levels. P/E works for stable, profitable companies with similar leverage. Price to book matters for banks and lenders. Using the wrong multiple can make a cheap company look expensive and the reverse.
Cleaning the numbers
One-off gains, accounting differences and different financial year-ends distort multiples. A comparison is only fair when the numbers underneath are prepared the same way.
Reading the range
The spread of multiples across peers often says more than the median. A wide range means the market is pricing real differences, and the analyst has to explain where the target sits in that range and why.
A worked example
Suppose you are valuing a mid-sized Indian specialty chemicals company. A quick screen returns twelve "chemicals" companies with a median EV/EBITDA multiple. On closer inspection, three are commodity producers with cyclical margins, two are much larger with global customers, and one had a one-off gain inflating EBITDA last year. Removing them leaves six genuine peers, and their median is meaningfully different from the original twelve.
Then the analyst has to place the target within that range. Is it growing faster than peers? Are its margins more stable? Does it depend on one customer? Each answer moves it up or down the range, and each should be written down.
Forward or trailing?
Trailing multiples use reported numbers and are verifiable. Forward multiples use estimates and better reflect growth, but they import someone's forecast. For a fast-changing business, forward multiples are usually more informative; for a stable one, trailing multiples are usually enough. Mixing the two across peers is a common and silent error.
When comps mislead
Comparables reflect market sentiment. If an entire sector is overvalued, comps will say the target is fairly priced at an inflated level. That is why I treat them as one input next to an intrinsic valuation, never as the answer on their own.
A habit worth keeping
Always show the peer list, the reasons for each exclusion and the full range, not only the median. A reader who can see the peers can judge the conclusion.
Used carefully, comparables are a reality check on a DCF: they tell you what the market is paying today. Used carelessly, they import other people's mistakes into your valuation.