Why Research Coverage Is Becoming Obsolete

Research coverage, the fixed list of companies a team can afford to follow closely, is becoming obsolete because it was never a strategy in the first place. It was a rationing device for scarce analyst time, and the thing being rationed is getting cheap.

For decades, an analyst could genuinely follow only so many names. Reading every filing, listening to every call, keeping a model current, and noticing when a footnote changed took real hours per company per quarter. So firms did the only sensible thing: they picked a list. This is my coverage. Everything on it, I know cold. Everything off it, I do not touch. That list, not the analyst’s skill, is what defined the edge of what a desk could see.

The coverage list is a budget, not a belief

It helps to see clearly what a coverage list actually is. It is not a considered view that these twelve companies are the only ones worth understanding. It is a budget. A person has a fixed number of hours, each company needs a certain number of those hours to be followed properly, and the list is simply how many names fit inside the hours available.

You can watch the arithmetic happen. Following a company well is not a one-time task. When a firm reports a quarter, you read the release, reconcile it against what you expected, listen to the call, note the exact words management used about the future, and update your model. Then you wait a quarter and do it again. Multiply that by the number of names on your list, and you hit a ceiling fast. The ceiling is why coverage lists are short.

Coverage was never a list of the companies worth knowing. It was a list of the companies you had time to know.

Everything outside that list falls into a blind spot. Not because those companies do not matter, but because there were no hours left. Smaller companies, names in adjacent sectors, a supplier or a customer of something you already own: all of it sits in the dark, watched by no one on the desk, until it becomes big enough or noisy enough to force its way onto the list.

What actually eats the hours

To see why the constraint is loosening, look at where the time goes. The expensive part of following a company is rarely the judgement. It is the tireless first pass: gathering the documents, reading them end to end, pulling the numbers into a consistent shape, and noticing what moved since last time.

Take management guidance as one concrete example. When Asian Paints pointed on a recent call to volume growth in the band of about 8-10% and an EBITDA margin band of 18-20%, and spoke of “maintaining our margin guidance,” a serious analyst does not just note the number. They record the exact language, because guidance is a forward claim with a hedge inside it, and then they check every quarter whether reality is tracking toward that band or drifting away, and whether the tone shifts from confident to cautious. That is valuable work. But most of it is watching: reading the next release, finding the guidance line, comparing it to the last one. The judgement, deciding what a drift actually means for the business, is a thin slice at the end of a thick slab of routine.

The same is true of the raw shape of a business. Reliance Industries, in the quarter ended September 2025 (Q2 FY26), reported four segments with completely different economics: Oil to Chemicals at roughly ₹1,60,600 crore of revenue on an EBIT margin near 9%, Retail near ₹90,000 crore on about 6%, Jio near ₹42,700 crore on about 52%, and upstream Oil and Gas near ₹6,100 crore on about 83%. Understanding a company like that means holding four separate stories in your head and tracking each one. The insight that the fattest margins sit in the smallest segments is judgement. The labour of pulling those four segment lines out of the disclosure, quarter after quarter, is not. It is exactly the kind of tireless reading that used to force the list to stay short.

When that tireless first pass gets cheap, the budget changes. The hours that used to be spent gathering and reading can be spent thinking, and the number of names a single analyst can stay current on stops being capped by how fast they can personally read.

Loosen the constraint and the list stops making sense

Here is the part worth sitting with. If the coverage list only ever existed because reading and monitoring were expensive, then cheapening reading and monitoring does not make the list bigger. It makes the whole idea of a fixed list start to dissolve.

Think about why you would ever choose in advance, and stick to, exactly fifteen names. You did it because following the sixteenth would have cost you attention you could not spare. Remove that cost, and the case for a hard boundary weakens. Why should a supplier two steps down your holding’s value chain be invisible just because it did not make the cut three years ago? Why should a company you might want to own next year be un-watched until the week you start looking at it?

The old answer was always the same: because there were not enough hours. Once that answer stops being true, “it is not on my coverage list” stops being a reason and starts being a habit.

This is a different claim from the one about static reports going stale. A report decays because a company keeps changing after you hit publish, a problem we cover in the death of the static research report. Coverage is a separate limit. Even a team that updates its work continuously has, until now, only been able to do so for the short list of names it chose. The static-report problem is about time within a name. The coverage problem is about how many names you can hold at all.

What replaces coverage

The replacement is not a longer list. It is a change in shape, from a list you pick to a surface you watch.

Under the old model, attention is spent up front. You decide who is on the list, and then you spend hours keeping each one current whether or not anything happened. Most of that effort is spent confirming that nothing changed, which is necessary and dull. Under the new model, the routine watching runs continuously and quietly across a much wider set of names, and the scarce human attention is spent only where something actually moved: a guidance band walked back, a new risk factor, a segment redefined, a number that broke from its own trend.

Coverage, in other words, stops being a gate you pass names through in advance and becomes a live field you monitor, with the analyst pulled in by exception rather than by calendar. This is the natural extension of always-on monitoring, which we make the case for in why continuous research is a competitive edge, scaled from a handful of holdings to a far broader universe. The discipline of watching many positions at once, described in how to monitor a portfolio of holdings, stops being a logistical strain and becomes the default state.

None of this removes the analyst. The question of whether AI replaces the job, which we take up directly in will AI replace financial analysts, has a clear answer here: the human still owns the judgement, the thesis, and the accountability. What the human sheds is the obligation to personally read every page of every name in order to stay current. The coverage list was a fence built around scarce attention. Skill was never the thing in short supply. Hours were. And it is the hours that are becoming abundant.

The quieter consequence

There is a second-order effect worth naming. A short coverage list does not just limit what you see. It shapes what you think is worth seeing. When your universe is fifteen large names, you start to believe those are the important ones, partly because they are the only ones you have ever looked at closely. The list becomes a lens, and the lens becomes a bias.

Widen the field and that bias has less room to hide. The company that would never have earned a slot on a crowded desk, the second-order supplier, the mid-sized name in a sector nobody was assigned, becomes visible by default rather than by special request. Whether any of that changes what a given investor concludes is entirely up to them. But the raw fact of being able to see it, without first spending a scarce hour to earn the view, is new.

Coverage was a good answer to a real constraint. The constraint is fading. It would be strange if the answer did not fade with it.