Managing Research Coverage Across Sectors

Institutional teams manage research coverage across sectors by treating attention as a fixed budget and spending it deliberately: they decide which sectors and companies deserve deep, continuous work, which get lighter monitoring, and which they are willing to ignore for now. The hard part is not analysis. It is allocation. No team, however good, can follow every company in every sector at full depth, so the real skill is choosing where to be deep, where to be shallow, and making sure every name still has an owner.

Most people imagine research as a stack of deep-dive reports. In practice, a buy-side team spends much of its energy on a management problem that comes before any single report: given a handful of analysts and hundreds of potentially relevant companies, who covers what, how closely, and how do you keep the whole map current. Get that wrong and you either spread everyone so thin that nothing is understood well, or you go so deep on a few names that a risk blindsides you from a company nobody was watching.

Coverage is a budget, not a wish list

The founding fact of coverage design is scarcity. A team has a finite number of analyst-hours per week, and every company that genuinely matters wants more of them than you can give. So coverage starts by admitting a hard tradeoff between two things you both want.

The first is depth: how well you actually understand a business. Depth means you have mapped its segments and the drivers underneath them, you know how it recognises revenue and where the accounting is soft, you have listened to several years of management commentary, and you can tell a real change from noise. It is expensive: it takes weeks to build and ongoing effort to maintain.

The second is breadth: how many companies you can touch at all. Breadth is what lets you spot an opportunity in a corner of the market you do not normally follow, or notice that a supplier to one of your holdings is signalling trouble.

You cannot maximise both with a small team. Every hour spent going deeper on one name is an hour not spent widening the net. Coverage design is simply the decision about how to split that budget, made on purpose rather than by accident. Teams that never make the decision explicitly end up making it implicitly, usually by defaulting to whatever is loudest that week.

Prioritise by what can actually hurt or help you

If attention is the budget, the next question is where to spend it first. The honest answer is not “the most interesting companies.” It is the companies where being wrong costs the most or being early helps the most.

A few lenses that teams use to rank names:

• Portfolio weight. A large position deserves closer coverage than a small one, because an error there does more damage. The biggest holdings should be the best understood, not merely the most talked about.
• Fragility of the thesis. Some holdings rest on a stable, well-proven business; others rest on a bet that is still playing out. The fragile ones need more frequent checking, because they have more ways to break.
• Rate of change. A sector in the middle of a shift, whether regulatory, technological, or competitive, moves faster than a sleepy one and burns more attention to stay current. Coverage should follow the change, not the alphabet.
• Asymmetry. A name you do not own but might, where a single quarter could open or close the case, can be worth watching even without a position, because the payoff to being ready is high.

None of this is static. Weights drift, theses mature or wobble, and sectors heat up and cool down, so the priority list has to be revisited on a schedule rather than set once. This is the same discipline that runs underneath how PMS firms research Indian stocks, where a small team covers a broad market and cannot afford to spend depth where it does not change a decision.

The tiered model: deep core, monitored middle, triaged tail

The practical shape almost every team converges on is a tiered coverage model. Rather than treating all names equally, they sort companies into a few bands, each with a different intensity of work.

Tier	What it contains	How it is covered
Core	Largest holdings and highest-conviction names	Deep, continuous work: full model, every result, every call
Watchlist	Candidates, smaller positions, key non-holdings	Lighter monitoring: headline numbers, guidance, big changes
Tail	The broad universe of possibly-relevant names	Alerts and triage only, pulled up when something moves

The core tier is where the classic deep research lives: a maintained model, a written thesis, and full attention on every quarterly result and management call. The middle tier is monitored, not modelled in full: you track the numbers that would change your mind and let the rest go. The tail is not really “covered” in the standing sense; it is scanned, so that when something material happens the right name surfaces and can be promoted into a higher tier for a closer look.

The point of tiers is that they make the depth-versus-breadth tradeoff concrete. Instead of a vague sense that some names matter more, the team has an explicit rule for how much each name gets, and a mechanism for moving a company up when it earns the attention. It is also the only realistic way one analyst stays on top of a large list, the same scaling problem covered in how to monitor a portfolio of holdings.

Depth versus breadth, decided rather than defaulted

The tiers only work if the boundaries between them are chosen and defended. The failure mode is tier creep: everything slowly drifts into the core because letting go feels risky, until the team is trying to deeply cover a hundred names and doing none of them well.

Good teams manage this actively. They cap the size of the core tier, so adding a name means demoting another. They set a rule for what earns promotion out of the tail. And they accept, out loud, that some sectors will be covered thinly or not at all, because pretending to cover everything is how you actually cover nothing.

This is where sector structure matters. Companies in the same sector share drivers, so covering one deeply lowers the cost of covering its neighbours, which is why coverage is usually organised by sector rather than scattered. An analyst who has mapped the economics of one lender understands the next one far faster, because the questions are the same even when the answers differ. That efficiency is also why the continuous version of this work compounds, a point argued in continuous research as a competitive edge: the second look at a sector is much cheaper than the first.

Handoffs: making a name survive the analyst

The quiet risk in any coverage system is that the understanding lives in one person’s head. When an analyst leaves or a name moves between people, the numbers stay but the context can walk out the door: why the thesis holds, what is being watched, and which risks would break the case.

Coverage that survives handoffs is different: the thinking is written down and kept current, not just the data. A few habits make a name portable:

• A current thesis, in words. A short, living statement of why the position exists and what would change it, updated as the story evolves rather than written once and forgotten. The idea of a thesis as a living document underpins the thesis monitoring checklist.
• An explicit watch list per name. The two or three things that would actually move the view, so a new owner knows what to look at first instead of rediscovering it.
• A shared record, not private notes. History, models, and reasoning that live where the team can reach them, so coverage is a team asset rather than a personal one.

When those exist, a name can pass between people with the context intact, and coverage does not reset to zero every time the roster changes. When they do not, every handoff quietly destroys years of accumulated understanding, and the team pays to rebuild it.

What to take away

Managing coverage across sectors is a resource-allocation problem wearing an analysis costume. The teams that do it well are not the ones that try to cover everything; they are the ones that decide, on purpose, where their scarce depth goes and are honest about what they are choosing not to watch closely. A workable system usually comes down to a few disciplines:

• Treat attention as a fixed budget and spend it where being wrong costs the most.
• Sort names into tiers, and defend the boundaries so the deep core stays small enough to actually be deep.
• Organise by sector so shared drivers buy you breadth without paying full depth twice.
• Write the thesis and the watch list down, so a name survives the analyst who covers it.
• Revisit the whole map on a schedule, because weights, theses, and sectors all drift.

Better tools change the arithmetic of this problem rather than its shape. When more of the routine monitoring and the surfacing of what changed can be handled with source-linked, point-in-time information that is calculated rather than guessed, an analyst can hold a wider tail at a given depth and spend more of their scarce hours on judgment. That is part of what a research terminal like Altys is built to widen, but the discipline of choosing where to be deep does not go away. Whether standing, fixed coverage lists are even the right unit anymore is taken further in why research coverage is becoming obsolete.

This article is educational. Altys Labs is not a registered research analyst or investment adviser, and nothing here is investment advice or a recommendation to buy, sell, or hold any security.