Sector Intelligence or a Good Story? How to Test a Sector Thesis

A sector story is easy to recognise: demand is rising, a new technology is arriving, or policy is changing. Sector intelligence asks the harder question: what would we need to observe before believing the proposed business effect?

The difference is not the length of the report. It is whether the explanation has a date, a mechanism and a test. A thirty-page outlook can still be mostly narrative. A short note can be useful if another researcher can inspect its evidence and understand its limits.

This is a general research method, not a sector-selection system. Our Sector Intelligence desk applies it to dated public publications, while the Global Intelligence desk separates their global arguments from Altys’s own interpretation.

Start with the claim, not the conclusion

“Demand is strong” is too broad to test. Demand for what, measured how, over which period? It could mean enquiries, bookings, units shipped, completed installations, retail sales or cash collected. Those observations are related but not interchangeable.

Rewrite the statement in a way that specifies the proposed effect. For example: “More customer orders should lead to higher delivered volumes over the coming reporting periods, provided production capacity and customer funding are available.”

That is still an assumption. But now it exposes the steps that could fail. It also prevents a researcher from later claiming that any positive number proved the original argument.

Give the thesis three dates

A publication date tells you when the author wrote the view. An observation period tells you when the evidence occurred. A forecast horizon tells you when the proposed effect is expected to appear.

These dates need separate fields. A recent article may rely on older operating data. A quarterly slowdown may coexist with a longer-term expansion thesis. Neither should be hidden under a single “latest” label.

When two reports appear to conflict, check these fields first. One may discuss next quarter’s utilisation while another discusses capacity additions over several years. Disagreement about the horizon is not the same as disagreement about the evidence.

Draw the business bridge

Choose a concrete sequence. In an equipment industry, it might be customer demand, orders, manufacturing, delivery, revenue recognition and cash collection. For a consumer business, it might be production, distributor shipments, retail sales and replenishment.

At each step, ask what source supports it and what constraint could interrupt it. The number of steps matters because a narrative often leaps from the first directly to the last.

Do not treat a positive development at one stage as proof of every later stage. A factory announcement does not prove utilisation. A contract does not prove profitable execution. Revenue does not prove collection. These are separate questions even when a presentation combines them into one growth story.

Keep three evidence types separate

Type	What it establishes	What it does not establish
Reported observation	A source reported a defined measure for a period	That the measure uses the definition you assumed
Management or publisher view	The author expressed that expectation on that date	That the expected outcome occurred
Analyst inference	A researcher proposes a mechanism or interpretation	That the interpretation is a disclosed fact

A quotation from management is useful for tracking guidance. It is not independent corroboration of that guidance. An outlook that cites the same company presentation is another interpretation of the presentation, not necessarily a second factual observation.

Likewise, three reports repeating one industry dataset do not provide three independent datasets. Their reasoning may add value, but the underlying evidence still has one source. Preserve that distinction in the research record.

Write the challenge before the next result

A thesis that cannot be challenged is hard to learn from. Define observations that would weaken the mechanism, not just a price move that would feel disappointing.

For a demand thesis, a challenge might be weaker retail sales alongside rising dealer inventory. For an execution thesis, it might be repeated delivery delays. For a cash-conversion thesis, it might be recognised revenue accumulating in receivables without the anticipated collections.

None of these is a universal verdict. A temporary timing mismatch may have a documented explanation. The point is to trigger review and identify what additional evidence is needed, rather than mechanically label a business good or bad.

Use a small evidence card

An evidence card can be a spreadsheet row, a research note or a document table. Include the exact claim, publisher, publication date, underlying observation period, business mechanism and next disclosure to check. Add a field for competing explanations and another for unavailable evidence.

Do not turn missing data into a neutral-looking number. A blank exposure field means the exposure is unknown. Zero means an observed or justified absence. Those are materially different statements.

Keep definitions stable when comparing reports. An industry shipment estimate, a company’s consolidated revenue and a segment’s order book cannot be combined into a single growth calculation without a defensible bridge. Document any transformation so another researcher can reproduce it.

Update the evidence without rewriting history

At the next review, retain the original card. Add the new observation and record whether it supports, challenges or leaves the assumption unresolved. If the definition changed, explain the change instead of presenting the series as continuous.

This makes the research useful even when the original view was wrong. You can see whether the error came from an unsupported premise, a broken transmission step, an unexpected external event or a horizon that was never specific enough.

Altys’s source-linked company research, guidance history and monitoring workflows support revisiting the evidence behind a view. Where outputs are exportable, Excel provides another way to inspect the inputs and calculations. This article does not claim that every step is automated or that a score replaces judgment. Request access to explore the workflow.

Good sector intelligence is not a more confident story. It is a clearer account of what is known, what is assumed and what would make us reconsider.

General research-process education only. No securities recommendations, targets, expected returns or portfolio allocations. Altys Labs is not a SEBI-registered Research Analyst or Investment Adviser.