CompanyGraph
How to decide whether a headline or price move changes the investment case—or merely changes today's attention.
Signal for which question?
Noise is not simply information that turns out to be false, and signal is not simply information that moves the share price. A signal is information that changes a defined decision: the expected cash flows, risks, financing needs, competitive position, or valuation at the horizon being considered. Noise is information that does not materially change that decision, even if it creates a large short-term price move.
The same event can therefore be signal for one investor and noise for another. A temporary input-price shock may matter to a leveraged producer with a covenant test next quarter but not to an unlevered company that can pass it through over several years. A regulatory proposal may be noise until it reaches the product, geography, and time period in which the company actually operates.
Price is a measurement, not a diagnosis
A share price records transactions and expectations at a point in time. It can move because of new information about the company, a sector or index flow, liquidity, positioning, or a change in the discount rate. The move may be important without identifying which cause dominates.
Research on price momentum finds continuation over specified formation and holding periods, but the NBER literature also examines competing explanations and later reversals. A rising price is therefore a reason to ask what changed, not evidence that the business became more valuable in the same proportion. Investor.gov advises investors to use public company disclosures and due diligence rather than relying on a single market signal.
Operating evidence has its own boundaries
Revenue, orders, margins, cash conversion, customer retention, inventory, and capacity each observe a different part of the business. A quarterly margin decline may come from a temporary input price, product mix, launch cost, or genuine loss of pricing power. Strong cash flow may reflect a working-capital release. Stable earnings may coexist with a customer concentration that makes the next contract decisive.
Calling a metric “fundamental” does not make it complete. The investor must identify the mechanism connecting it to future outcomes and what would weaken that connection. A lost contract is more informative when the customer represented a repeatable workflow and the company lacks replacement capacity. A small earnings miss is less informative when the variance is timing and the underlying orders, prices, and unit economics are unchanged.
Examples of the distinction
Suppose a company misses an analyst estimate by two cents while renewing its major customers, holding prices, and delivering the same product roadmap. The miss may be noise for a long-horizon operating thesis, although it can still reveal forecasting or communication problems. If the same company loses a customer representing a fifth of revenue and cannot replace the capacity, that is more likely signal because the future cash-flow path and concentration risk have changed.
A three-percent market decline after an interest-rate announcement may be noise for a specific business if its contracts, debt costs, and demand are unaffected. The same announcement can be signal for a highly levered builder whose financing and project pipeline depend on refinancing. The event is not classified in the abstract; its relevance is established through exposure.
Time reveals but does not solve the problem
Time can help distinguish a temporary observation from a persistent change, but there is no universal waiting period. A product recall, cyber incident, or covenant breach can be signal immediately. A slow loss of technical relevance may take years to appear in revenue. A short-lived price move can still matter if it triggers a margin call or prevents a needed equity raise.
Looking backward also creates hindsight bias. An outcome that became obvious later may not have been identifiable when the decision was made. Preserve the information available at the time, the alternatives considered, and the test that would have falsified the interpretation.
How attention becomes analysis
- State the decision: define the horizon and what would cause a buy, hold, sell, or risk-limit change.
- Map exposure: connect the news or price move to customers, products, debt, capacity, regulation, or competitors.
- Separate observation from inference: record what was measured, what management claimed, and what you infer from it.
- Find the next test: identify an order, renewal, cash collection, margin, financing, or regulatory event that can confirm or weaken the explanation.
- Check alternatives: ask whether sector flows, liquidity, accounting timing, or a different mechanism explains the same movement.
- Set a review date: decide when new evidence should change the assessment rather than reacting to every update.
Noise is not a synonym for “small” and signal is not a synonym for “dramatic.” The useful distinction is whether information changes the expected future of the business or the risk of owning it, under a stated question and horizon.