News Count is the number of recent news headlines we found and scored for a ticker over the last seven days. It is a measure of attention, not of sentiment: it tells you how loudly the market is talking about a company, and says nothing whatsoever about whether the talk is good or bad.
Prices move on information, and information arrives as news. A company that generated three headlines this week and a company that generated thirty are in genuinely different states, whatever the headlines said. The busy one has something happening: an earnings report, a lawsuit, a product launch, an activist investor, a rumour. The quiet one does not.
That matters for a forecaster because news volume is a decent proxy for how much fresh uncertainty is being priced. Elevated attention tends to accompany elevated volatility, in both directions. It is the difference between a stock the market has stopped thinking about and one it is actively re-evaluating.
It is emphatically not a direction signal. Thirty headlines about a fraud investigation and thirty about a blowout quarter produce the same News Count. For direction you need the companion metrics: news sentiment scores the tone of those headlines, and sentiment trend tells you whether the tone is turning.
The pipeline is short and worth knowing in full, because every step shapes the number you see.
We query a public news feed for headlines about the ticker, keep only those published in the trailing seven days, and cap the result at twenty. Each surviving headline is then passed to a language model that scores its tone. News Count is the number of headlines that came back with a usable score.
Three consequences follow directly, and none of them are hidden from you here. First, twenty is a ceiling, not a coincidence: any ticker in a genuine media storm reads twenty, and so does one merely having a busy week. The metric saturates, and above the cap it stops carrying information. Second, it counts scored headlines, not fetched ones — if the scoring model chokes on a headline, that headline silently does not count. Third, there is no deduplication: a single wire story picked up by eight outlets counts eight times.
Example (illustrative — invented numbers chosen to show the mechanics, not a live ticker). Company A shows a News Count of 20. Company B shows 4.
The tempting reading is "A is five times as newsworthy as B." The defensible reading is much weaker: "A is at or above our ceiling, and B is quiet." A might have had one significant announcement that thirty outlets rewrote, or twenty genuinely distinct stories — the number cannot distinguish those, and it would read 20 either way even if the true count were 200.
What you can legitimately do with this: notice a change. A ticker that normally shows 3 and today shows 20 has had something happen to it. That is a prompt to go and look at what — not a conclusion in itself.
Syndication inflation is the big one. Financial news is heavily recycled: one press release becomes a dozen near-identical stories within an hour. Because we do not deduplicate, that single event can consume most of the twenty slots. The count measures coverage volume, which correlates with — but is not the same as — the number of things that actually happened.
The cap cuts both ways. It protects us from unbounded cost and it destroys information at the top of the range. Any comparison between two tickers both reading 20 is meaningless. Comparisons in the low single digits are more trustworthy than comparisons near the ceiling.
Large companies are structurally noisier. A mega-cap gets written about constantly for reasons that have nothing to do with any new information. Comparing raw News Count across companies of very different sizes tells you mostly about their size. Comparing a company to its own recent normal is the more honest use.
And a zero is ambiguous. It usually means a genuinely quiet week. It can also mean the news feed returned nothing for that ticker symbol — a plumbing failure and a quiet market look identical from here. If a ticker you know is in the news reads zero, believe the news, not the counter.
The number of news headlines about a ticker, published in the last seven days, that we found and successfully scored for tone. It measures attention, not direction — it says nothing about whether the news was good or bad.
Because we cap the number of headlines we process per ticker at twenty. Any ticker at or beyond that ceiling reads 20, so a reading of 20 means "at least twenty" and comparisons between two tickers both showing 20 carry no information.
No. A scandal and a record quarter generate headlines just the same. Use the news sentiment score for tone and the sentiment trend for whether the tone is improving or deteriorating.
Yes. There is no deduplication, so one wire story republished by several outlets counts once per outlet. This is the main reason a single event can push the count to its ceiling.
Browse all S&P 500 tickers to see this metric applied to individual companies.
Educational research only — not investment advice.