Learn · The Twelve Measurements — Lesson 12 of 12
Social sentiment in markets, explained
After January 2021, no serious market observer dismisses retail chatter as noise: coordinated attention moved prices double digits in days, and it did so visibly, in public, before the moves completed. The question stopped being whether social data matters and became how to measure it without being fooled by it.
What it measures
Reads Reddit and StockTwits activity, scored relative to each ticker's own trailing median rather than against other tickers — a thousand posts is quiet for one name and a spike for another. The smallest share of the composite, alongside volume surge.
Farlens reads two venues with usable structure — Reddit's finance communities and StockTwits, where posts are ticker-tagged — and scores each name's discussion against its own trailing baseline. A thousand posts is a quiet day for NVDA and an five-alarm spike for a mid-cap; only the deviation from a name's own normal means anything. Both the volume of discussion and its tone feed the reading.
Why it moves prices
Attention precedes participation. Before retail money moves, retail conversation moves — the chatter spike is upstream of order flow, which is upstream of price. That chain is why the measurement earns a place at all. But the relationship is two-sided: moderate, building discussion has historically accompanied genuine interest, while explosive spikes have more often marked crowded, late-stage moves — maximum conversation tends to arrive near maximum commitment, which is precisely when the marginal buyer is running out. The measurement grades the spike's shape, not just its size.
How to read it
Positive: discussion is leaning constructive relative to normal. Negative: leaning negative, or spiking in a way that has historically accompanied crowded positioning.
It carries the smallest share of the composite, tied with volume surge, and the modesty is deliberate: chatter is the noisiest data in the stack and the only kind that can be manufactured cheaply. Read it as colour on the participation story — loudest when it diverges from price, as when a falling name's discussion turns constructive, or a rising name's turns to exit talk.
How Farlens uses it
Social chatter carries 5.2% of the Farlens composite (weight 0.6 of 11.6 across all twelve measurements). Coverage: Needs enough discussion to measure. Thinly discussed names, most non-US listings and indices are skipped. Bare tickers that are common words (NET, NU, ARM, COIN) need special handling to avoid false matches.
Two conventions matter here. First, a measurement that cannot be observed for an instrument is skipped, never counted as zero — a zero would read as “neutral” and quietly dilute the composite. Second, on any given day a contribution is the measurement's score times its weight, divided by the total weight of the measurements present that day — so the published breakdown always sums to the score beside it. The full stack, with every weight published, is on the signals page.
Why it is absent from the example reading
Farlens publishes one frozen example reading — AAPL on 22 July 2026 — and this measurement is not in it. That is the honest answer, not a gap: Needs enough discussion to measure. Thinly discussed names, most non-US listings and indices are skipped. Bare tickers that are common words (NET, NU, ARM, COIN) need special handling to avoid false matches.
This is a deliberate property of the methodology. A measurement that cannot be observed for an instrument returns nothing, and the composite is computed from the measurements actually present that day — their weights rescale so the reading still sums correctly. Treating “unmeasurable” as “neutral” would quietly dilute every score it touched.
Common words make treacherous tickers: NET, NU, ARM and COIN appear in millions of posts that have nothing to do with the companies, so the pipeline matches deliberately narrowly for such names. Missing some genuine chatter beats counting the phrase “fishing net” as market sentiment — the same honesty rule as everywhere else: no reliable measurement, no reading.
What it cannot tell you
Chatter can be gamed — pump groups exist precisely because attention moves thin names — which is why this measurement is small, baseline-adjusted and never a lead voice. It reads English-language, US-centric venues, so its coverage of Gulf names and most non-US listings is honestly thin to absent. And tone classification is imperfect: sarcasm, memes and irony defeat any classifier some of the time. The design response to all three is the same — weight it modestly and skip it when it cannot be measured.
Frequently asked
Did meme stocks prove social data works?
They proved attention can move prices — and also that by the time chatter peaks, much of the move has often happened. Both halves of that lesson are built into how the spike's shape is graded.
Which platforms are read?
Reddit's finance communities and StockTwits — venues where discussion is public and ticker-tagged. Platforms without usable structure would add noise faster than signal.
Why is it the smallest weight alongside volume surge?
Because validation, not enthusiasm, sets the weights. It adds real information at the margin and would mislead as a headline measurement — the weight reflects that balance.