It is the balance of bullish and bearish calls that the analysts CryptoQuant tracks make about Salesforce (CRM) in their public posts on X, CryptoQuant QuickTakes, Seeking Alpha and Substack, and in the licensed ratings of sell-side analysts. In the last 7 days, 48 analysts posted on or rated Salesforce (CRM); 3 of their posts and ratings were directional calls (bullish or bearish; neutral posts and hold-type ratings are not calls), as of 2026-10-21 12:00 UTC; the Analyst Consensus Index reads 92% bullish / 8% bearish: a 7-day average in which each of the 25 analysts with a live directional stance on 2026-10-21 counts once (23 bullish · 2 bearish). The bear and bull cases and the narratives above are built from those posts; narratives found only in news coverage are listed separately under From News.
Every post by a tracked analyst is labelled per asset as bullish, bullish-nuance, bearish, bearish-nuance or neutral by a language-model pipeline. The index then gives each analyst one vote on each day: the recency-weighted average of their directional posts from the last 30 days (a post counts in full for 14 days, then fades out by day 30), weighted by the analyst's accuracy score; neutral posts do not vote. On equities the latest bullish or bearish sell-side rating of each covered analyst votes the same way for up to 90 days (hold-type ratings do not). It runs from −100 (everyone bearish) to +100 (everyone bullish); the percentage on this page is that index mapped to 0–100%. The chart plots a moving average of the daily index whose window is the selected period (30 days on the default view) and colours it by that average's rolling 90-day z-score. Read the methodology →
Analysts on X, Seeking Alpha and Substack enter the roster through a reviewed selection: reach, activity, coverage of a tracked asset and explicit calls, read by a person before anything is loaded. Sell-side analysts enter through the licensed ratings feed on the equities they rate. Every analyst is then followed the same way: all of their public posts are collected and labelled. Company accounts, news feeds and direct stakeholders of an asset are not listed as analysts (news is shown separately under From News). Accuracy is scored per asset against the market phases that followed each call. To be ranked an analyst needs at least 50 days of sentiment data, participation in at least 2 bull and 2 bear market phases, an active posting history (not inactive for 6+ months), at least a year between their first and last post and at least 10 directional calls on the asset. The Top 1%, Top 5%, Top 10% and Top 25% badges are the analyst's percentile inside the scored pool (Top 1% only for the #1 of a pool of 100 or more); an analyst with fewer than 30 directional calls carries no badge.
New posts are collected from X every 4 hours, from news sources every 30 minutes and from Seeking Alpha and Substack twice a day, and labelled right after collection; on equities the licensed sell-side ratings are polled every hour. The daily index, the narrative clusters and the bear / bull summaries are rebuilt once a day from 00:00 UTC, and the index is recomputed again after each platform-feed and ratings run. An hourly index is kept for BTC, ETH and SOL. The figures on this page are served from a cache that refreshes within the hour.
Consensus View lists narratives shared across several posts that include at least one tracked analyst. Standalone Analyst Views lists individual analysts whose own directional calls in the period lean one way (at least one bullish or bearish call, 60% or more in the same direction), one card per analyst, quoting the post that most directly discusses the asset. From News lists narratives that come only from news outlets.
No. Everything on Analyst Consensus (sentiment labels, consensus indexes, narratives, accuracy scores and the analysts’ own posts) is provided for information only. It is not investment, financial, legal or tax advice and not a recommendation to buy, sell or hold any asset; past accuracy of an analyst does not predict future results. Terms →
