Fisher made phone calls. The modern analyst has a richer, faster, and more treacherous toolkit, treacherous because each new source has a characteristic lie you must know. Here is the professional stack, what each reveals, its limit, and the part that matters most for you: the free approximation a solo learner uses to reconstruct the same signal without a five-figure subscription.
What does each source reveal and where does it mislead?
| Tool | What it reveals | Its characteristic limit | The free solo-learner approximation |
|---|---|---|---|
| Expert networks (Tegus, AlphaSense, GLG, Third Bridge) | Structured interviews/transcripts with ex-employees, customers, competitors, industry experts: the fastest way to primary qualitative insight | Expensive (institutional pricing); experts self-select and may be out of date; compliance/MNPI risk; you hear what the expert believes, not always what is true | Read what free experts already published: competitors' earnings calls and investor decks (rivals describe your target better than it describes itself); industry-association reports; trade press; sell-side initiations (often free via broker portals); podcasts (Acquired, Business Breakdowns) that are recorded expert interviews |
| Channel checks | Real-time demand and pricing at the point of sale: inventory, discounts, footfall, availability, staff morale | A snapshot, easily unrepresentative (one store, one week, one city); confirmation bias in what you go looking for | Do them yourself. The biggest free edge an individual has. Visit the stores (a Tanishq and a Kalyan and a local goldsmith in one afternoon); pose as a customer; note discounts, footfall, what staff push; call dealers as a prospective buyer |
| Glassdoor / AmbitionBox / Blind / Indeed | Culture, attrition, morale, management quality (Fisher 7–9), often the earliest signal of decay | Selection bias (angry leavers over-post); ratings inflate/deflate around events; gameable by HR | Free by design. Read the trend and the text, not the star average; weight consistent specific complaints (e.g., "sales targets forcing channel-stuffing") over generic gripes; compare the target vs its direct peers on the same site |
| App store reviews & ratings | Customer sentiment, product quality, and the trajectory of a digital product; complaint themes | Vocal-minority bias; review-bombing; irrelevant for non-app businesses | Free. Read the recent 1- and 2-star reviews for recurring specific failures; track the rating trend across app versions; for Titan, the CaratLane and Tanishq/Encircle apps; for Copart, the buyer app |
| Web-traffic & alt-data (SimilarWeb, Sensor Tower, Similarweb, credit-card panels, satellite) | Traffic/engagement/download trends as a revenue proxy; footfall from geolocation; even parking-lot satellite counts | Proxies drift from revenue; coverage gaps; expensive at the useful tiers; can badly mislead on B2B or omnichannel firms | Free tiers + public proxies: SimilarWeb free tier and Google Trends (search interest for the brand/category as a demand proxy); LinkedIn company page (employee-count growth by function, a free hiring-trend alt-data feed); job-posting counts; download-rank trackers' free views |
| Social / forums (X-FinTwit, Reddit, MouthShut, Trustpilot, YouTube) | Unfiltered customer and (sometimes) employee voice; the bear case in the wild (see steelmanning the bear case) | Noise, astroturfing, pumpers and shorts talking their book | Free. Mine for specific recurring complaints and for the strongest bear points; ignore price-target chatter |
Advertisement
What rules apply when using these tools?
Three rules govern the whole toolkit, and they matter more than any single source:
- Triangulate; never trust one source. A single Glassdoor rating is noise; a two-year decline in Glassdoor and rising attrition disclosures and a wave of LinkedIn exec exits and customers complaining about service is a signal. The mosaic is the method: each tile is weak, the assembled picture is strong.
- Know each source's built-in lie. Angry ex-employees over-post; review-bombers cluster; web traffic is not revenue; a channel check is one store on one day. You are not looking for a source without bias, because there isn't one. You are correcting for the bias you know each carries.
- Separate signal from tradable-but-illegal. Scuttlebutt assembles non-material public facts into a material mosaic, and that is legal and is the whole game. Being handed one material, non-public fact by an insider is not scuttlebutt; it is a crime (the ethics thread). The line: a customer telling you the stores feel busy is scuttlebutt; a manager telling you unreleased same-store-sales numbers is inside information. Stay on the mosaic side of the line, always.
Can you do this without paid tools?
The liberating truth for a solo learner: the free stack recovers most of the paid stack's signal if you are diligent. You cannot afford Tegus, but you can read the competitor's call (where a rival describes your target's weaknesses for free), walk the stores (a channel check no institution's analyst has time to do as often as you), and mine Glassdoor and app reviews (the same raw text the paid tools repackage). What you lack in budget you make up in legwork and time-horizon.
Fisher built a fortune on shoe-leather; the internet made the shoe-leather cheaper, not obsolete.