Primary Signals: Getting Underneath the Analysts

Analyst research is secondary by nature. Someone gathered, interpreted, and packaged information, and the package reached you after passing through their franchise and their tilt. The most powerful move a practitioner can make is to get underneath the package to primary signals, the raw traces of activity that analysts themselves are reading. You will rarely have the resources of a research house, but you can often check the analysts against reality more cheaply than you think.

Start with what "primary" means here. A primary signal is evidence of activity that has not yet been interpreted by an intermediary: a regulatory filing, a permit application, a job posting, a shipment record, a public tender, an interconnection queue, a patent, an earnings-call transcript. These are the fingerprints of what companies are actually doing, as opposed to what they or their analysts say they are doing. They are messy, incomplete, and require work to read, which is exactly why they are valuable, because the work is a barrier that keeps most people reading only the packaged version.

Consider a few concrete examples. If you want to know whether data-center construction in a region is really accelerating, the analyst report and the broker narrative are two views, but the power-utility interconnection queue and the local permit filings are primary. A surge of large-load interconnection requests is companies putting their intentions on the record with a utility, which is harder to fake than a press release. Hiring is another window: a company staffing up a particular function is voting with payroll, and job postings are public. Earnings calls are gold, because executives are legally constrained in what they can say and analysts on the call push on the specifics; the transcript often contains harder information than any downstream report. Import and shipping records, customs filings, and procurement tenders reveal physical flows of goods that marketing cannot easily inflate.

The discipline is to use primary signals not to replace analyst research but to test it. When an analyst forecasts explosive growth in a category, ask what primary signal would have to be true if that forecast were right, and then go look for it. If the hiring, the filings, the shipments, and the tenders are consistent with the forecast, your confidence rises and it rises on independent evidence. If the forecast is soaring while the primary signals are flat, you have found something important: either the forecast is ahead of reality, or the growth is taking a form the primary signals you chose do not capture, and either way you now have a sharper question than you started with.

There is a reverse discipline too. Primary signals without a frame can mislead as badly as analyst reports without scrutiny, because a single filing or posting is anecdote, and anecdote generalizes poorly. The value comes from volume and pattern: not one permit but the trend in permits, not one job posting but the shape of hiring across a sector. This is where analyst research and primary signals genuinely complement each other. The analysts provide the frame, the category, the baseline, the historical trend. The primary signals let you check whether the frame still fits the present. Neither is sufficient alone, and the practitioner who holds both is far ahead of the one who has only the packaged report or only a handful of raw data points.

A practical caution on effort. Chasing primary signals has a cost, and the skill is proportioning the effort to the stakes. For a routine decision, reading two or three analyst views against each other may be enough. For a decision where a lot rides on the answer, and especially where the consensus analyst view is being sold to you by someone who benefits, the investment in primary signals pays for itself, because that is exactly the situation where the packaged view is most likely to be shaped and most expensive to get wrong. Learn to feel where the stakes justify getting underneath, and get underneath there.

The habit to build is a reflex of suspicion aimed not at the analysts, who are mostly competent and honest, but at the layer of interpretation itself. Every packaged number was once a set of primary traces, and those traces are often still findable. The practitioner who reads them, even partially, even occasionally, stops being purely a consumer of other people's conclusions and starts being able to check them. That capacity, more than access to any single source, is what separates a reader who is moved by coverage from one who reads it.