Two people read the same annual report. One comes away reciting the profit figures; the other comes away having spotted an inflection. Why?
The Lag Is Structural
Annual reports arrive once a year, quarterly reports every three months. Before a number lands on the income statement, it has to travel a whole chain in the real world: production, shipping, sale, collection. Every step takes time. By the time the report confirms a boom, the ground may already have cooled; by the time it confirms a decline, the capacity may already have cleared out.
This is not slow disclosure. It is structural — the balance sheet records the past, while prices trade the future. Research notes look fresher, but most of them merely reshuffle the same old numbers: secondhand lag.
So the lag sits there. What closes the gap?
What to Look At
There is a system that puts this rule in hard terms: every heavy position must be visited in person. What to look at? Four settings, four answers.
At factories, look at capacity. At BYD’s plant in Anhui: are the lines running at full load, are the logistics trucks queuing? The report will not admit any of it for another six months.
At stores, look at foot traffic. For restaurant brands, queue length and table turnover are visible at a glance. Stand outside a bubble-tea shop for ten minutes; counting cups is more honest than reading ten research notes.
At hospitals, look at usage. Whether a drug or a consumable is actually selling — prescriptions and inventory speak first.
At distribution channels, look at inventory. For coaches, the delivery rhythm of overseas dealers; for consumer goods, how much stock the channel is carrying. At Xiang Piao Piao, at Tims, at Dong-E — the water level of channel inventory touched the inflection point before the report did.
The purpose of fieldwork is not to collect evidence for a conclusion already held. It is to feel real supply and demand, real consumer behavior. Fieldwork that starts from a conclusion is tourism, not research. Ten years in front of the candlestick chart teaches less than one day standing on a shop floor.
At this point the objections arrive — more than one of them.
Two Rounds Against
First round, the privilege objection: ordinary investors have no way into factories or meetings with executives; field research is an institutional luxury.
Half right — access is not universal. But the objects of research are observable public settings: store traffic, channel inventory, hospital usage, delivery rhythm. Anyone can go and look. No inside sources, no executive meetings. A single data point decides nothing; crossed readings from many points are what count as perception. The privilege objection defeats a posture, not the substance.
Second round, the sample objection, and it is stronger: what you see with your own eyes is itself a sample. One busy store does not represent the country; one full workshop does not represent an industry. Sample bias is more dangerous than reporting lag.
True — and the conclusion is drawn the wrong way. The remedy for sample bias is not retreat into lag; it is more samples. Three stores in one city, ten stores in three cities, factories crossed with stores, hospitals, channels. A financial report is one large sample plus one quarter of delay; fieldwork is many small samples plus real time. The slow one is the past, the fast one is the present — holding both beats holding either.
The One Beat That Matters
One question remains: reports plus research notes are far easier. Is the legwork worth it?
Worth it — because the legwork buys exactly that beat. Research notes are secondhand lag; statements are the past of the past. Channel inventory called the inflection before the report did, and the beat it gained is the margin of safety.
A statement answers: how did the business score on last period’s exam. Fieldwork answers: is this business still alive right now.
One boundary: no stock picks, no price targets. The companies named here serve as scene-setting examples only, not as any current buy or sell judgment.
