Satellite imagery vs. expert networks for tracking capacity expansion
Every industrial analyst covering cyclical names has run the same playbook: book a few expert network calls with former plant managers or regional sales reps, ask leading questions about utilization, and try to triangulate whether a competitor is adding lines before the next earnings call confirms it. It works, sometimes. It also has a shelf life, because expert networks and satellite imagery aren't solving the same problem, even though they get lumped together under "alternative data."
What an expert network call actually gives you
A good channel check gets you intent and sentiment. A regional sales manager might tell you a customer mentioned a new supplier is quoting higher volumes, or that a plant has been running weekend shifts since March. That's useful color, and it often arrives months before any public disclosure.
But it's secondhand, and it's bounded by who picks up the phone. The consultant on the call left that plant eighteen months ago. The compliance wall at most expert network firms keeps you from asking anything that sounds like it's targeting nonpublic information, which means the questions get vaguer the closer you get to the thing you want to know, like whether a specific site has broken ground on a new line. You end up with a probability, not a fact. Three corroborating calls feels like confirmation, but it's still three people's impressions of a plant they may not have walked through recently.
What imagery gives you instead
A quarterly image of the site itself doesn't care what anyone remembers or is willing to say. If a new building pad is poured, graded, or framed between one pass and the next, it shows up. If the yard is stacking more finished-goods inventory or raw material than it was two quarters ago, that's visible too. The site either has new structure on it or it doesn't, and that's a different kind of answer than "I heard they might be expanding."
Where imagery falls short is intent. A cleared lot doesn't tell you whether the company is building a warehouse or a production line, and a stalled project doesn't tell you whether that's weather, permitting, or a capital allocation decision made in a boardroom you'll never see into. That's exactly the gap a channel check fills: someone who's talked to procurement or local contractors can usually tell you what's going up, and why the pour stopped in January.
Where this leaves an analyst building a model
Primary research and imagery each catch what the other misses. Expert network calls are strong on motive and context but weak on current physical state. Imagery is strong on physical state, repeatable, and doesn't degrade as a source relationship ages, though it stays silent on why.
That means imagery works less as a replacement for channel checks than as a way to stop running them blind. If you already have a short list of sites where a competitor might be expanding, a per-site image pass tells you which of those sites show ground-breaking or new line construction, so the next round of expert calls can focus on the two or three where something is visibly happening instead of a dozen where nothing's changed. It also gives you a record to go back to when a call conflicts with another call. The image doesn't have an opinion.
For a sell-side or buy-side team covering a handful of named competitor sites, the cost of imagery tracking is small relative to the cost of a wrong capacity call baked into a model a quarter too early or too late. It's also one of the few inputs in this stack that doesn't thin out the way a source network does when someone changes jobs or stops returning calls.
If you're tracking specific industrial sites and want a standing view of what's being built there before it shows up in a press release, Capacity Expansion Tracking follows named plants quarter over quarter and flags what changed on the ground.
Pair that with your existing channel checks and you'll spend less time on calls to sites where nothing moved.