The Recurring Failure Modes
An empiricist learns more from a failure than from a success, and it is worth being precise about why. When a project succeeds, the causes are overdetermined. Maybe the site was genuinely good, maybe the host agreement was well-crafted, maybe the opposition happened to be disorganized that year, maybe the commission had a pro-development majority for unrelated reasons. You cannot easily separate the decisive factor from the lucky ones. A failure is more honest. When a project dies at the planning commission, the cause of death is usually identifiable, because something specific broke the deal, and the record of the hearing often tells you what it was. So let me lay out the failure modes that recur, drawn from the pattern rather than from any single dated case, because the point is the shape of the mistake, not the scoreboard of a particular county.
The first and most common failure mode is pricing consent at zero. The developer treated land-use approval as a formality, sequenced it last, and arrived at the hearing having done no real community work. This is the empty-room assumption applied to a room that is no longer empty. The tell is a project that looks strong on every quantified axis, power, fiber, price, incentives, and collapses at the one axis nobody modeled. When you read the autopsy, the fatal line item was always there; it was simply never entered into the spreadsheet. The lesson is the one from the first module: consent is a constraint with observable proxies, and refusing to price it does not make it cost nothing.
The second failure mode is answering the wrong concern. Here the developer did do community work, but built the package around the wrong fear, a generous offer aimed at issues the community did not actually hold, while the issue it did hold went unaddressed. The classic version is spending on visible amenities while ignoring water or electricity-rate concerns that were the real driver of opposition. This is a listening failure upstream of a spending failure. The money was real; it was simply aimed at the wrong target because engagement was informational rather than investigative. A dollar spent on the concern residents do not have is worse than no dollar at all, because it signals that you were not paying attention.
The third failure mode is gaming the permitting path and getting caught. The developer found a procedural route that minimized public input, filing quietly, characterizing the project vaguely, choosing the path with the least discretionary review, and the community discovered the maneuver. The result is not merely a lost vote but a poisoned relationship, often expressed as a moratorium that stops every subsequent project too. This is the failure mode with the longest blast radius, because it converts a single project's problem into a jurisdiction-wide barrier. The lesson is that procedural cleverness is a loan against future consent, and the interest rate is punitive.
The fourth failure mode is timing the market wrong on opposition. The developer assumed the opposition would be as disorganized as it was a few years ago and was met instead by residents who knew the comprehensive plan cold, coordinated with groups in neighboring counties, and executed the continuance-and-study delay tactic with precision. This is a failure of reading the room's current state rather than its remembered state. Opposition capability has risen sharply and unevenly; a developer working from an outdated model of how sophisticated the resistance is will be repeatedly surprised, and surprise at a public hearing is expensive.
The fifth failure mode is the site that simply should not have been built there. This one deserves honesty. Not every loss is a process failure that better community relations could have salvaged. Sometimes the water genuinely is not there, or the grid genuinely cannot absorb the load without raising rates, or the place is genuinely unsuited to a large industrial campus, and the community's refusal is the correct answer to a real constraint. Reading this failure mode correctly matters because the wrong lesson, that the project just needed better messaging, sends the developer to lose again at the next unsuitable site. The right lesson is that consent sometimes fails because the underlying project is bad, and the honest move is to hear the no as accurate information about the site rather than as an obstacle to be overcome.
What unites these failure modes is that each one is a constraint that was observable in advance and was not priced. That is the empiricist's consolation and the empiricist's discipline in a single idea. The losses are legible, which means they are teachable, which means the cost of the education has already been paid by someone else. Your job is to read the autopsy before you file, not after you lose. In the next lesson we turn the same lens on the wins, carefully, and ask what they actually demonstrate once we strip out the luck.