Ask a long-tenured site manager why the monitoring well in the northeast corner was abandoned in 2004, and you will usually get an answer in about four seconds. It kept silting up. The driller hit a sand seam. Two rounds of unusable data, and the team gave up on it.
Now go look for that in the file. You will find the well in the 2003 sampling table and absent from the 2005 one. Nothing explains the gap.
This is the shape of the institutional knowledge problem for site owners. The documents record conclusions, and the reasoning behind those conclusions was never written down anywhere.
Two kinds of knowledge live on every site
The first kind is documented. Reports, lab certificates, permits, regulatory correspondence, sampling tables. It is auditable, it transfers with the file, and it is what a purchaser's counsel will ask for during diligence.
The second kind is tacit. It is the accumulated judgment of everyone who has worked the site:
- Why a delineation stopped at a particular boundary rather than continuing
- Which historical result the team never trusted, and why
- What the regulator said on a phone call that shaped the next three years of work
- Why one remedial option was dropped before it ever appeared in an options analysis
- Which of the four spellings of a location name refers to the same well
None of that is secret. Most of it was discussed openly, decided sensibly, and simply never made it into a deliverable, because deliverables document what was found rather than what was considered and set aside.
Tacit knowledge is also the kind that leaves. A consultant loses a recompete. A project manager retires after 28 years. A corporate reorganization moves a portfolio to a different business unit. The documents stay exactly where they were. The interpretation of them walks out the door.
What the loss actually costs
For a corporate portfolio holder, the bill arrives in familiar forms.
Work gets repeated. A new consultant re-investigates a boundary that was already characterized, because the file does not explain why the previous team stopped where it did. The data supported the decision. The rationale was verbal.
Conservative assumptions replace evidence. Faced with an unexplained gap, a reviewer assumes the worst case. That assumption flows into the cost estimate, the reserve, and eventually the disclosure. Uncertainty is expensive even when the underlying site is fine.
Old errors get reinforced. A result that everyone on the original team knew was a lab artifact gets cited in a new report as though it were reliable. Nobody left knows to flag it. Ten years on, it is part of the accepted site narrative.
Diligence takes longer than it should. A transaction timeline built around three weeks of file review stretches when the file review produces more questions than answers, and there is no one left to ask.
None of these are catastrophic on their own. Across a portfolio of a hundred sites and a thirty-year holding period, they compound into real money and real schedule.
Why the usual fixes do not hold
The standard response is a knowledge transfer exercise: interview the departing expert, write it down, file it somewhere.
The intent is sound and the execution rarely survives contact with reality. Exit interviews happen after the person has mentally moved on. The resulting memo lands in a folder nobody opens, because nobody knows it exists. And the exercise only captures what the departing person thought to mention, which is not the same as what the next person will need to know.
The deeper problem is that tacit knowledge is organized by situation, not by topic. It surfaces when someone asks the right question at the right moment, and the right moment is usually years after the person who knew the answer has gone.
Capture it where the question gets asked
A more durable approach starts from a different premise: the place to record context is against the evidence it explains, not in a separate document.
Statvis builds a queryable knowledge base from a site's full document corpus, with every answer linked to the source page it came from. That corpus handles the documented half of the problem. The tacit half is addressed by letting context attach directly to the record.
- Page notes let a professional annotate the specific page where a questionable result appears. The note travels with the evidence and surfaces whenever that page turns up in a search, rather than sitting in a memo nobody retrieves.
- Location deduplication resolves the naming inconsistencies that experienced staff carry in their heads. When MW-4, MW-04, and Monitoring Well 4 are recognized as one location, that knowledge stops depending on tenure.
- Timeline reconstruction distinguishes when an event occurred from when it was reported, so the sequence a long-tenured manager holds intuitively becomes explicit and checkable.
- Cited answers to plain-language questions mean the successor can ask what the predecessor would have been asked, and get an evidence-backed response instead of a shrug.
This does not capture everything. Some judgment is genuinely irreducible and belongs in a conversation with an experienced professional. The goal is narrower and more achievable: make sure that the questions with documented answers get documented answers, so the limited time of experienced people goes toward the questions that actually require them.
The practical test
Pick one site in your portfolio where the primary contact has changed in the last five years. Ask the current manager three questions about decisions made before their tenure, and see how many can be answered from the file alone.
Whatever that exercise reveals is your institutional knowledge position. It is worth knowing before a transaction, a regulatory request, or a retirement makes it urgent.
The documents are an asset you already own. The context that makes them mean something is an asset too, and it is the one most portfolios are quietly losing.