● Method
The discipline that matters in this work is not finding things. It is being rigorous about the difference between what you know and what you have assumed — and being willing to report that the answer is nothing.
We start from the timeline the family or the agency already has and take it apart. Each point gets sorted into one of three buckets: confirmed by a record or a document, reported by a person, or assumed because it seemed obvious at the time. The third bucket is usually larger than anyone expects, and its contents have often been quietly driving the whole search.
At the end of intake we tell you whether open sources can realistically add anything. Sometimes the honest answer is that the case turns on phone records or bank data that only a subpoena reaches, and that no amount of public-data work will substitute. We would rather say that at the start than bill for a month of it.
Everything we pull is recorded with where it came from and when it was captured. The date is not bookkeeping — it is load-bearing. Elevation data may have been flown years ago, land cover is updated on a multi-year cycle, and crowd-sourced camera registries are a snapshot that was already incomplete on the day it was taken.
A finding built on a source that predates the disappearance has to carry that caveat all the way through to the final document, because the ground may simply no longer look the way the data says it does.
Geospatial and temporal modelling over real road, terrain and hydrology data. Then the part that actually determines whether the work is any good: we try to break our own result.
Every ranking gets checked against independent evidence — usually current satellite imagery — before anyone sees it. We ask what a scoring rule is really rewarding, because a plausible formula can quietly rank the wrong thing first. A model that favours proximity to a road, for instance, will put roadside ditches at the top of the list; those are exactly the places that get searched in the first hour, so ranking them highly is worse than useless.
When a check changes the answer, the change goes in the record. We would rather hand you a short list we have tried to destroy than a long one we have only admired.
Deliverables are built for the person standing at a treeline with a phone, not for a boardroom: ranked search areas with coordinates, maps, and printable worksheets that say what to look for at each location and what would rule it out.
Every product carries a confidence ledger — what is established, what is inferred, what remains unknown, and what would change the conclusion. With your consent we route findings to the agency of jurisdiction. We do not act on them ourselves, and we do not publish them.
A limit worth stating plainly A ranked search area is a screening tool, not a prediction. It says “this ground is worth the hours” — never “they are here.” Anyone who tells a family otherwise is selling something.
There is no charge to ask, and we will tell you honestly whether open sources can still add anything to the case.
