Designing a Survey That Can Actually Find Something

Ravi Nemade
October 14, 2025

The survey that was designed to fail

Picture a narrow gold vein, five meters wide, sitting under a reconnaissance grid with lines 500 m apart and samples every 100 m. The probability of a single sample landing on that vein is close to zero — not because the crew did anything wrong in the field, but because the survey was never built to find something that size.

This is the uncomfortable truth about soil geochemistry: the survey is usually won or lost before the first sample bag is filled. Everything downstream — QA/QC, interpretation, drill targeting — depends on decisions made at the design table.

What decides success before fieldwork begins

A handful of planning-stage factors determine whether a survey has a real chance:

Geological understanding

the wrong deposit model produces the wrong survey design entirely.

Sampling medium selection

the wrong horizon (see Post 2) produces a weak or misleading signal.

Target size

a grid too wide simply misses the anomaly.

Pathfinder selection

analyzing the wrong elements makes a real anomaly invisible in the data.

Terrain understanding

transported cover can invalidate assumptions the whole survey was built on.

Orientation survey

skipping this step is one of the most common causes of failed exploration programs.

None of these are field execution problems. They’re design problems, and design problems don’t show up until the results come back wrong — by which point the budget is already spent.

Grid spacing has to respect the target, not the budget

This is where the math gets unforgiving. If an anomaly is 50 m wide and your sample spacing is 100 m, it’s entirely possible for the pattern to look like: sample — anomaly — sample, with no sample point ever intersecting the target. That’s not bad luck. That’s a false negative built into the survey design.

Target geometry drives the numbers directly:

The general rule: sampling interval should be less than half the expected anomaly width — preferably closer to a quarter of it. Skipping this calculation is one of the most expensive mistakes in exploration, because it’s invisible until someone asks why a promising-looking area came back “clean.”

Why sampling lines should cross the strike, not follow it

Sample lines run parallel to geological strike stay inside the same rock unit the whole way — host rock the entire traverse, with little contrast to detect. Lines run perpendicular to strike cross host rock, alteration zone, mineralized zone, and back to host rock, maximizing the chance of catching real contrast in the data.

This is a simple orientation decision, but it’s one that gets overlooked surprisingly often on greenfield ground where the structural grain isn’t obvious yet — which is exactly why geological mapping has to come before, or alongside, survey layout.

Getting this right across varied terrain is part of the daily work on our Tanzania licenses — structural orientation looks different walking a shear zone near Nzega than it does across a graphitic horizon in Lindi, and survey design follows the geology, not a template

When reconnaissance is a waste of money

Reconnaissance surveys — wide-spaced, low-cost, first-pass — have their place, but they fail predictably when the expected anomaly is smaller than the sampling interval. Beyond target size, reconnaissance can also fail because of:

Deep cover, where the signal never reaches the surface at all.
Urban or infrastructure contamination, which masks natural patterns.
Transported overburden, where there's no meaningful soil-bedrock relationship to sample.
The wrong exploration model, where the crew is searching for a deposit type that isn't there.

The underlying lesson holds across all of these: survey scale must match target scale. A reconnaissance grid designed for a broad porphyry system will reliably fail to detect a narrow gold vein system, regardless of how carefully the fieldwork is executed.

Download the Investor Deck →
to see how this scale-matching discipline shapes our program design across the Kahama, Nzega, and Geita gold clusters.

Designing around cover: transported ground and buried targets

Two related design challenges deserve their own attention, because they’re common across greenfield terrain like much of Tanzania’s exploration ground.

Transported alluvium

Before designing a soil survey over transported cover, the first task is establishing the thickness of that cover and the nature of the transport — river, wind, or glacial. Standard soil sampling can fail outright here, because soil chemistry may bear no relationship to the underlying bedrock. Alternatives include sampling deeper horizons (C horizon or saprolite), auger drilling to reach residual material, interpreting the geochemistry of the transported media itself for transport direction, and integrating geophysics — often essential in these settings. The governing question before interpreting any anomaly: what is the provenance of this soil?

Deeply buried deposits

A soil survey can still be effective over a buried deposit, but effectiveness drops with depth. Shallow deposits (0–50 m) are often detectable; deposits beyond 100 m become much harder. Residual cover transmits signal better than transported cover. Mobile pathfinder elements — arsenic, antimony, mercury — can still reach the surface even when the target commodity itself stays put at depth. And structural pathways like faults and fractures can act as conduits, letting some genuinely deep deposits still generate a detectable surface signature.

Design first, drill later

A perfectly executed field program built on a flawed design will still miss the target. This is why, on every new license — whether it’s a fresh gold target near Jomu or an early-stage graphite assessment in Tanga — the design phase gets as much rigor as the fieldwork itself.