Why depth, not breadth

Afrimintel covers 113 worked African mining dossiers — each with a named dossier-level primary source, anchored by a field-structured intelligence-grade core — across 13 mineral provinces — producing mines and advanced-development projects alike, every figure traced to a named primary source. This is a deliberate editorial choice, not a coverage limitation. Here is the case for it.

The benchmarks Afrimintel is measured against

Senior institutional readers come to Afrimintel from a reading list that includes:

PlatformAfrican deposit coverageDepth model
Wood MackenzieSeveral thousand globally; African subset in the hundredsMulti-analyst desks per commodity; subscription priced for institutional buyers
S&P Capital IQ Mines (formerly SNL)Tens of thousands globally; African subset in the thousandsDatabase-first, with analyst overlay on Tier-1 assets
CRUCommodity-cohort coverage at depth on copper, aluminium, iron ore, fertilisersSpecialist-by-commodity desks
Benchmark Mineral IntelligenceBattery-supply-chain focused; African scope concentrated on Li, Co, Ni, graphiteSpecialist EV / battery vertical
FastmarketsPricing-anchored coverage; African scope where commodity volumes warrantPrice discovery anchor
USGS Mineral Resources OnlineFree; thousands of African entriesSpatial-reference + colonial-era inventory; minimal current operator-level intelligence
BGS AfricaCountry-level mineral atlases; topical reportsContinental-survey-grade; not deal-level

A senior reader will note that Afrimintel's 113 worked (dossier-grade and intelligence-grade) dossiers remain *substantially fewer* than what Wood Mackenzie or S&P offer in this space. That is correct, and the depth-over-breadth case holds regardless of the count. The relevant question is what Afrimintel does with each dossier, and why depth-over-breadth is the governing editorial choice.

The case for depth over breadth

The institutional decision-aid test is decided on the depth tier, not the breadth tier. A capital committee evaluating an investment does not need a database of 1,800 entries; it needs an authoritative, primary-source-verified, field-level-provenanced screening artefact for the asset under decision. The first question is whether the artefact for that one asset is defensible, not whether the platform also has 1,799 others.

Specifically, depth-over-breadth makes sense at this stage because:

1. Primary-source verification does not scale linearly

Field-level provenance — every reserve, grade, operator interest, recent-event entry traced to a named, dated, citable primary document — requires reading the underlying technical reports, not auto-ingesting trade-press summaries. The Kamoa-Kakula 31 March 2026 NI 43-101 Technical Report is approximately 200 pages plus appendices; the Kabanga July 2025 FS Technical Report is similar; the Sukari S-K 1300 TRS effective 31 December 2025 is similar again. The unit cost per dossier of doing this properly is significant. The cost of doing it shallowly across 1,800 entries — which is the only way to scale to that count without a multi-analyst desk — produces *Sourced* claims in name only.

2. The Quality Standard is non-negotiable, and it constrains scale

Afrimintel's three-state model (Sourced / Derived / Absent — no fourth state, published at /quality-standard/) explicitly disallows the synthesis-from-summary pattern that lets large databases scale. Each Sourced claim requires a named, dated, citable primary document. Each Derived claim requires a published methodology and inputs that are themselves Sourced or Derived. Absent is preserved as the honest disclosure when neither tier applies. The framework is incompatible with rapid horizontal expansion; that is by design.

3. The set is not arbitrary — it is positioned across decision-relevant axes

The worked dossiers cover specific commodity-province-jurisdiction combinations that match the decision-shape map institutional capital allocators actually face — with the six intelligence-grade dossiers anchoring the set: Tier-1 copper sulphide (Kamoa-Kakula), critical mineral pre-FID (Kabanga nickel, Ngualla REE), disputed-tenure capital allocation (Manono lithium district), force-majeure recovery operator (Balama graphite), potash under a changed owner (Colluli). Around that core, the dossier-grade tier extends the decision-shape coverage across producing Tier-1 gold, PGE and further critical-mineral assets. The set is curated for *decision-shape coverage* rather than for asset-count maximisation, and it now spans producing Tier-1 assets through advanced-development projects carrying bankable feasibility-study NPVs.

4. The depth tier is where the field-level provenance contract lives

The intelligence-grade dossiers are the only tier where Afrimintel makes the strongest data-quality commitment: every numeric and factual claim is one of three states (Sourced / Derived / Absent), with the source named and dated, and the claim subject to Audit Log tracking when corrected. Spatial-reference entries are *positionally accurate* but do not carry the same provenance contract; they exist to provide geographic context, not to support deal-level decisions. Research-grade entries sit between the two — operator and commodity verified, deeper structural data not yet completed; operator-attestation entries carry operator-disclosed figures without an independent reporting-code framework. The tier system is itself published; the data-quality contract is therefore tier-dependent and explicit.

5. The 13 mineral provinces are scored, regardless of asset count within

The province-scoring framework operates at the geological-terrane level, not the asset level. All 13 African mineral provinces are scored on the published Mineral System Potential / Investment Climate / Composite Score formula, with the formula and inputs published on the methodology page. The province scoring covers 100% of the relevant African geological terrane, even where individual asset coverage is thin. This means a senior reader interested in, say, the West African Birimian or the East African Rift gets the structural geological framing for the entire province, regardless of whether Afrimintel's IG-tier coverage of specific assets in that province is one or ten.

What 1,800 would buy, and what it would cost

To be honest about the trade-off:

What 1,800 would buy: the ability to support broader screening mandates where institutional analysts ask "show me everything currently producing in West Africa" or "show me all African Cu development projects above 10 Mt." For a Wood Mac–style analyst with that workflow, Afrimintel at 113 worked dossiers still does not replace a multi-thousand-asset database.

What 1,800 would cost: either (a) a multi-analyst desk with the editorial capacity to maintain primary-source verification at that scale — which is a scale of operation Afrimintel would reach only after substantial institutional engagement — or (b) the abandonment of the field-level provenance contract, the three-state Quality Standard, and the per-record audit cadence. (b) is not a trade Afrimintel will make; (a) is on the deferred roadmap, sequenced behind commercial validation.

The correct comparison

Afrimintel is not directly comparable to Wood Mackenzie or S&P Capital IQ Mines at coverage breadth, because that is not the comparison being made. The correct comparison is to the screening artefact a senior analyst would produce for one asset, in one decision context — and what that artefact would look like once produced.

The Case Studies section publishes worked demonstrations: Kabanga Pre-FID DFI Investment Brief, Manono Disputed-Tenure Capital Allocation Brief, and Loulo-Gounkoto Post-Settlement Re-Entry Brief. Each is a real, dated, primary-source-verified institutional screening on a live decision. The case for depth-over-breadth at this stage rests on what these dossiers contain, not on what the asset count is.

How the count will evolve

To be honest about the constraint as well as the choice: the worked set has grown from an original core of a few dossiers to 113 records across the dossier-grade and intelligence-grade tiers, of which six are held at intelligence-grade — field-structured reserve/resource against a named primary source and export-verified — with a published verification queue of candidates moving toward that bar. Depth is not uniform across the worked set, and the platform does not claim it is. The intelligence-grade six carry the full bankability and worked-economic treatment; the dossier-grade records are each anchored to a named dossier-level primary source — a recognised reporting-code technical report or audited operator disclosure — with their field structuring, bankability panels and province scoring sequenced behind editorial capacity. What every record shares is the contract: each figure is Sourced (named, dated source), Derived (published method), or honestly Absent — no fourth state, no number invented to fill a row. The tier grew by reading more reports against that contract, not by lowering it. Where a secondary source is used to surface a figure, the standing obligation is to reconcile it to the issuer’s latest primary disclosure; the audit log records where that reconciliation has caught and corrected stale or mis-vintaged numbers.

The intelligence-grade tier is intended to grow. Each new IG dossier requires reading the underlying technical reports, applying the field-level provenance contract, registering the dossier with the audit-overdue badging mechanism, and triggering the next-audit-due date. Growth pace is bounded by editorial capacity and Quality Standard compliance — not by an arbitrary target. The deferred roadmap publishes the candidate-asset list and the conditions under which each moves from research-grade to intelligence-grade.

Growth continues to be sequenced with editorial capacity through the role architecture published on /about/, and its shape follows institutional engagement: assets requested by named institutional readers move into the tier first. The count is not a target to be maximised — it is the running total of assets that have cleared the primary-source bar, and it will only ever move at the pace that bar permits.