Pick up an Easter egg in a British supermarket this week and there's a reasonable chance the wrapper says "chocolate flavour" rather than chocolate. In the United States, the egg is still called chocolate — but it costs 14% more than it did a year ago and contains less cocoa than it once did. The product has changed on both sides of the Atlantic. The reasons are the same — and they run deeper than last year's cocoa price spike.

The Easter chocolate on shelves this week is the downstream consequence of decisions made months ago, under supply conditions that the available data did not fully illuminate. That gap — between what official statistics capture and what satellite intelligence now reveals — is what this piece is about.

A quiet reformulation

When cocoa prices hit nearly $13,000 per tonne in late 2024, manufacturers faced a straightforward choice: absorb losses, raise prices or reformulate. Most did some combination of all three. In the UK, McVitie's Club bars, Penguin, Toffee Crisp and Blue Riband each fell below the legal threshold for milk chocolate — not through any sleight of hand, but because the economics of cocoa butter had made the original recipe untenable. The word "chocolate" was replaced, on pack, with "chocolate flavour." These were rational responses to an irrational market. Across the Atlantic, Hershey, Mondelez and others navigated the same pressure through price increases, smaller pack formats and reduced cocoa content. US chocolate prices are up 14% in early 2026 compared with the same period in 2025. Volume is down. Consumers are paying more, buying less — and in many cases receiving less cocoa in what they do buy. None of this is the story of bad actors. It is the story of a supply chain making decisions under conditions it could not fully see.

Cocoa futures have retreated from their 2024 peak of nearly $13,000 per tonne to around $6,000. Analysts broadly forecast a market surplus for the coming season. The received view is that the worst is behind us. That view deserves scrutiny — not because the supply picture is necessarily worse than forecasts suggest, but because the data infrastructure underlying those forecasts has significant gaps that satellite intelligence now makes visible.

A base that has not been fully measured

Ghana and Côte d'Ivoire account for roughly two-thirds of global cocoa supply. The figures that anchor supply models, inform lending decisions and feed futures pricing in Ghana derive primarily from COCOBOD, the country's cocoa regulatory body, which registers approximately 1.9 million hectares of farmland. These are the numbers that move through the market — from West African first miles to commodity desks in London and Chicago.

Treefera's satellite analysis — using a 10-metre resolution deep learning model trained to detect active cocoa canopy — identifies 2.67 million hectares of producing cocoa in Ghana. That figure is consistent with independent peer-reviewed research that has found official estimates to structurally undercount the producing base. This is not a criticism of COCOBOD; ground-based registration at that scale is genuinely difficult. It is an observation about what the measurement infrastructure was built to capture and what it was not. When stress assessments and disease risk models are applied to registered area rather than actual producing area, the error compounds through every downstream calculation — whether that calculation is made in Accra, London or Chicago.

What the regional picture shows

Headline production figures for 2024 showed a meaningful rebound from the prior El Niño-pressured season. Belt-wide, actively producing cocoa area expanded approximately 6% — Bono up 10.6%, Ahafo 18.8%, Ashanti 4.4%. Those numbers were genuine. Satellite monitoring confirmed the recovery.

It also showed something the aggregate figures did not resolve. Western North contracted by 21% over the same period, losing approximately 64,000 hectares while conditions improved across every neighbouring region. Western North carries the highest documented burden of Cacao Swollen Shoot Virus Disease in Ghana. A government survey found 69% of surveyed farms there diseased and unproductive. The most severe CSSVD strains reduce yields by 50 to 70% and kill trees within two to three years. Improved rainfall does not reach them. A regional divergence of that kind — recovery across most of the belt, structural contraction in the most disease-affected area — is precisely the signal that matters for supply planning and capital allocation. It is not visible in national aggregates.

By January 2025, 90 days into Ghana's Harmattan season, Treefera's area-weighted weather stress score stood at 0.78 on a zero-to-one scale where 1.0 represents maximum historical stress. The prior season, which produced roughly 530,000 metric tonnes against a COCOBOD target of 800,000, closed at 0.61. Approximately 40% of Ghana's national tree stock is now over-aged or CSSVD-affected — a portion of the base that does not respond to better growing conditions regardless of how favourable the season becomes elsewhere in the belt. Analyst surplus forecasts for 2025/26 range from 79,000 to 287,000 metric tonnes. That spread — nearly fourfold on a single season — reflects genuine uncertainty about a physical base that national-level figures do not resolve.

What this connects to

The reformulation decisions now visible on Easter shelves from London to Los Angeles were capital decisions before they were procurement ones. The compound chocolate market — products using fat substitutes rather than cocoa butter — is projected to grow from $4.36 billion to $7.16 billion by 2032. Once a manufacturer has validated consumer acceptance of a reformulated product, the decision calculus for reverting is steep. Barry Callebaut, one of the world's largest chocolate processors, reported sales volumes down 22% in the three months to November 2025. The structural adjustment is already priced into investment plans on both sides of the Atlantic.

The wrapper change on a UK Easter egg and the price increase on an American one are not separate stories. They are the same story at different stages. Both trace back to decisions made under conditions that official data could not fully see. The gap between what national statistics capture and what high-resolution satellite intelligence reveals is not a technical detail. It is where supply chain risk actually lives — and where better visibility changes what is possible, for procurement teams, for the capital markets that underwrite them and for the farming communities in West Africa whose livelihoods sit at the origin of all of it.

All Treefera figures are derived from a 10-metre deep learning model validated against independent CERSGIS ground truth data across multiple Ghanaian districts. Model F1 score: 0.932. Weather Stress Score is area-weighted by confirmed producing cocoa at district level, on a 0–1 scale where 1.0 represents maximum historical stress. Not financial advice.