Genesis

A soil data platform for the bio-sourced industries. It measures the health of farmland and the impact of farming practices, from a single plot up to an entire group. This case study follows its data explorer.

The first challenge was to give clear answers at every level of aggregation. At the finest, a single plot. At the widest, a whole group with all its brands, supply chains, production methods and subsidiaries. In between, products, farms, and any set of plots worth looking at together: a group of farms, plots sharing a practice or a context, plots growing the same crop. There is even a score for wheat, worldwide.

The second was the score itself. Each plot gets a Genesis score, built from many indicators, themselves computed from several raw measurements. A farm’s score is the average of its plots weighted by their hectares, and scores keep rolling up from there to a brand, a supply chain or a group. The rating had to work at every one of these scales and stay easy to read at all times.

Nutri-Score and energy performance certificates were the obvious references, but you can’t average a letter grade. Star ratings, on the other hand, hold up: on Airbnb or Google, the mean of thousands of reviews still reads at a glance. So the final rating takes something from each. It is a number out of 10, which averages cleanly, shown on a half-circle gauge with one of five labels, from Critical to Optimal, in the colour codes everyone knows from food and energy labels. From 6 up, a score counts as good. A plot gets a whole number and a gauge cut into ten segments. Anything aggregated gets one decimal and a continuous arc, so the shape alone tells you whether you’re looking at a single field or at a whole supply chain.

The hardest call came later. The explorer had grown into the biggest feature of the product, with dozens of filters and ten types of charts. Key account managers used it heavily, but the end users it was meant for found it intimidating and barely touched it. So we took it out. It moved to an admin space for the Genesis team only, and the production app lost its data manipulation tools. It came back with a much simpler navigation, built around the supply chains that mattered most to each client, and the filters then returned little by little, in subtler ways.