Farmcrowdy Foods / MeatHub — Fresh Food Commerce During COVID

Context

In April 2020, Lagos was in lockdown. Wet markets — Mile-12, Oko-Oba — were the primary source of fresh meat for most urban households. Those markets were crowded, informal, and now either closed or too dangerous to visit. People who needed to feed their families had no trusted alternative.

Farmcrowdy had a beef processing facility capable of handling 100 bulls per day and a logistics network already running. The question was whether trust — the kind required to pay for fresh meat you can't see, from a company you may not know, delivered by someone you've never met — could be designed fast enough to matter. We had about four weeks.

The three-channel problem

The assumption at the start of the project was that we were designing a mobile app. We were not. The people who needed fresh meat in Lagos in April 2020 included middle-class households with smartphones and reliable data — and they included domestic staff, market traders, and households where the person doing the food shopping was not the same person who owned the smartphone. A single-channel app would reach one part of that market.

The product launched across three surfaces simultaneously: the Farmcrowdy Foods app for Android and iOS, a WhatsApp ordering flow for users without app access, and a phone hotline for voice orders. Each surface served a different user profile. Each one needed the same core trust signals translated into the right format for that surface.

In the app, trust came through hygiene photography, sourcing information, and an order tracker. On WhatsApp, it came through response speed, structured message templates, and a confirmation number the user could refer back to. On the phone, it came through the call handler's ability to give a price, confirm a delivery window, and repeat the order back correctly. Designing for WhatsApp and phone meant designing for environments where I had no control over the visual layer. The trust had to live in the copy, the timing, and the process.

What the research found

The research question was: what makes someone trust a fresh food delivery service they have never used, for a product category (raw meat) where trust failures have physical consequences? The findings were not what I expected. Hygiene certification and sourcing provenance — the things food delivery platforms usually lead with — were secondary concerns. The primary concern was: what happens if something goes wrong?

Users wanted a clear answer to "can I return this?" before they wanted to know where the beef came from. The return and dispute path was the biggest trust signal, not the origin story. That changed the onboarding priority: return policy and contact information moved to the product detail page, not the checkout. The WhatsApp flow began with a message that named the return window before it named the price.

How those findings changed the interface

The product detail page led with the fulfilment promise — delivery window and return policy — before the product description. This felt counterintuitive; food platforms lead with photography and descriptions. But for a first-time buyer, the fulfilment promise reduces risk. Photography closes the sale after the risk has been removed.

The order confirmation on WhatsApp arrived within 90 seconds of a completed order. Users in test sessions did not trust confirmations that took more than two minutes — the speed of the confirmation was read as a signal of operational competence.

The Meathub physical shops were partly a logistics decision and partly a design decision. For customers within walking distance of a shop, collection was offered as an alternative to delivery — serving the household member who wanted to see the product before accepting it. A visible, clean physical location made the digital product more credible: the shop was a trust anchor for the app.

Outcomes

143 orders in the first month of operations. 3,567 unique orders by September 2020. 831 unique customers. $33,000 in transaction revenue from March to September 2020, entirely through organic growth. (Source: Farmcrowdy September 2020 Company Profile.)

What I'd do differently

We designed the three channels in sequence — app first, then WhatsApp, then phone — under the four-week timeline. The phone hotline script ended up reverse-engineered from the app flow rather than designed from how a voice conversation actually works. I would have designed the phone channel first, since it has the least room to hide a bad decision, and adapted the app and WhatsApp flows from that constraint instead of the other way around.