Through 2018-19, with Jason Thomas, I cofounded Sow – a marketplace for groceries from local farms. Over 9 months we designed a brand, tested several prototypes, and raised funding. Eventually shutting down after failing to find a sustainable business model.
Hey! I’ve got a project I’m submitting to a pitch competition. If it wins I’ll be turning it into a company. Think you could help with a logo?
The project was Sow. An online marketplace for locally grown food. It needed a visual identity to help tell the story of the company, and to look like enough of a thing for people to take it seriously.
In short, the pitch won! That September I wrapped up projects in Nepal and relocated to Columbia, South Carolina to take the prize money and work on it full-time.
With the overarching goal of making it easier to buy local food, we set out on a mission to learn all we could about local farming. Over the next month we:
As we dug, themes started emerging. Collecting and affinity mapping our findings helped lead us to concrete insights and potential problems to solve.
One thing we heard time again was that farmers struggle to track and communicate their inventory, both internally and to customers. As a result, food was going to waste and farms were missing out on business.
We built a simple app for City Roots, a local farm suffering from this. Harvesters would log what they harvested; packers would log what product was packaged and shipped. In-between, front office would have live data on inventory levels, and could have the confidence to sell higher quantities as orders came in.
The app started as a super low-fidelity prototype: pen and paper system workers used to tally product in and out. This let us test the mechanics of the process cheaply, and we jumped into code only once the kinks were ironed out on paper.
The app was quick to build. We had City Roots set up with it in a week or so. There were teething problems galore to sort out, but they quickly started getting value from it.
Previously, if a chef or wholesaler called with an order, front office would have to call them back after physically going to the store room and counting how much stock they had. Now, they could see their availability onscreen, as they took the call, and could close the sale there and then.
Though the inventory tracker had proved useful in the short tests we’d run, we felt we’d trayed from the original goal of making it easier to buy local food. We continued learning and brainstorming other routes we could take.
Spending so much time in and around one farm highlighted another interesting nugget. Local farmers typically delivered to restaurants themselves – driving into town once or twice a week – requiring a $100-$150 minimum order to make a drop. We saw an opportunity here. Could we aggregate consumer orders to meet this minimum, and have farmers deliver to the consumer, no middleman required?
This led to the idea of a farmers market in a box, a self-service farmstand stocked with locally sourced food, delivered by farmers, for the neighborhood.
Over the next two weeks, and many trips to Home Depot, we built our first prototype. It was basic: a fridge and freezer from Craigslist, a tablet and card reader to accept payments, and a wooden frame to hold it all together. A kind friend let us host it on her porch.
We filled the box with meat, eggs, and butter from 5 local farmers, made an announcement on the neighborhood mailing list, and went knocking on doors to announce the box’s opening. We opened for business on October 30th, and had a steady stream of curious neighbors trying it out over the following days.
What followed was a rapid education in the challenges of running a small business. Supplies ran out, chickens stopped laying, things needed fixing, the fridge frosted up, and customers needed updating. We quickly had to figure out systems for checking stock levels and ordering ahead of time. We built in redundancy for popular products such as eggs. We set up a mailing list to update neighbors with stock levels and new arrivals.
Weeks passed as we worked to iron out the kinks and gather feedback from our first customers. We learned a lot from these initial users, and quickly accumulated a long backlog of improvements to make to the product.
Customers’ biggest complaint was that they couldn’t know with any certainty what would be in stock in the box, so they couldn’t rely on it for their normal shopping. In an effort to address this, we built a Progressive Web App listing stock levels for everything in the box. Pulling live stock data from the the box and presenting it through a mobile-friendly UI.
When we launched it for the neighborhood, we saw a good response in revenues. Regulars appreciated being able to check if their favorites were in stock. People new to the concept had something they could check out from their phone, without the commitment of walking the few blocks to see the box themselves.
The box worked great as a first experiment, but it was far from financially viable for us or the farmers supplying it. This micro-retail model would’ve needed 10x the weekly sales volume to be viable, which didn’t seem likely without a major rethink. The outlook looked much rosier though if we could take regular subscription payments from a small group of committed neighbors. 30 neighbors with a weekly recurring grocery order would make an enticing proposition for a farmer, and it would be well worth their while to deliver.
With this we set about sketching ideas for how we could take and deliver subscription orders for a neighborhood. We sketched concepts for bigger and better communal refrigerators, to live in parks and other public spaces. We began talks with city officials, getting approval to build a prototype in the neighborhood park.
The larger prototype would’ve required more cash, and so we turned to look for investment. We put together a pitch deck and concepts for future versions of our app, where neighbors could browse a catalog of products and order from their local farmers. We’ve since been pitching to grants and accelerators, raising the funds to build this prototype.