Work

Find A Farm

A two-sided marketplace I co-founded that let farms sell straight to the people around them, and let those people buy without posting in a Facebook group and hoping. The product worked, and the company still failed. This is the write-up of why.

Role
Co-founder
Concluded
January 2021
Footprint
United Kingdom and United States
Outcome
Wound down and debriefed
900%
increase in Google searches for "grocery apps" since 2019
Cited in business plan
3 in 5
Brits buying groceries online
Mintel
66%
consumers predicted to buy groceries via app by 2024
Business of Apps, projection
Supply side · web appFarmsInventory, point of sale, shipping options, collections scheduler.
The platformFind A FarmSubscriptions from farms plus a transaction fee.
Demand side · mobile appBuyersFarm locator, filtered search, favorites, online purchasing.
Free£0
  • 10 active items
  • 7.5% transaction fee
  • 14-day payout, £100 minimum
Standard£29 / month, or £290 a year
  • Unlimited items, featured badge
  • 7.5% transaction fee
  • 7-day payout, £50 minimum
Pro£59 / month, or £590 a year
  • Everything in Standard
  • 5.5% transaction fee
  • 7-day payout, no minimum
Two products, one platform. Pricing was weighted so transactions earned more than memberships, which is why the first hundred farms got a year of Standard free.

The problem we picked

People were tired of posting on Facebook and asking around town to find a farm they could trust. On the other side, the farm side had no solution at all. Our read was that this was one problem with two faces.

So we built two things. A mobile app for buyers, with a farm locator, filtered search, favorites, and online purchasing. And a web app for farms: product inventory, point of sale, shipping options, and a collections scheduler.

The business model

Revenue came from farm subscriptions plus a transaction fee. A free tier capped at ten active items with a 7.5% fee and a fourteen-day payout. Standard at £29 monthly or £290 yearly added unlimited items, a featured badge, and a seven-day payout. Pro at £59 monthly or £590 yearly cut the fee to 5.5% and removed the payout minimum entirely.

The pricing was deliberately weighted so that we earned more from transactions than from memberships. That is why the plan gave the first hundred farms a full year of Standard for free. We wanted volume on the platform more than we wanted subscription revenue on day one.

What worked

The product worked. That is the line from the debrief and it matters, because plenty of startups never get that far. The team had a strong skillset, everyone knew a little of everything, and being split across two countries gave us wider reach faster than a single-market team of the same size. The lack of funding forced decisions we would not have made with money in the bank.

What went wrong

On January 8, 2021 we sat down and wrote out fifteen mistakes. Grouped, they fall into four buckets.

Mistakes 03, 04, 12, 14, 15

Foundation and expectations

We did not build a solid foundation before starting. There was no leader running the company, the founders held different expectations of each other, the business plan was not realistic, and it contained no cost analysis at all.

Mistakes 01, 02, 08

Technology choices

The marketing site should have been on Squarespace or Wix rather than built by us. The app should have been React rather than the platform we chose. And a separate merchant app split our effort.

Mistakes 06, 07, 09

Onboarding farms

This is where the product lost. The farm-facing UX was hard to navigate, bank details were awkward for merchants to set up, and the onboarding flow was too simple to carry a non-technical user.

Mistakes 05, 10, 11, 13

Growth and operations

We messaged people from a Facebook profile. We bought an email list without checking lead quality first. E-commerce across two countries was harder than expected. And when trouble came, we reacted to the situation instead of pushing through it.

We didn't make a solid foundation before starting.Mistake twelve, wind-down debrief

What I took from it

The two failures that actually killed it were not technical. No leader and different expectations are organizational problems, and no amount of shipping fixes them. Everything in the technology bucket cost us time. Those two cost us the company.

The one that still shapes what I build is the onboarding bucket. We asked farmers who had never sold online to navigate a merchant app, set up bank details, and manage inventory, and we gave them a thin onboarding flow to do it with. That gap between what a tool assumes and what a non-technical user can do is the same gap Slingshot is built to close.

What it used