Project Writeup
7 min read
From Farm Data to Growth Strategy
How a crop-quality project for One Acre Farm expanded into a broader effort across operations, purchasing, pricing, retention, and customer growth.
One Acre Farm already had years of useful information. Planting spreadsheets tracked crops, varieties, beds, fields, dates, quantities, and notes. Sales records showed what customers actually bought.
The problem was not a lack of data. It was that the data rarely made its way back into decisions.
“I collected data for years but then there was no use for it, so I just stopped.”— Farmer Mike
That became the starting point for the project. I originally set out to make the farm’s planting and crop-quality data easier to use. Working directly with the farm, however, revealed a broader problem: information was fragmented not only between planting and sales, but across the customer experience as well.
The project gradually became an exercise in connecting both sides of the business: what the farm grows and how customers buy it.
Starting with the operating problem
One Acre Farm already relied on Google Sheets, so replacing its workflow with an entirely new system would have created more friction than it removed. Instead, I built around the process the farm already used.
The crop logger brings existing planting records into a mobile-friendly interface, lets the farm review crops by field and bed, and records quality using consistent inputs rather than disconnected notes. That created a more structured feedback loop between what was planted and what happened in the field.
The next step was connecting that operational data with sales. Planting records can tell the farm how much it grew. Crop-quality logs can show how the harvest performed. Sales data provides the missing signal: how much customers actually wanted.
Together, those datasets can help answer more useful questions:
- Which crops were repeatedly planted above demand?
- Where did strong crop quality fail to translate into sales?
- Which crops regularly sold through?
- Where could planting better reflect purchasing behavior?
- Which fields or varieties consistently performed better than others?
The point was not to create dashboards for their own sake. It was to turn information the farm already generated into decisions it could use next season.
Finding a second source of friction
Working on the data side of the farm also made me look more closely at how customers purchased produce. One Acre Farm primarily operates through a CSA model. The model is valuable because customers commit to a share, but that same structure can create friction for someone who is not ready to pay for a large portion of a season upfront.
Could the farm preserve recurring customer relationships while reducing the commitment required to start buying?
I explored a more flexible purchasing model in which customers could buy produce individually, choose quantities based on what they wanted, and make smaller recurring purchases instead of beginning with one large seasonal payment.
I prototyped the experience in Shopify to work through what the model would actually require: product listings, changing crop availability, individual pricing, subscriptions, fulfillment, and the customer checkout flow.
The prototype was deliberately an experiment rather than an immediate migration. The goal was to make the alternative concrete enough for the farm to evaluate before taking on the operational cost of changing platforms.
It also changed how I thought about pricing. A fixed CSA share packages availability, quantity, and commitment together. An individual-product model separates them, giving the farm more flexibility to adjust crop prices and availability while giving customers more control over what they purchase.
The important part was not Shopify itself. It was testing a different pricing and purchasing model against a real source of customer friction.
Choosing where to pursue growth
I also looked at how One Acre Farm could reach more customers. Instagram was one possible acquisition channel, but adding a channel also means producing content, maintaining it, and creating another workflow for a small team.
Rather than assuming that more social media was automatically the right answer, we focused on an audience the farm already had: existing customers.
Every CSA pickup already created a recurring physical touchpoint. The farm had a distribution channel sitting inside the bags it was already handing out, so I designed a small card to turn that moment into a customer-growth experiment.
The card directed customers toward actions useful to the farm:
- leave a Google review,
- join the newsletter,
- share the farm with someone else,
- and return for another purchase.
The ideas went through feedback with the farm and were revised before printing. We then paired the final cards with a $5 newsletter promotion, connecting a physical customer touchpoint to a measurable digital funnel:
CSA bag → card → newsletter sign-up → automated offer → redemption
Instead of treating the card as marketing collateral, the farm could look at backend behavior and ask whether it actually changed anything.
Building measurement into the experiment
The promotion is recent, so I do not yet have enough evidence to claim that it increased retention or revenue. That is also part of the project.
Rather than calling an experiment successful because it launched, I wanted the farm to be able to measure what happened next. The signals I would watch include:
- new newsletter subscribers,
- promotion redemptions,
- new Google reviews,
- subsequent purchases from participating customers,
- and changes in repeat-purchase behavior.
The next step is to use the backend data to see how effective the promotion actually was. That closes the same loop the project began with: collecting information is only valuable if it changes a future decision.
From a software problem to a business problem
The original request could have remained a crop-logging application. Instead, working closely with the farm showed how decisions about operations, demand, purchasing, and retention affected one another.
The technical pieces were useful because they made these questions easier to answer. They were not the end goal.
What I learned
What I found most valuable was moving between the field, the data, and the customer experience without assuming that the first proposed solution was necessarily the most important one.
Sometimes the answer was software. Sometimes it was a different purchasing model. Sometimes it was a printed card in a bag.
The common thread was identifying friction, making a hypothesis, building the smallest practical way to test it, and creating a path to learn from the result.