What I Learned Testing, Starting and Launching A Food Delivery and Meal Prep Business

From key assumptions to custom tech, here’s what it took to build Southfork.

Andy Blechman 20 Oct 2020 • 8 Min Read


I first got into the food delivery and meal prep game in 2014, while I was in business school at Columbia. Most days I would listen to podcasts on the ride to school, and the vegan athlete Rich Roll’s was a favorite. Eventually, I was convinced that a plant-based diet was worth trying.

At the time there weren’t a ton of local restaurants in Harlem that I could order delivery from. In order to have a good meal on campus, I had to bring my own food or I’d be stuck eating terrible cafeteria food. But preparing meals each day from home took time that I didn’t always have and I figured I wasn’t the only one who was itching for healthier options. So I started to take orders from my classmates for fast-casual restaurants downtown—Sweetgreen, Little Beet Table, Dig—and get great food delivered right into the lobby of the school.

I featured one vendor per day, and gathered orders through Google forms the day before. After some initial testing, I found that taking pre-orders and issuing cut-off times was a much easier system for me to operate than an on-demand model. I was getting upwards of 60 orders every day and collecting payments via PayPal. The order volume was big enough that it counted as catering for the restaurants, which meant that the vendors handled delivery. I thought I had something.

After I graduated, my wife and I decided to leave New York. We had been in the city for nine years, felt it was time to try living somewhere new, and ultimately settled on Atlanta where she’s from. At the time, food delivery wasn’t as readily available like it was in NYC, so it seemed like there was an opportunity to offer high-quality, healthy lunch food to working professionals. I was inspired by some of the meal service companies that were popping up in Manhattan at the time such as the now-defunct Maple. Service providers like DoorDash and Uber Eats had yet to arrive in Atlanta, so the market didn’t have strong delivery options, or saturation.

Step 1: The Testing Phase

Thanks to my wife’s brilliant marketing mind, we came up with the name Southfork. I started the business with four key assumptions that I could test.

  1. I could sell to tenants and employees on an office building to office building basis to create awareness.
  2. I would be able to generate a significant number of orders from each building.
  3. We would deliver once per day to each building, right at lunch time. (By batching orders, my unit economics would be really high: one delivery person cost $18 per delivery and my profit on each meal was $4–5, so I only needed 4–5 meals to break even.)
  4. I would hire a chef and we would make our own food from a commissary kitchen.

Then, I got to work. Here’s what happened next:

  1. After a long and grueling search for a like-minded chef and an affordable kitchen space, I realized that the path of least resistance was to partner with restaurants to provide food, so I could prove that the idea worked. I convinced three restaurants to work with me to supply a mix of their best-sellers and healthier iterations of pre-existing dishes.
  2. I started by delivering meals to the office building where we worked.
  3. I persuaded my co-working space, Industrious, to let Southfork be the official lunch provider of our Midtown Atlanta location and promote Southfork with a daily email to all members.
  4. I built a basic website on Squarespace to take orders and start testing.

The tests worked. Each time I ran a test, I got about 20 orders, which allowed me to make a profit.

Our logo

Step 2: Building The Tech

After a thorough testing period, I decided to invest in technology. My test runs revealed that there was a market for my idea, and also that I needed a better system. While Squarespace was a great starting point, it lacked the core functionality that I required. For example, manually pulling menu items off the website every morning at 10:30 AM and compiling data was both inefficient and unsustainable.

I was hoping to use an off-the-shelf service like Shopify, but most e‑commerce technology companies aren’t built to support pre-orders and food delivery, so I decided to build a custom site. (Side note: this is one of the core reasons we built Bottle!) That’s when I met Will, my co-founder in Bottle. A friend from school connected me to the agency he was freelancing for, and he was brought on to help build out the backend. We built a killer website, which got a ton of positive feedback.

Step 3: Acquiring Users

While the tech was being built, I set out to acquire customers pre-launch. Here’s how I did it:

  1. I hired an outsourcer from Upwork to put together a list of companies and their key contacts at 30 target office buildings.
  2. I used YesWare to send automated emails that felt personal to each of these contacts.
  3. Over the course of three months, I met with dozens of company admins and property managers to pitch Southfork as their lunch provider.

Step 4: Hiring Drivers

Once I had five buildings signed on, the next step was to hire drivers. Here’s how that went:

  1. I was able to find drivers through Craigslist for two-hour time slots with guaranteed hours throughout the week. (One of the advantages of Southfork’s model was the predictability.)
  2. I assembled a team of four drivers using an automated vetting process: candidates would submit a Google form I linked to on Craigslist and I reviewed their answers then set up an interview if I felt good about them, and finally, submitted their information for a background check. We paid each driver an hourly wage for 1–2 hours per day. (At this point in Southfork’s history, I had hired someone to help me with operations.)
  3. Each Sunday, we coordinated the delivery schedule based on driver availability.
Southfork’s pita sandwich

Step 5: Figuring Out Food Production

The goal was always to make our own food. Before I moved to Atlanta, I had hired a chef to put together 20 recipes. But as we got deeper in the weeds, I was having trouble securing the right talent, getting introduced to local suppliers (I was still new to the city), and finding a budget-friendly kitchen space.

So we doubled down on partnering with restaurants. I pitched every top restaurant in Atlanta and got buy-in from five well-known brands, including Rumi’s Kitchen and The General Muir. The reason they worked with us was because we brought them demand and order volume before lunch, when their staff was on-site but not busy yet.

Step 6: Launching

In order to launch, we needed to build awareness. Back in 2015, social media wasn’t as essential as it is today, but if I were launching now, I would have started by building an Instagram following pre-launch. Here’s what I did do:

  1. I worked with a local PR firm, which helped me land a great article in Atlanta Magazine. (Even if you don’t have budget for a PR firm, I suggest reaching out to small, local papers to tell them what you’re up to, as press coverage goes a long way.)
  2. I entered a pitch competition called Startup Riot and won 15 meetings with investors and experienced entrepreneurs, which led me to many great conversations with smart, well-connected people. This also led to a decent amount of buzz in the community.
  3. I sent out an official launch email to the list of buildings and employers that I had previously reached out to.

These were the methods and tools that allowed me to test, start, and successfully launch a food delivery and meal prep business. While Southfork never got to the size I hoped it would, the silver lining was that I met Will, found my way to Bottle, and we have built a great company today that supports so many local food brands.

This is the first article in a series about building a meal delivery business from the ground up. Stay tuned for further blog posts on finding a food production facility, creating recipes and hiring culinary talent, building a budget and cash flow model, and growing an audience on Instagram and Facebook.