Every full-service operator eventually gets asked the same question. A regular customer plans an offsite. A local company needs lunch for 40. The kitchen can do it, the food is right, so someone in the group says yes. The order goes out, the customer is happy, and the next week two more orders come in. Six months later the restaurant is running a catering business by accident, out of the same kitchen, with the same team, on the same P&L. And nobody can tell whether it is making money.
I spent five years scaling a $30M Michelin-recognized Bay Area group from three to five locations. The catering channel we built inside that group grew into an enterprise line serving Stanford, Google, Apple, Meta, LinkedIn, Salesforce, Cisco, Adobe, and Nvidia. It became a real business unit. But it only became one when we stopped treating it as an extension of the restaurant and started treating it as its own operation. This is what that took.
Catering is a different business, not a different service
The fastest way to hurt both your restaurant and your catering channel is to run them as one thing. The two look similar from the outside. Same kitchen, same recipes, same brand. But operationally almost every variable is inverted:
- Demand is contracted, not walked in. A la carte demand is probabilistic. Catering demand is a purchase order. That changes forecasting, scheduling, and inventory.
- Batch size flips. A la carte makes one plate at a time. Catering makes 40 to 400. The labor curve per unit is completely different.
- Delivery is a cost line, not a revenue line. Restaurants sell dine-in. Catering sells delivered. Vehicles, drivers, and packaging become real P&L items.
- Failure mode is public and permanent. A missed reservation costs one table. A missed catering drop costs one corporate account, and corporate accounts talk to each other.
Once you accept that catering is a separate business, the four prerequisites below are obvious. Skip any of them and the two businesses fight each other inside the same building.
Fig. 1 · The four preconditions.
1. A separate P&L, from day one
The most common mistake is to run catering revenue into the same P&L as restaurant revenue and split nothing else out. When that happens, three things get quietly hidden.
First, catering labor gets buried inside restaurant labor. A prep cook spends four hours on a catering order in the morning and six hours on lunch service, and the payroll system does not split those hours. Catering labor as a percent of catering revenue is invisible.
Second, catering food cost blends with restaurant food cost. If catering uses a different protein blend or a different portion spec, you cannot see it. The blended food cost looks fine and the catering food cost is running 8 points high.
Third, and worst, delivery and packaging cost get filed under "operating supplies" or "other" and never get attributed to the orders that generated them. A catering drop that looks profitable at $2,400 of revenue is actually breaking even after $180 of packaging, $75 of driver time, and $40 of vehicle wear that never made it to the cost side.
The fix is simple and unavoidable: from the first day, catering has its own P&L. Its own revenue line, its own COGS, its own labor, its own delivery cost, its own packaging cost, its own gross margin. You do not need a separate legal entity. You just need a chart of accounts that separates the two channels cleanly, and a bookkeeper who does the split correctly every week.
2. A separate production window
Catering cannot be produced during a la carte service. Every operator learns this the hard way, usually after their expo blows up on a Friday because the prep team was assembling a 200-person drop while tickets were firing.
The fix is a dedicated production window, usually 5am to 10am, with its own cooks, its own equipment, and its own workflow. Catering gets produced, packed, and staged before the doors open for lunch. Then the a la carte team walks into a clean kitchen with all the catering already out the door or on the delivery pad ready to load.
The two channels stop competing for the same 12 square feet of prep table. Both quality and both speeds go up.
Shared production windows produce misses in both channels. The catering plates lose freshness and the a la carte plates lose tempo. Split the window and both improve immediately.
3. Dedicated packaging and vehicles
Catering packaging is not restaurant takeout packaging in a bigger box. It is a different category of product. It has to hold heat for 45 minutes to two hours in a car. It has to stack without crushing. It has to look good on a corporate conference room table with 40 people looking at it. And it has to have a hot-cold barrier that prevents the salad from wilting under the pasta.
Undercapitalizing on packaging is the single most common way catering programs bleed margin invisibly. Cheap packaging degrades product quality on arrival, which produces the one thing catering cannot recover from: a corporate contact posting in a Slack channel that the food showed up cold.
Vehicles matter for the same reason. A catering order that has to go across the Bay Area in a chef's personal Corolla is not a scalable operation. Real catering channels need at least one dedicated vehicle by the time revenue passes about $500K a year, and typically two vehicles once revenue is over $1M. The math is not exotic. It is depreciation, fuel, insurance, and a driver, all loaded into the delivery line of the catering P&L.
4. A real sales function, with a CRM
Catering orders do not walk in the door. They come from a specific human at a specific company who trusts a specific person at your restaurant. If nobody on your team owns that relationship, you will have a catering channel that maxes out at whatever inbound demand happens to find you, which is usually a small fraction of what the market could support.
The sales function is a person, not a job description bolted onto a general manager or a hostess. Their week looks like this: 60 percent outbound, 30 percent servicing existing accounts, 10 percent post-event followup and case study collection. They own a CRM with every corporate contact, every recurring order, every dietary preference, every past complaint, every renewal date.
At Zareen's the enterprise catering channel served nine of the largest tech companies in the region. Every one of those accounts started with the same pattern: a warm introduction, three sample deliveries at cost, a small first order, and then a monthly cadence for years. None of that happens without a person whose calendar is built around it.
The unit economics that make it work
When the four prerequisites are in place, catering economics are attractive. Food cost typically runs 24 to 28 percent of catering revenue, versus 30 to 34 percent in a la carte, because batch production is more efficient with trim and yield. Labor as a percent of catering revenue can run under 15 percent because the labor curve is nearly flat above 40 covers. But packaging runs 4 to 6 percent of catering revenue, and delivery runs another 6 to 9 percent, both of which do not exist in a la carte.
Net contribution margin on a well-run catering channel lands in the 22 to 30 percent range, roughly 4 to 8 points higher than a comparable a la carte contribution. That gap is what makes the whole thing worth building.
When not to launch a catering channel
Three situations where catering is the wrong move, at least for now:
- Your a la carte operation is not stable. If food cost, labor cost, or service quality are still drifting in the restaurant, adding catering will make all three worse. Fix the restaurant first.
- You do not have kitchen capacity in the morning. If your prep team is already running from 6am to close, there is no production window to carve out. Fix the prep schedule first.
- You do not have the capital for packaging and vehicles. A catering channel launched on borrowed retail packaging and a manager's personal car is a P&L accident waiting to happen. Wait until you can capitalize it properly.
How the enterprise accounts actually got won
People assume the tech-campus catering accounts came from a pitch deck and a sales cycle. They did not. Every one of the enterprise accounts we won at Zareen's followed the same four-step pattern, and the pattern is worth naming because it is repeatable in almost any market with a corporate base.
Step one: a warm introduction from an existing guest. A regular dine-in customer who worked at Meta mentioned that their team lead was tired of the incumbent caterer. That was the entire pitch. The name was on the reservation list, the food was already something they trusted, and the introduction was over dinner, not over LinkedIn.
Step two: three sample deliveries at cost. Not free. At cost. Free samples signal desperation. Cost-priced samples signal seriousness. The receiving contact could taste the food, see the packaging, watch the delivery timing, and evaluate us as a real operational partner before any money changed hands.
Step three: a small first order for a specific team meeting. Twenty covers. Perfectly executed. On time, hot, correctly labeled, with a clear post-delivery followup email the same afternoon. This first order is the audition. Get it wrong and the door closes for a year.
Step four: a monthly cadence for years. Once the account trusts you, the recurring pattern is remarkably stable. A well-run corporate account at a Bay Area tech campus turns into $80K to $250K of annual revenue for the next three to five years, with attractive margins and predictable production schedules.
None of this happens without a person who owns the pipeline, works it every day, and closes on the operational execution rather than on the pitch. That is why the sales function is prerequisite four, and why hiring the wrong person for the seat is the single biggest cause of stalled catering channels I have seen.
What I got wrong the first time
I launched catering at one of our units too early. The a la carte P&L was still leaking and I convinced myself catering revenue would help pay for the fix. It did not. It made the fix harder because the same team was now stretched across two operations, and the shared P&L hid which one was actually dragging on the other.
The other mistake: I hired a catering operations manager who came from the floor instead of from account management. She was great at hospitality and at the kitchen. She was uncomfortable making outbound calls, and the sales pipeline stayed thin for a year. The next hire came from a corporate account management background and the pipeline tripled inside two quarters. Lesson: the catering operations manager is a sales role first and an operator second.
The point
A catering channel inside a restaurant is a real second business, not a side hustle for the same team. Give it its own P&L, its own production window, its own capital, and its own sales function. Do that and it becomes the highest-contribution line in the group. Skip it and you get an accidental catering business that grows into a margin problem you cannot see.