The Bay Area corporate catering market is not a food market. It is a reliability market that happens to serve food. When a Google Foods coordinator orders a hundred boxed lunches for a 12:15pm session at their Mountain View campus, what they are buying is the certainty that ninety-eight of those boxes arrive labeled correctly, at the right building, by 11:55am. Taste matters. On-time matters more.
I spent five years building an enterprise catering channel out of a $30M Michelin-recognized Bay Area restaurant group. At peak we ran standing catering with Stanford, Google, Apple, Meta, LinkedIn, Salesforce, Cisco, Adobe, and Nvidia, across a workforce of 215 that had to keep the a la carte side of the house running at the same time. This is the operating discipline that made that channel work, and the parts of the business I did not appreciate until I got them wrong.
What "enterprise catering" actually is
Enterprise catering is not restaurant food served in aluminum trays. It is a separate operating discipline that shares a kitchen with your restaurants and shares almost nothing else. The customer, the pricing model, the labor curve, the packaging, the delivery window, the failure mode, and the recovery playbook are all different. Run enterprise catering like a scaled a la carte kitchen and you will lose money on every large order and never be able to tell why.
The clearest way to see the difference is to look at what each side is optimizing for:
- A la carte restaurant optimizes for cover velocity and the guest's next visit. Failure mode is a slow ticket. Recovery is a comp.
- Enterprise catering optimizes for on-time drop and account retention. Failure mode is a late arrival at a booked meeting. Recovery is a refund plus a lost quarter of standing orders.
The math is different, the risk shape is different, and the operating rhythm is different. Everything below is written for the catering side of the house.
Fig. 1 · Enterprise catering and a la carte are different operating disciplines.
The nine campuses, and what each one taught me
Every tech campus has a slightly different personality, and the operating discipline that works at one does not automatically transfer to another. Here is the short version of what each one taught the team, roughly in the order we onboarded them.
Stanford
The Stanford business is a mix of standing faculty club catering, department events, and one-off donor dinners. The buyer mix is admin staff and event coordinators, and they care about consistency across a semester. Late once, you still get invited back. Late twice in a month and someone else gets the next request for proposal. Stanford taught us that account continuity outranks any single order.
Google Foods runs its own vendor rotation and its own scoring system. Every drop is graded on packaging, temperature at arrival, labeling accuracy, and driver conduct. Miss one and the score follows you into the next slot allocation. Google taught us that our catering ops manager needed to think in vendor scorecards, not tickets.
Apple
Apple's Cupertino catering runs on tight badging and tight time windows. Drivers get held at security if paperwork is wrong. Kitchens learn the hard way that a driver who is late to the badge desk is not late to the meeting, they are absent from it. Apple taught us to run dispatch backward from badge time, not meeting time.
Meta
Meta's Menlo Park campus is enormous and the internal map matters. A dropped order at the wrong building is functionally a missed order, because nobody has 20 minutes to walk it across campus. Meta taught us to build a per-building drop protocol with named contacts, and to confirm building numbers on every order rather than trusting the address on the purchase order.
LinkedIn's catering ops team is one of the sharpest I worked with. The standing orders are large and repeatable. The tension is that a repeatable order becomes invisible to the kitchen, and invisible orders drift on quality. LinkedIn taught us to never let a standing order run more than four weeks without a fresh kitchen taste-test on the exact menu the client is receiving.
Salesforce
Salesforce events run larger and more theatrical. Executive lunches, board dinners, offsite kickoffs. The margin is better and the presentation stakes are higher. Salesforce taught us that a catering ops manager needs to be part sales lead, part event captain, part logistics dispatcher, in the same phone call.
Cisco
Cisco's San Jose campus operates on long-standing recurring calendars, and their internal buyers value predictability the way an airline values it. Cisco taught us to build our production calendar off their recurring meeting calendar rather than off order acknowledgements, which cut our own lead-time stress by about half.
Adobe
Adobe cares about brand-consistent presentation. The catering has to look like the brand it came from, from the boxing to the labels to the tent card at the buffet. Adobe taught us that catering packaging is not overhead. It is the visible half of the product.
Nvidia
Nvidia was the fastest growing account we ever added. Order volume tripled inside twelve months as their campus workforce scaled. Nvidia taught us that a catering channel can absorb rapid demand growth only if you separate the catering kitchen from the a la carte kitchen physically, and only if the catering ops manager owns the growth curve on a weekly rolling forecast, not a monthly rearview.
Every account is a different scorecard. The operator's job is to know which scorecard you are being graded on this week, and to run the drop against that one.
The kitchen has to be split
This is the single lesson I wish I had learned earlier. Running catering out of the same physical line as your restaurant kitchen looks efficient on paper and is expensive in practice. Two things break at once.
First, the a la carte line loses focus when a 200-box order lands in the middle of a Friday lunch. The cooks are trying to fire tickets and pack boxes at the same time, and both suffer. Second, catering quality drifts because the production window is being interrupted by ticket firing. You cannot batch and portion accurately when the pass is calling you every ninety seconds.
The fix is a dedicated catering production area, ideally with dedicated catering hours that end before the a la carte lunch rush begins. In our largest unit we ran catering production from 4am to 10am on the same physical footprint, then broke down, sanitized, and handed the space back to a la carte prep by 10:30am. The cooks who came in at 11am for service walked into a clean line, not a catering aftermath.
If you cannot get physical separation, get temporal separation. If you cannot get temporal separation, cap the size of catering orders the a la carte kitchen can accept on a service day. Any operator who has tried to hold both without one of those three constraints will tell you the same story.
Logistics is the hidden margin lever
Nobody talks about catering logistics because it is the least fun part of the business. It is also where the margin lives. A well-run catering delivery operation runs at a 2 to 3 point higher contribution margin than a poorly run one, on the exact same food, at the exact same price, because the losses hide in vehicle utilization, driver overtime, and re-runs of missed drops.
Five things actually matter:
- Route the drops backward from the meeting time. Start with meeting start, subtract fifteen minutes for setup buffer, subtract badge and dock time, subtract drive time on that day of week at that hour. That is your dispatch time.
- Use dedicated catering vehicles, not the chef's SUV. Wrapped, refrigerated, and inventoried with setup gear at the start of every shift. A driver who has to load setup gear at dispatch is a driver who arrives late.
- Every drop has a named onsite contact and their phone number on the packing sheet. A driver arriving at a building without a phone number is a delivery that will be late. This one item cut our late-drop rate by roughly a third.
- Track on-time percentage at the driver level and the client level. Not just the aggregate. An aggregate 96 percent on-time can hide a client running at 87 percent and about to walk.
- Debrief every late drop the same week. Cause, driver, corrective action, and whether the client noticed. If the client noticed, the debrief includes a followup call from the catering ops manager before the client calls us.
Fig. 2 · Meeting time defines dispatch time, not the other way around.
The corporate ops contact is your real customer
The person who signs the invoice is not the person who decides whether you get the next order. In enterprise catering the real customer is the corporate ops contact who books the food, and their measure of you is whether they get in trouble internally when you deliver.
This changes the account management job in three ways.
First, your catering ops manager needs a real client CRM. Not a spreadsheet. A working system with contact history, order history, delivery notes, allergens, and a running list of preferences that survive the loss of any one team member. Ours held about eleven hundred active contacts across the nine campuses at peak, and it was the single most valuable operational asset we built on the catering side.
Second, followup is a first-class part of the job. Every order gets a followup within twenty-four hours. Not an automated email. A human call or a personal note from the account owner. This is what separates a catering vendor from a catering partner, and partners get the standing business.
Third, the corporate ops contact gets briefed before every menu change. If they hear about a new menu from their director instead of from you, you have already broken the relationship. Menu changes get a one-page briefing, a taste sample delivered to the office, and a phone call. Every time.
Pricing that reflects the real cost of a drop
Most restaurant operators price catering as a slight markup on the restaurant menu. That is a losing pricing model for enterprise catering, and it is why so many restaurant catering channels never actually make money.
The real cost of a catered drop is not the food. The food is roughly 32 to 36 percent of the price, same as the restaurant. What breaks the P&L is the invisible cost stack that a restaurant menu never carries:
- Dedicated catering production labor
- Packaging and disposables
- Vehicle, fuel, insurance, and driver labor
- Onsite setup labor when required
- Setup gear replacement (chafers, sternos, serving utensils)
- Account management labor that keeps the client on file
Add those in and a catered lunch drop needs to price at roughly 1.4 to 1.6 times the equivalent restaurant menu to hit the same contribution margin. Not because the food is more expensive. Because the delivery of the food is a real cost that has to be paid for on the same invoice.
When we repriced our catering menu properly, our contribution margin on catering moved from roughly break-even to a healthy positive line inside two quarters. We lost approximately zero standing accounts in the process. Nobody in enterprise catering leaves over a fifteen percent price shift when the reliability is real.
Quality standards that survive transit
Restaurant food is plated and served in the same building. Catering food is plated, packed, driven, held, and served in a different building thirty to ninety minutes later. Every item on the catering menu needs to be tested against that transit and holding window, and any item that does not survive it comes off the menu.
The test we ran on every new catering item:
- Prep, pack, and hold for the maximum expected time between production and service.
- Transport in a real catering vehicle, on a real route, at the actual time of day.
- Set up as the client would set up.
- Taste, plate, and photograph.
- Compare against the same item on a restaurant plate at the pass.
If the transit item does not land inside a defined tolerance of the restaurant version, the recipe changes or the item does not go on the catering menu. This process cost us a handful of favorite dishes and saved us dozens of client complaints.
Fig. 3 · The nine standing accounts, at peak.
The catering ops manager is the whole channel
Every restaurant group that ever built a working corporate catering channel had a specific person at the center of it. Not the chef. Not the general manager. Not the owner. A dedicated catering operations manager whose entire job was the channel. Hire this person before you take the tenth account, or the tenth account will break the channel.
The role is a hybrid that does not exist anywhere else in a restaurant group. Part account manager, part event captain, part logistics dispatcher, part quality auditor. They live in the CRM, they live in the dispatch board, and they are on the phone with corporate ops contacts for a large part of every day. In a healthy channel they are also the person the drivers know by name, the person the packing line trusts, and the person the chef listens to when a menu item is not surviving transit.
The single biggest hiring mistake I have seen operators make in this role is trying to promote a strong server or a strong sous chef into it. Neither is the right profile. The right profile is a hospitality-native project manager: someone who is comfortable with a spreadsheet, comfortable with a client phone call, comfortable with a delivery van running late in traffic, and comfortable making a decision without asking. They exist. Pay them properly and they will build the channel around them.
What I got wrong the first time
Three mistakes I would not repeat.
I let the head chef own catering menu design for too long. The head chef thought like a restaurant chef, and restaurant chefs do not naturally design for the transit window. We ended up rebuilding half the catering menu after two quarters of transit complaints. Now I put a catering-first sous chef on that job from day one.
I underinvested in the CRM. For the first eighteen months our catering ops manager ran the account book out of a shared spreadsheet. When she went on maternity leave, we lost about six weeks of institutional memory and two accounts. The lesson is that a CRM is not a nice-to-have. It is the account book that has to outlive any one person.
I priced off menu instead of off drop cost. For the first year we priced catering at a fifteen percent markup on the restaurant menu, and the numbers looked fine at the top of the P&L until you subtracted the driver labor, the vehicle costs, and the packaging. Once we repriced against real drop cost, contribution went positive and stayed there.
The point
Enterprise catering for tech campuses is a great channel for a Bay Area restaurant group, but only if you run it as its own operating discipline. Separate kitchen space, dedicated catering ops leadership, real dispatch logistics, a working CRM, honest pricing that reflects drop cost, and a menu tested for transit. Get those six things right and the channel becomes the highest-margin, highest-retention part of the group. Get any of them wrong and you are running an expensive way to lose money in an aluminum tray.
The single sentence I would give any restaurant operator building a corporate catering channel from scratch: the customer is not buying lunch. They are buying not-getting-in-trouble-at-work. Deliver that and the food will be graded gently. Fail on that and the food will not save you.