Your restaurant customer database Zayos, what to keep and how to keep it
Published July 2026 · Updated July 26, 2026
A restaurant customer database is one list of every diner who has ordered from you, with a way to reach them and a record of what they bought. If you cannot pull the last 90 days of orders by name in under a minute, you do not have one.
Free, 60 seconds, no card.
Can you answer these three questions before your lunch rush?
Skip the feature lists for a minute. Sit at your own desk and try to answer these. Whatever tool you already pay for, the answers tell you exactly what you have.
Who ordered from you in the last 90 days?
Not how many orders. How many people, and what are their names. If the answer lives in a stack of tickets, a phone that texts back, or a marketplace dashboard you cannot export, the answer is that you do not know.
Which of them has not come back in 60 days?
This is the single most valuable question a restaurant can ask its own data, because a lapsed regular is the cheapest sale in the business. Answering it requires order dates tied to a person, which is exactly what most restaurants never keep.
Can you email or text those people this afternoon?
Having the record and having permission to use it are two different things. A list you cannot legally or practically contact is an archive, not an asset.
Most independents fail all three, and it is not a discipline problem. It is that the systems in a restaurant are built to close tickets, not to keep people. Nothing in the building has the job of remembering a diner.
What actually belongs in a restaurant customer database?
Ten fields. That is the whole thing. Every one of them earns its place by making a specific decision possible, and a database missing the last two rows will get you in trouble or make you look like a stranger to your own regulars.
| Field | What it lets you do | Source |
|---|---|---|
| Name | Turns a transaction into a person your staff can greet and your email can address. | Checkout |
| The channel you can use most often at the lowest cost, and the one that survives a phone number change. | Checkout | |
| Phone | The channel with the fastest response, and the one your kitchen uses when an order goes sideways. | Checkout |
| First order date | Tells you whether someone is new or is a regular you have been undercounting. | Order record |
| Last order date | The field every useful segment is built on. Without it you cannot define lapsed. | Order record |
| Order count | Separates a one-time delivery from a Tuesday habit worth protecting. | Order record |
| Lifetime spend | Ranks the book. The top decile is who you protect when something goes wrong. | Order record |
| What they order | The line items make the difference between a generic promo and a message that reads like you know them. | Order record |
| Channel and location | Web, QR, counter, which store. Tells you where the relationship was formed. | Order record |
| Contact permission | Whether they agreed to hear from you, and when. This is the field that keeps you out of trouble. | Checkout consent |
Notice what is not on the list. Birthdays, dietary notes and favorite table are pleasant, and a fine-dining room should keep them, but no independent has ever lost money for not tracking them. Last order date has decided more marketing outcomes than every soft field combined.
Where do customer records leak out of a restaurant?
A restaurant serving 3,000 orders a month can end the year with a few hundred usable records. The gap is not one big hole, it is six small ones running all day.
The phone order nobody wrote down
A regular calls, orders the same thing they always order, pays at pickup, and leaves no trace. The oldest and largest leak in the business. Most restaurants lose more customer records to the phone than to any app.
Marketplace tickets
DoorDash, Uber Eats and Grubhub hand the kitchen what it needs to cook and hand off the order. You typically get a first name and a delivery note. No email, no reachable phone, no permission. Those diners are not in your database and never will be.
Guest checkout with no record kept
Some ordering systems collect an email to send the receipt and then treat it as a transaction detail rather than a customer record. The data passed through the building without stopping.
Three systems, three lists
Reservations in one tool, online orders in another, catering in a spreadsheet. The same person exists three times and is a regular in none of them, so nothing about them looks important.
The POS list that is really a receipt log
A point of sale can tell you a card ending in 4412 bought lunch. That is a payment record. It becomes a customer record only when it carries a name, a contact method and permission, and most of the time it does not.
The staff phone
Catering leads and large orders often live in one manager’s text thread. When that manager leaves, so does the book.
The pattern behind all six: a record only survives if capturing it is part of an action someone was already taking. Any process that depends on a busy person doing an extra step during a rush will not run on a Friday night, and Friday night is where your customers are.
How do you prove you own the database?
Run the export test. Log in to whatever system holds your customers and try to download the list right now. Not a report, not a chart. A file with names, email addresses, phone numbers and last order dates, on your computer, in under five minutes. If you cannot, the honest description of your situation is that a vendor holds your customer relationships and lets you look at them.
Then ask three contract questions, in writing, before you sign anything:
- Can I export the full customer list, with order history, at any time and without asking support?
- If I cancel, do I keep the data, and for how long can I still get it?
- Does the vendor, or anyone they work with, market to my diners using this data?
That third question is the one people forget to ask. A customer list you share with a platform that also serves your competitors is not the same asset as a list only you can see.
Which segments are worth building first?
A database earns nothing until it is cut into groups you can speak to differently. Four are enough for a single location, and each one is a rule about dates and counts, so you can build them the day the data starts landing.
New, one order
First order in the last 30 days, order count of 1.
The highest-return message a restaurant can send is the one that turns a first order into a second. This is a small group with a short window, and the offer barely matters. Being remembered does most of the work.
Lapsed regular
Three or more orders lifetime, nothing in the last 60 days.
These people already liked you enough to come back twice. Something changed: a move, a bad night, a competitor. A single specific note beats a blast, and the response rate on this group is usually the best in the book.
Top spenders
Top 10% by lifetime spend.
Worth knowing by name. This is who gets the first catering slot, the call when the special they love is back, and the personal apology when an order goes wrong.
Catering and large orders
Any order above your normal ticket ceiling, or any order flagged catering.
Office buyers repeat on a calendar, not on a craving. One record with the buyer, the office, and the date is worth more than a hundred single-entree records.
How does the customer book work on Zayos?
Zayos is commission-free direct ordering for independent restaurants, and the customer database is not a module you add. It is the by-product of taking orders on a channel you own. Every direct order, whether it came from the branded site, a QR code at the table, or a reorder link, writes a record with the ten fields above, including the consent the diner gave and when they gave it.
Records match on phone and email so the same person does not appear three times, and order history hangs off the person rather than the ticket. Last order date is maintained automatically, which means the lapsed segment is always current instead of being a spreadsheet somebody rebuilt in March. You can export the book whenever you want, and it is yours whether you stay or leave.
The honest limit is the one every platform shares. Orders that arrive from DoorDash, Uber Eats or Grubhub can be ingested into the same kitchen tablet through our Otter integration, so your line sees one feed instead of a tablet farm, but those tickets bring no diner identity with them. Every record in your book will come from a channel you own. That is the whole argument for direct ordering, stated as a data problem instead of a fee problem.
Pricing is flat: Operator $499 a month, Operator plus Marketplace $599, Concierge $699, each per location, month to month, no setup fee. The customer book is included on all three, with no per-contact charge, which matters because per-contact pricing quietly punishes you for succeeding. The only per-order cost is a service fee the diner pays at checkout, $0.99 pickup, $2.99 delivery, $0 dine-in, 10% catering. The restaurant keeps 100% of food revenue and 100% of tips. See pricing and how the fee works.
We are a Fort Lauderdale company, built with operators here. Naya Grill in Pompano Beach and West Palm Beach runs its direct ordering on Zayos today and keeps $48k+ a year that used to go to commissions. Storefronts are built for Shishka Lebanese Grill, La Vie Mediterranean, Yummy Grill, aura, Courtyard Cafe, Mr. Smoke across 13 locations, Yalla Market and Tap That Ash, each opening ordering at its own launch. Most operators are live in under two weeks, with the first direct order inside the first week. More on how it works.
The questions owners actually ask.
Is my POS customer list the same as a customer database?
Can I get customer data from DoorDash or Uber Eats orders?
How do I know if I actually own my customer database?
How many customer records does a restaurant need before it is useful?
What about duplicate records for the same person?
Do I need permission to email my own customers?
Terms defined plainly in the operator glossary.
Find out how many diners you are handing away.
The free AI report reads your public ordering channels and shows how much of your volume goes through a channel that keeps the customer. About 60 seconds, no card, no call.
Questions first? Talk to us.