Restaurant reporting that ties out, inside Zayos.
Published July 2026 · Updated July 26, 2026
Restaurant reporting should answer one question: which channel made money last week, after everything it kept. Read net revenue by channel, effective take rate, repeat rate, average ticket and refund rate. Anchor every report to your restaurant timezone, then tie one closed day to the bank.
What are you actually trying to find out on a Monday?
Most owners do not want analytics. They want three answers before the prep list gets written: did we make money last week, which channel made it, and is anything drifting. Everything a reporting screen does that is not one of those three is decoration.
It is harder than it sounds because a restaurant with delivery apps is running four or five businesses with four or five ledgers. Each app has its own dashboard, its own definition of an order, its own payout schedule and its own idea of when a day ends. Nobody built them to agree with each other, or with your bank.
So the job is not more charts. The job is one place where each channel is measured the same way, in your timezone, on money that actually arrived.
Why do none of my dashboards agree with each other?
Gross versus net
A marketplace dashboard usually shows what the diner paid. Your bank shows what arrived after commission, delivery fees, advertising and funded promotions, minus adjustments such as chargebacks and clawbacks that often land weeks later as one unexplained lump. Two different numbers about the same order, and reading the first as your revenue is how a busy month turns into a thin deposit.
Timezones and cutoffs
A platform that closes its day at midnight UTC will move your Friday late night orders into Saturday. If one report is anchored to your restaurant timezone and another is not, the two will never tie out and you will spend a Monday hunting a difference that is really a clock.
Refunds land later than orders
A Saturday order refunded on Tuesday sits in Saturday for the sales report and Tuesday for the bank. Any report that does not tell you when a refund was recorded against when the order was placed will drift a little every week.
Everybody defines an order differently
Cancelled before acceptance, cancelled after, partially refunded, a duplicate the guest placed twice. Two systems will count those four events differently, and the gap shows up as a percent or two you cannot explain.
None of these are conspiracies. They are ordinary engineering decisions made by companies that never had to reconcile with each other. Your job is to stop chasing the difference and start measuring one number the same way every week.
Which five numbers are worth reading every week?
- 01
Net revenue by channel
What actually hit the bank, split by direct web, direct app, and each marketplace. Not what the diner paid. This is the only number that answers where your money comes from.
- 02
Effective take rate per channel
Everything a channel kept, divided by what the diner paid. Marketplace commission is usually 15% to 30% before the extras and 25% to 35% blended once delivery, advertising and processing stack. Compute your own rate, do not quote a brochure.
- 03
Repeat rate
Of the customers who first ordered in a given month, how many ordered again within the next 30, 60 and 90 days. It moves slowly, so read it monthly, and it tells you whether the marketing worked or the discount just rented a stranger.
- 04
Average ticket, by channel
Watch this whenever you launch a promotion. Order count up and average ticket down usually means a discount reached people who were already ordering.
- 05
Refund and cancellation rate
Track it per channel and per location. A location drifting upward is almost never a customer problem. It is a prep time problem, a stock problem, or a menu that promises something the kitchen stopped making.
Compute the take rate yourself rather than quoting a headline percentage. Divide what a channel actually paid you for a period by what diners paid on that channel for the same period. The answer is often several points worse than the commission rate in the contract, which is the entire reason the blended number exists. There is more on that arithmetic in the Florida commission benchmark and the delivery commission statistics.
How do you check a report is telling the truth?
Ten minutes, once, on any system you are evaluating or already running.
- Pick a closed day at least a week old. Recent days are still moving because of refunds and payout timing.
- Tie the day to the deposits. Report total for that day, against what landed for that day. Write down the difference and find out what it is made of. If nobody can tell you, that is the finding.
- Check the clock. Ask which timezone the day boundary uses. If the answer is not your restaurant timezone, every late night comparison you make will be slightly wrong.
- Find one refunded order. You should be able to see the order date and the refund date, and see which day each one landed in.
- Try to export it. A number you cannot get out of a system is a number you are renting.
Run that on every vendor demo. It takes ten minutes and it separates reporting from a slideshow faster than any feature list.
What does Zayos actually report?
The main reporting screen carries revenue, orders and refunds across five ranges: today, yesterday, the last 7 days, the last 30 days and month to date. Every range is anchored to the restaurant timezone and stays correct across daylight saving changes, because a report that quietly shifts an hour twice a year is worse than no report.
Channel mix
The channel breakdown separates direct web and direct app from Uber Eats, DoorDash, Grubhub and Postmates, so the question the whole platform exists to answer, how much of my volume is mine, has one screen. Marketplace orders arrive through Otter, the aggregator that pulls every app into one kitchen tablet, so they are counted next to your direct orders instead of inside somebody else’s dashboard.
Retention, not just revenue
The cohort report answers the question most restaurant dashboards skip: of the customers who first ordered in March, how many came back in April, in May, in June, and what has each cohort been worth since. That is the number that tells you whether a promotion bought a customer or rented a stranger.
More than one store
A cross location leaderboard ranks every store on revenue, orders, average ticket, refund rate and top item, and colors each metric against the group average so an outlier is visible in a glance rather than in a spreadsheet. Scope is honored throughout: a manager pinned to two stores sees two stores, an owner sees the group. Mr. Smoke has 13 locations built on this structure, with ordering opening at launch.
Reports that come to you
Scheduled reports email a chosen recipient list on a cadence you set, which is the only version of reporting that survives a busy service. The report an owner reads is the one that arrives.
All of it is included in the plan. Operator $499, Operator plus Marketplace $599, Concierge $699, per location per month, month to month, no setup fee. Naya Grill in Pompano Beach and West Palm Beach runs on Zayos today and keeps over $48,000 a year that used to leave as commission, at roughly 3,000 orders a month.
Reporting questions owners ask
Why do my delivery app numbers never match my bank deposit?
Which restaurant reports actually matter every week?
How do I know whether a report is telling me the truth?
What reporting does Zayos include?
Can a manager see reports for only their own store?
Do I own the underlying data?
Start with the number you do not have.
The free AI report reads your current ordering setup and returns your channel mix and what it is costing, in about 60 seconds. No card, no call.
Or read how onboarding runs, most operators are live in under 2 weeks.