How Kei Concepts made $4,266 more in four weeks
They swapped a printed text menu for video on every item at their Westminster coffee house, on iPads at the counter and a QR code on every table. Everything below comes from Kei's own Toast reporting for the four weeks that followed, including the one target the pilot missed.
- +$4,266
- more revenue than the same orders would have taken before
- 26,836
- orders measured, every one of them counted
- 1 in 5
- guests opened the menu on their phone. That is 21.6% of orders
At a glance
Kei ConceptsWestminster, California- Shop
- Kei Coffee House
- Pilot window
- August 25 – September 21, 2025
- Point of sale
- Toast
A PDF, and a lot of questions
Kei Coffee House sells drinks whose names do not describe them.
A Pistachio Quattro Latte is a lovely thing and an unhelpful phrase, and the menu guests were handed said exactly that and nothing else. A name, four ingredients and a price. The gap was closed by staff, one guest at a time, all day.
What a guest saw before
What a guest sees now
- A text-only list, the same on paper and on the wall
- No photograph of anything
- Names that describe an idea, not a drink
- Anything else you wanted to know, you asked the counter
- Every item with its own video, playing
- Top Sellers and Staff Picks, chosen by the shop
- An iPad at each cashier station, and a QR code on the table
- What is in it, written down, before anyone has to ask
Four weeks, measured on the till
Before the menu went in, an order at Kei averaged $16.05. That number is the line everything below is measured against.
It is a per-order figure on purpose. How many people walk through the door swings with the season, the weather and whatever was spent on advertising that month, and none of that tells you whether the menu did anything. What guests put on a single order does.
Each bar is that week’s orders multiplied by how much its average check beat the $16.05 baseline. Together they make +$4,266, the figure at the top of this page.
Week 4 is shown after adding back the $3,319.70 of back-to-school discounts Kei chose to run that week, so all four weeks measure the same thing: what guests chose to order. Counting those discounts as lost revenue, week 4 comes in $2,586 under the baseline instead, and the four-week total is +$946. Both numbers are in the table below.
| Week | Menu scans | Orders | Scanned | Average check | Versus baseline |
|---|---|---|---|---|---|
| Week 1Aug 25 – 31 | 1,498 | 6,777 | 22.1% | $16.41 | +$2,440 |
| Week 2Sep 1 – 7 | 1,502 | 6,800 | 22.1% | $16.10 | +$340 |
| Week 3Sep 8 – 14 | 1,257 | 6,270 | 20.0% | $16.17 | +$752 |
| Week 4Sep 15 – 21 | 1,541 | 6,989 | 22.0% | $16.15$15.68 as rung up | +$734−$2,586 as rung up |
| Four weeks | 5,798 | 26,836 | 21.6% | $16.21 | +$4,266 |
Week 4 ran $3,320 of back-to-school promotions. Those discounts were a marketing decision the menu had nothing to do with, so the check-size column adds them back to show what guests chose to order. Exactly as the till rang it, week 4 came in under the baseline and the four-week total is +$946 rather than +$4,266. Both numbers are on this page because only one of them is the answer to the question the pilot asked, and a reader is entitled to decide which.
What the four weeks say
Every week beat the baseline
Not once, and not on average. All four. The lift is 16¢ an order, which is small until you multiply it by 26,836 orders.
Scanning missed its target
The pilot wanted 25% of orders to come with a scan and got 21.6%. The iPads at the counter carried the result; the QR codes on the tables under-delivered.
One month is one month
Four weeks in one café is a real result and a small sample. It is enough to act on and not enough to promise, which is why nothing here is presented as an average across restaurants.
The money arrived as things people had not planned to buy
A bigger check is either more items or dearer ones. Kei’s own item reporting says it was the first.
And specifically the small things nobody walks in for. Fries, popcorn chicken and a slush, the items you add because you just watched one. They sold 31% more between the two periods.
Yogurt slush: Blueberry stands in. The strawberry version is a modifier Toast cannot report separately.
Strawberry matcha latte: Regular and double combined, because the double was split into its own item mid-period.
+40%
Strawberry Yogurt Slush
+22%
Popcorn Chicken W/ Aji Sauce
−12%
Beef Pate Chaud
−51%
Banana Toffee Latte
Why the drinks fell
Because this comparison runs from summer into autumn, and the two items that halved are iced drinks.
Between them they lost 52% of their sales. No one should read a café selling less iced matcha in November as a verdict on its menu.
Take those two out and the rest of the tracked menu was 3% up. Flat to slightly better, in a quarter when fewer people came in at all.
Why the check is the honest number
Total sales move for a dozen reasons a menu cannot claim credit for. Average check moves for very few.
That is why the pilot measured it, and why the four-week result and this item comparison can point in different directions without either being wrong. One is measuring what people ordered. The other is measuring how many of them showed up.
What the staff said
All 4 front-of-house staff working the pilot answered a survey at the end of it. Four is a small number, so the findings below are counted in people rather than turned into percentages.
- All four
- 3 of 4
- 3 of 4
“I get far less questions of what components are in a drink.”
Front of house, Kei Coffee House
“It is way nicer to have a visual representation of the menu rather than trying to explain items based off of a text based menu.”
Front of house, Kei Coffee House
And what was worse
Two of the four said taking an order took slightly longer. The reason one of them gave is the most useful sentence in the survey:
“I think the pictures that were introduced provided a lot more attraction to certain drinks that had names that may not sound as appetizing as they look, and as a result increased the time customers spent to find a drink they may like.”
Front of house, Kei Coffee House
Guests spent longer choosing, and spent more. Those are the same sentence told from two sides of the counter.
Run the same four weeks in your shop
We will build your menu first, from your real items, and you can look at it before you decide anything.
How this was measured
- Where the numbers come from
- Kei Coffee House’s own Toast reporting: weekly orders, average order value, QR scans and a discount report. Plus a survey of the 4 front-of-house staff who worked the pilot. PlatePost did not measure any of it.
- The window
- August 25 – September 21, 2025, at the Westminster location only.
- The baseline
- $16.05, Kei’s average order value before the menu went in.
- How extra revenue is worked out
- For each week, its orders multiplied by the difference between that week’s average check and the $16.05 baseline. The four weeks are then added together. No modelling, no projection and no annualising.
- The adjustment
- Week 4’s $3,319.70 of back-to-school discounts are added back before its average check is calculated. That is the exact figure from Kei’s Toast discount report, not a rounded one. Without that adjustment the four-week total is +$946 and week 4 itself is −$2,586.
- The item comparison
- June to August 2025 against September to November 2025. A different and much wider window than the pilot, spanning a change of season, and overlapping the pilot’s first week. It is shown to say which items moved, not to size the result.