AdverDia Group  /  Nassau, The Bahamas

Now you can see the floor.

In virtually every restaurant, the person with the most influence over the guest experience is the server. Their performance is the least measured thing in the building.

0:00

A guest presses. The clock starts. The assigned server clears it on arrival, and that elapsed time is the backbone of everything else.

The POS measures the money. The KDS measures the kitchen. Nothing measures the floor.

Ask any GM who their best server is and you get a confident answer in two seconds. Ask them to prove it with a number and the room goes quiet.

That isn't a management failure. Food cost, waste, labour percentage, ticket times, cover counts and average check are all tracked to the decimal, because the tools exist to do it. For the floor, they don't. So it gets assessed through three unreliable channels.

The manager's walk

“How is everything?” Guests are polite, the sample is tiny, and it's biased toward whichever tables a manager happens to reach.

The complaint

Someone has to be unhappy enough to speak up. Most never come back and never say why. You lose the customer and the lesson.

Tips and sales

Pooled, lagging, skewed by check size. A weak server on a great table still looks fine on paper, and you find out weeks later.

What gets measured gets managed. The floor is barely measured, so it's coached on feel, rewarded on tenure, and corrected only after the guest is already gone.

One centerpiece on the table. Two things being measured.

Two instruments sit inside the same object. One reads your operation, one reads your guest. They measure different halves of the same dinner, and neither substitutes for the other.

The button

Measures response

  • Response time, per server, per tableNot a shift average you have to interpret.
  • Repeat pressesHow often a guest had to ask twice. The tell no other system catches.
  • Peak-hour behaviourMeasured against that server's own baseline, so you see who falls apart at 8pm.
  • TimeoutsPresses that were never cleared at all.
  • No app, no download, no language barrierThe guest at your table may not speak English. They can still press a button, which is why your data is close to complete rather than a self-selected sample.
The button is fact. It happened or it didn't, and the clock doesn't have an opinion.

The QR panel

Measures satisfaction

  • One scan, three fields, ten secondsA 1–10 rating, one open question, and an email address.
  • One question only“How was your experience?” Guests won't complete a survey after dinner. They will answer one question.
  • Covers the meal, not just the serviceThe half the button cannot see. Today a kitchen problem and a service problem look identical to a GM. Here they separate.
  • Opted-in guest emails, owned by youCollected at the table, exported to the CRM you already run.
The QR is judgment. It's how the guest scored you, in their own words.

Fast service that guests didn't enjoy is a kitchen problem. Slow service that guests forgave is a staffing problem. You cannot tell those apart with one instrument.

Décor, not a gadget.

The button sits inside an object that belongs on your table. Custom manufactured, finished to match your room. The QR panel is designed into the same piece rather than propped against the salt.

Candle holder

Dinner service, low light.

Bud vase

Brunch, terrace, garden rooms.

Table number stand

Counter service, pool, casual.

Reserved-sign block

Fine dining, cabanas, day beds.

Finished to your room

Material, finish and form are specified per outlet. A steakhouse and a beach bar don't get the same object.

Built for a dining room

Weighted, wipeable, and heavy enough not to walk. Charged on a cycle your team can run alongside candle service.

The QR is part of the piece

Etched or printed into the object itself, so it reads as part of the setting rather than as a table tent.

Your POS already knows who has that table.

Nothing works until each press is tied to the right server. BUTN takes that link from the check, which means nothing new to maintain during service.

1

A guest presses

The clock starts and the request routes to the server assigned to that table.

2

The check is the assignment

When a server opens a check on a table, your POS already knows who is responsible for it. The press scores against that server automatically. Nobody has to build a section map, update it on cuts, or maintain it through breaks. Where a POS connection isn't available, sections can be assigned by hand instead.

3

The server arrives and clears it

Cleared at the table, which stops the clock and logs the elapsed time against the table and the server.

4

The floor sees its own numbers

Your kitchen got faster when ticket times went on a screen. Nobody had to be lectured about it. Measure the floor, show it the same way, and service improves before management has to intervene. Then measure again and prove it moved.

A bad night is fixable while the guest is still sitting there.

A rating at or below your threshold pushes straight to the manager on duty with the outlet and table number attached. They walk over. They fix it. The guest leaves having been looked after instead of leaving quietly and never coming back.

Every alert and every intervention is logged, so you can see how many were actioned and by whom rather than assuming it happened.

Manager alert  8:42pm 4/10

Ember & Tide  /  Table 14  /  Section B

Food was great but we waited a long time for drinks and had to ask twice.

Actioned 8:47pm by D. Rolle

Four lines instead of three hundred comments.

Ratings are a number. What guests write is where the information is, and today it dies in a spreadsheet nobody opens. Every response is sorted into stable categories — food quality, service speed, staff courtesy, cleanliness, atmosphere and value — so the data stays trendable. Emergent detail sits underneath, so a spike about one dish or the music volume is still visible.

Ember & Tide, week of 12 May

141 responses, average 8.6, up from 8.3. Problem incidence 9.2%, down from 11.8%. Six recovery alerts fired, five actioned at the table.

DimensionMentionsDirectionSentiment
Staff courtesy47upPositive
Service speed31upNegative
Atmosphere18flatMixed
Food quality11downNegative

Connected issue

Service-speed comments are almost entirely about the first drink, not the food, and cluster on Friday and Saturday second seating. Response times run above this room's own baseline in the same window and problem incidence hits 17%, nearly double the average. Points to bar throughput or floor coverage on those two shifts.

Emergent theme

Food-quality mentions concentrate on two dishes and have fallen since the 3 May menu change.

Reported weekly for a single room and daily across a property, and suppressed below a minimum sample so a summary is never built on a handful of comments. Sample data shown.

Nobody has to be told to hustle when the numbers are on the wall.

One screen in the service area carries the open requests, the state of every position, and the shift's top performers. Servers start racing each other because the metric is visible, individual and fair. The board shows the top and the shift average; anything below that goes to the manager's dashboard, privately.

BUTN   Ember & Tide
Avg wait 1:04Longest 2:47Requests 1478:47

Open requests — 4

Table 12Section B / James / pressed twice
2:47
Table 3Section A / Tanya
1:12
Table 21Terrace / Maria
0:18
Table 7Section A / no check open
0:41

Floor — 34 positions

Section A

Section B

Terrace

Open Seated New Waiting Overdue

Tonight's top

1  Maria0:42
2  Tanya1:01
3  Devon1:06
Shift average1:04

The same layout carries a dining room, a cabana row or two hundred lounge chairs. Sample data shown.

From one restaurant to forty outlets, the same two instruments.

A single-room owner and a resort F&B director ask the identical question at the end of service: who was fast, who wasn't, and what did the guest think about it. The only difference is how many rows are in the report.

Ember & Tide, Tuesday 12 May

Service closed 10:41pm

Presses

147 / 112 covers

Avg response

2:38 down 0:03

Timed out

2 / 147

Recovery alerts

2 / 2 actioned

Second seating, 8:15 to 9:30pm

Response ran 0:36 above this room's own baseline. Nine of the twelve slowest presses of the night fell in that window.
ServerAvgUnder 60sRepeatPeakScore
Maria0:4294%3%+0:0892
Tanya1:0181%7%+0:2484
James1:2863%14%+1:1168
Shift average1:0479%8%+0:3481

Read the last two columns together. James isn't just slower on average; he's asked twice more than four times as often as Maria, and he falls apart at peak. That's a coaching conversation with specifics in it. Scores are normalised for section size and table count, so a server working twelve busy tables is never ranked against one working four quiet ones. Sample data shown.

The annual study tells you where you finished. BUTN tells you where you're heading.

Your property makes real decisions on annual third-party guest satisfaction studies. They're credible, externally comparable and worth the attention they get. They're also slow: a full year of guest stays, published months after that window closes. A dinner that disappointed someone in June can surface the following July.

BUTN is the leading indicator for the same two dimensions your F&B leadership is already accountable for, food and beverage and staff service, reported from your own rooms while the year is still in play. And it tells you something no property-level score can: which room is responsible. A single score for the property is a verdict. A score per room, per shift, per day, is a plan.

One room. Ninety days. A report on your own numbers.

We install in a single high-visibility restaurant, train your managers, and run for a full quarter. At the end you get a before-and-after on response times, repeat presses, timeouts, guest ratings, problem incidence and recovery alerts actioned, set alongside your own covers and average check across the same window.

If the numbers move, expand. If they don't, you've spent one quarter finding out.

What we handle

  • Hardware and custom centerpieces finished to your room
  • Installation and the service-area display
  • POS connection, scoped with your IT team
  • Manager and staff training
  • Support on the island, same-day device replacement

What you provide

  • One room
  • Access to your POS for server assignment
  • Thirty minutes of your GM's time at kickoff