Skip to content
Reserver
Start free
Horizonte 27 · Madrid, Spain

Fine dining with an always-full waitlist and no-shows under control

Horizonte 27, 22 seats in Madrid, moved from managing the waitlist on WhatsApp to a system with pre-auth and automatic alerts. Result: no-shows from 18% to 4%.

Global no-show rate
4%
before: 18%
Waitlist → booking conversion
62%
before: ~25% (manual)
Weekend effective occupancy
96%
before: 82%
Out-of-hours calls answered
100% (AI)
before: 0 (voicemail)

Starting point

Horizonte 27 is a 22-seat gastronomic restaurant in central Madrid. Single tasting menu, bookings 2-4 weeks in advance, average ticket of €180 per cover. The experience is tuned to the last detail — except for reservation management, which until a year ago ran on WhatsApp and a paper diary.

The concrete problem

With only 22 seats and demand 2-3x higher, the waitlist was part of the business model. When someone cancelled:

  1. The maître d’ saw the cancellation (sometimes with hours of lag).
  2. Opened the “Horizonte 27 Waitlist” WhatsApp group.
  3. Messaged the first in line.
  4. If no reply in 30 minutes, messaged the next.
  5. Frequently this happened during service — while the kitchen was saturated.

Result: ~25% of freed seats didn’t get filled in time. At €180 per cover, a lot of money evaporates.

The second problem was no-shows. Despite the high ticket and VIP profile, 18% of larger parties failed — especially bank holidays. “We were embarrassed to ask regulars of 10 years for a card,” the chef explains. Consequence: accepting the loss.

Why they changed

The trigger was a December night when two group reservations cancelled last-minute (13 covers) and they only managed to cover 4 before service. Estimated loss that night: ~€1,600.

The implementation

Reserver Plus configured in 3 weeks, prioritising three things:

1. Real pre-authorization for everyone

€50 per person, applied to all bookings without exception. The delicate part was communicating it to regulars. Solution: a personal email explaining the situation, the policy, thanking them for their understanding. Response: zero drop-offs. 4 questions. 2 congratulations for professionalising.

2. Automatic waitlist

Strict configuration:

  • Registration via widget only (phasing out WhatsApp gradually).
  • When a seat opens, automatic email+SMS alert to the first in line.
  • Acceptance window: 60 minutes for bookings under 24 h; 4 h for bookings more than 1 day out.
  • No response? Moves to the next automatically.

3. Voice AI

Dedicated number active outside hours (8 pm to noon the next day, plus all Monday, the closing day). The bot handles typical inquiries: availability, edits within the 24 h policy, tasting menu info, allergen policy. For complex decisions (party of 10+ on a special date, alternative menu request), it escalates to the maître d’ via internal WhatsApp.

4. Custom VIP tags

20 years of paper guest records moved to the CRM with custom tags: “Chef’s friend”, “Visiting sommelier”, “10+ year guest”, “Severe X allergy”, “Press”. These drive behaviour: some skip pre-auth, others get priority on the waitlist.

Results at 6 months

  • Global no-show: 18% → 4%.
  • Waitlist conversion: 25% → 62%.
  • Weekend effective occupancy: 82% → 96%.
  • Additional monthly revenue from better waitlist conversion: ~€9,000.
  • Out-of-hours calls: from voicemail to 100% answered by AI.

Unexpected effects

Three not-anticipated upsides:

Better online reviews

By drastically reducing “I went and they cancelled my booking because the waitlist got messy” experiences, Google rating rose from 4.6 to 4.8. Several reviews specifically mentioned the professionalism of management.

Less maître d’ stress

The maître d’ had inherited anxiety from carrying the list mentally. Since the system does it, she sleeps better (her literal words). Side effect: higher-quality attention on the floor.

Data they didn’t have

For the first time they know with certainty what percentage cancels, how many get replaced via waitlist, how long gaps stay open. This data is starting to inform policy: they’re considering lowering pre-auth to €35 because the data shows €50 was more than sufficient to deter.

What the chef would recommend

  1. “Communicate the change with a human voice.” A generic system email would have annoyed. A personal signed email explaining it didn’t.
  2. “Don’t fear pre-auth with loyal guests.” Regulars respect professionalisation. Those who don’t are less loyal than you think.
  3. “Voice AI replaces voicemail, not people.” They didn’t put it on during service hours — important calls had to reach a human.

Note: Illustrative case. “Horizonte 27” isn’t a specific customer; the numbers reflect typical patterns in high-demand gastronomic restaurants. If you run a fine dining venue with waitlist conversion problems, write to us: we can design a specific configuration for your case.

“The moment I stopped checking WhatsApp every night to see who had cancelled was liberating. The system alerts the next in line; I sleep.”

Chef Álvarez · Owner and head chef
More stories

See how other restaurants apply it.