Meta Gap Analysis

Web app for MetaHotelsA daily dashboard that shows a hotel group where it loses direct bookings to sites like Booking.com, how much it costs, and what to do about it.

Product designWeb app development
Data processingAI features
Meta Gap Analysis command centre and world map on a desktop and a tablet

/ Overview

When a room is cheaper on Booking.com than on the hotel's own website, guests book there and the hotel pays commission on a booking it could have had directly. Across dozens of hotels, that adds up fast. MetaHotels already collected the right data, but it arrived every day as huge spreadsheets nobody had time to read. We built Meta Gap Analysis to turn those files into clear dashboards, daily emails and AI-written action plans.

/ Project details

Client:MetaHotels

Industry:Hotel software - revenue

Platform:Secure web app (desktop & tablet)

What we did:Product design, Web app development, Data processing, AI features

/ Results in numbers

47+

Hotels in
one place

8

Booking sites
compared

6

Data sources
read every day

80x

Faster review charts
(~406 ms to ~5 ms)

/ The problem

Lots of data, no answers. Revenue managers needed four simple answers, quickly, for every hotel.

/ What was going wrong

Where are we losing?

Which hotels, which booking sites, which countries and which devices.

How much is it costing?

Lost revenue in money, not just percentages.

Why is it happening?

Is it mobile-only pricing, booking timing, country-based prices or a booking site undercutting the hotel?

What should we do today?

A clear next step for each hotel, sent to the person responsible.

/ What we did

Raw files in, clear actions out, every day. The data now arrives, gets checked, is turned into dashboards and lands in managers' inboxes without anyone lifting a finger.

/ The work

Automatic daily data

Six data sources are collected every morning, matched to the right hotel, checked for duplicates and converted into one currency.

A command centre

One screen answers 'how are we doing right now?': money lost, bookings lost, and whether the hotel's price beats, meets or loses to each booking site.

Deep-dive views

Worst offenders, an interactive world map of where losses come from, device and date patterns, and screenshots as evidence.

Price-match and booking tracking

See how often the hotel's own 'we'll match any price' tool works, and where guests drop out on the way to booking.

Guest review insights

Scores from Google, TripAdvisor and booking sites, boiled down to what to fix first and what's worth promoting.

AI action plans

Pick a hotel and dates, and the app writes a prioritised to-do list across pricing, direct bookings, SEO, ads and reputation.

Daily emails and PDFs

Managers get a report every morning, and any dashboard can be saved as a print-ready PDF in one click.

/ Before and after

The difference, in plain terms. Where MetaHotels started, and where it is today.

/ What changed

Seeing price gaps

Before
Reading through raw spreadsheets by hand
After
One place for 47+ hotels and 8 booking sites

Lost revenue

Before
Rows of data nobody could act on
After
Losses shown in money, so teams fix the costliest first

Next steps

Before
No clear action for each hotel
After
AI action plans and daily emails to the right people

Daily work

Before
Big files landing in storage every day
After
Processed, saved and reported automatically

Speed

Before
Review charts took ~406 ms
After
~5 ms, with booking charts nearly twice as fast
/ For the technical readerHow it was built, and the tools we used

Ingestion

Six daily feeds from Azure Blob Storage (website parity, price-match reports, guest funnel, hotel services, Google Hotel Ads parity, reviews) with file-pair validation, hotel matching, idempotent dedupe, batch audit logs and nightly FX conversion. Admins can re-run and backfill without a developer.

Performance

Indexes and queries tuned against live measurements: review trends ~406 ms to ~5 ms, booking-funnel queries roughly halved. Two-layer caching with Redis, invalidated after each ingestion, falling back to the database if Redis is down.

Security

Role-based access with hotel-level data scoping on every API, reCAPTCHA v3, per-IP rate limiting, account lockout, CSRF origin checks and short-lived sessions.

AI & reporting

Per-hotel context assembled from parity, price-match, funnel and review data, streamed to GPT-4o for prioritised actions with impact/effort ratings; React Email daily reports; headless-Chrome PDF export of any filtered dashboard.

Delivery

Grew from v1 to v4.4 through additive database migrations with no historical data lost. SMTP, thresholds, branding and AI settings are configurable from the admin console.

Built with

Next.js 16React 19TypeScriptTailwind CSSshadcn/uiRechartsD3-geoPrismaPostgreSQLRedisAzure Blob StorageNextAuth.jsOpenAI GPT-4oReact Email
Photography
3D Models
Development
Illustrations
Fashion
Digital Art
Packaging
Motion
Illustrations
Video Production
Photography
3D Models
Development
Illustrations
Fashion