OBVIO Hotels (Accor) + IDeaS G3

OBVIO replaced spreadsheet-driven forecasting across 44 Accor-branded properties with IDeaS G3 starting in 2022. The deployment sits inside the broader Accor + IDeaS global revenue-management partnership (Feb 2024), which is being rolled out across 5,000+ Accor properties in 110+ countries.

IDeaS' broader footprint is 31,000+ properties in 169 countries, OBVIO is one lit node in a much larger network, and the published numbers are programme-level, not single-property heroics.

G3 produces a default rate recommendation per room-type and rate-code, optimising price, restrictions, and overbooking jointly. Inputs: booking pace, on-the-books pickup, market demand signals, historical seasonality, and competitive positioning. The system re-optimises the entire annual pricing book daily.

Operationally, this shifts the revenue manager from daily 3-month + weekly 12-month manual reviews to G3-flagged "high-impact dates only", with the system holding decision authority on day-to-day rates and the human focused on exceptions (conferences, sports, weather, political disruption, events where training data goes stale). The pattern is: AI recommends; the human decides on what matters.

• ~20% RevPAR uplift since 2022 across the OBVIO programme
• +14% ADR year-on-year
• +9.5% RGI (Revenue Generating Index) for 2019–2023
• Forecast accuracy is the lever Arnaud Fougeront, Director of Commercial & RM, calls out by name: 'The real strength of the tool for me is the forecast accuracy.'

The clean separation of AI forecast vs human decision. Staff retain override authority; the scorecard measures the system as a whole, model + manager, not the model in isolation.

The transferable architectural pattern is forecast-and-optimise ML paired with a named decision owner who reviews exceptions: conferences, sports, weather, political disruption, events where training data goes stale.

Accor's scale gives IDeaS data advantages specific to the Accor network. A single boutique property will not replicate Accor's performance curve on day one; the RMS takes months to learn a property's demand pattern.

The numbers are a portfolio result over multiple years; your own ramp will be slower and lumpier. RGI is a market-share index that depends on a defensible comp set; if your comp set is wrong, the metric is meaningless.

• Is there ~6 months of clean PMS data and a working rate-shopping feed? If not, fix that before piloting.
• Do you currently run a manual revenue process you can compare against? Without a baseline, ROI is unmeasurable.
• Is a named revenue manager prepared to override the model on edge events, or will the team default to 'what the system said'?
• Whose comp set, and how is it agreed? RGI moves with the comp set definition.

IDeaS is a SAS Institute company and the longest-tenured incumbent in hospitality RMS. G3 is its current flagship: 100+ demand models running classical machine learning (an ensemble of demand-forecasting models with per-property auto-weighting feeding a dynamic-programming optimiser), not an LLM.

The model class matters: G3 is auditable, the inputs are tabular, and the failure modes are statistical, not hallucinatory.