Gains:
- Ability to calculate and interpret revenue management indicators such as RevPAR, ADR and occupancy with artificial intelligence support
- Ability to have artificial intelligence draft and evaluate dynamic pricing scenarios based on demand, season and competitor data
- Being able to understand that the price decision belongs to the human being, taking into account the market, brand position and ethical boundaries, and that artificial intelligence only produces scenarios.
The same room in the same hotel can be sold for 1,800 TL on a Tuesday night, 3,200 TL on the weekend, and 4,500 TL during a festival week. This is not an inconsistency, it is a method: it is called revenue management. Its goal is simple: to sell the right room to the right guest, at the right time, at the right price, through the right channel. In this unit, you will learn how to use artificial intelligence (AI) to calculate revenue indicators, interpret demand and competitor data, and generate dynamic pricing scenarios. The critical principle is from the outset: the price decision is a very high-risk decision; AI produces scenarios, humans determine the price.
Three key indicators of revenue management
Before we talk pricing, let's clear up three numbers.
Occupancy rate: Number of rooms sold divided by the number of rooms available for sale. If 96 rooms are sold in a 120-room hotel, occupancy is 80% (96 / 120).
ADR (Average Daily Rate): Room revenue divided by the number of rooms sold. If 96 rooms were sold for a total of 230,400 TL for one night, ADR = 230,400 / 96 = 2,400 TL.
RevPAR (Revenue Per Available Room): Room revenue divided by the number of salable rooms; alternatively occupancy × ADR. In the example above, RevPAR = 230,400 / 120 = $1,920, or 80% × 2,400 = $1,920. RevPAR is the heart of revenue management because it combines both occupancy and price into a single number.
Why are both important? Because high occupancy is not always good. Selling rooms very cheaply and achieving 100% occupancy may be worse than selling slightly more expensive and achieving 85% occupancy and higher RevPAR. AI can calculate and display these two scenarios side by side when you give the data; Which one you choose depends on your strategy and market knowledge.
indicator
formula
What does it measure?
occupancy
Sold room / Sold room
How full are you?
ADR
Room revenue / Rooms sold
How much did you sell it for on average?
RevPAR
Room revenue / Salable room
How well did you evaluate your capacity?
TRevPAR
Total revenue / Salable room
Efficiency including non-room revenue
What is dynamic pricing
Dynamic pricing is when the room price is not fixed but changes according to demand, season, remaining availability, day, events and competitor prices. It's the same logic we're used to when it comes to flight tickets. Inputs include: demand signals (search, booking rate), pickup (how quickly reservations come in for a given date), lead time (time between booking and stay), competitor prices, local events calendar, and your cost base.
AI is a powerful helper here, but in a limited role: when you give inputs like demand, occupancy target, competitor range, and cost, AI can chart expected occupancy and RevPAR scenarios for different price points, explain the logic, write a pricing justification paragraph. What it cannot do: know the real market situation, see the competitor price live and determine the final price. These are your jobs.
Attention: Telling AI to "determine tomorrow's price" means making it decide on an issue for which it does not see the market. Ask the AI for a price “scenario,” not a price “decision”; You decide with your market knowledge.
Ethical and legal boundaries
Dynamic pricing has its limits. When making personalized pricing, applying discriminatory prices based on the guest's personal characteristics (e.g. device, location) carries both ethical and legal risks. Exorbitant price increases during extraordinary periods such as disasters and epidemics are seen as opportunism and may be sanctioned. Additionally, price transparency and consumer legislation require avoiding hidden fees. When AI recommends a price, human filtering that respects these boundaries is essential.
Step by step: Creating price scenarios with AI
- Prepare the indicators. Anonymously collect current occupancy, pickup rate, remaining days, cost base and competitor price range.
- Specify the destination. Do you prioritize RevPAR, occupancy, or filling by a certain date?
- Ask for a script. Tabulate expected results for different price points.
- Verify. RevPAR = occupancy × ADR calculation manually; Verify competitor data from the real source.
- Decide. You choose the final price with your market knowledge and enter it into the system.
three mini cases
Case 1 — Reading RevPAR correctly. One resort boasted 98% occupancy for the weekend, but profits were below expectations. The manager gave anonymous occupancy and ADR data to AI; YZ commented, "RevPAR dropped because high occupancy came with low ADR" and tabulated two alternative price scenarios. The human chose the 90% occupancy target with a slightly higher price; RevPAR rose 11% the following weekend. AI gave the account and the scenario, the decision was up to the human.
Case 2 — Unverified competitor data. A revenue expert asked AI, “What is the price of competing hotels today?” The AI “gave” confident prices for three hotels even though it had no access to live data. This was all fabrication; The specialist lowered the price accordingly and lost unnecessary revenue. The right way was to get the competitor price from a real price monitoring tool or manual check, use AI only for interpretation.
Case 3 — Ethical boundary. One team asked the AI for an "automatic higher price for guests calling last minute" logic. AI produced a blueprint; but the manager realized that this could systematically punish some guests and undermine the principle of transparency. The logic has been translated into a last-minute schedule that is clear and consistent to all guests. AI suggested it, the ethical filter came from humans.
Weak prompt / Strong prompt
Weak prompt:
Determine our room price for tomorrow and write down the price of your competitors.
This claim is doubly flawed: the AI knows neither your actual demand nor the live competitor price; fabricates both.
Powerful prompt:
Your role: revenue management assistant. The price decision is mine; You create a scenario. Data (anonymous, real): 120-room city hotel, 78 rooms are full for tomorrow, the remaining 42 rooms, pickup has accelerated in the last 3 days. The cost base is ~900 TL per room. I give the competitive range: 2,200–2,800 TL. My goal: Maximize RevPAR. Task: State reasonable occupancy assumption for price points of 2,200 / 2,500 / 2,800 TL (write that it is an assumption), calculate the expected RevPAR and put it in the table, show the calculation. Competitor DO NOT make up the price; Use the range I gave.
Template: income indicator calculation:
Anonymous data: room sold [x], room sold [y], room revenue [z]. Task: calculate occupancy, ADR and RevPAR with their formulas, comment in a single paragraph. Just work with the numbers I give; fabricate other data.
Template: pricing justification text:
Context: I set the price [x] TL for [date] because of occupancy [%], pickup [status], event [present/absent]. Task: translate this decision into a clear 4-sentence justification note for the team. Adding a new number; Be limited to the data I gave you.
Template: seasonal price strategy outline:
Context: [property type], upcoming season [high/low], target [RevPAR/occupancy].Task: draft a framework that proposes different pricing logic for weekdays/weekends/event days. I will put the concrete price; You list the logic and ethical/legal considerations.
Common mistakes
- Let AI make price decisions. AI does not see the market; People determine the price with market knowledge.
- Asking the competitor's price from the AI. AI cannot access live price; If asked, he will make it up. Get it from the real source.
- Just looking at the occupancy. High occupancy may mask low RevPAR; Read both together.
- Crossing the ethical/legal boundary. Exorbitant price increases and discriminatory prices create risks; human filter is required.
- Mistaking the assumption for reality. Have the AI's occupancy assumption marked as "assumption" and test it with your own data.
Tip: In each price scenario, ask the AI to explicitly write out its occupancy assumptions. So you can always see the question "what assumption is this scenario based on" and compare it with reality.
In summary
Revenue management is the art of optimizing revenue per salable room (RevPAR), not occupancy. Three indicators — occupancy, ADR, RevPAR — are the language of this business. In dynamic pricing, AI generates scenarios, calculates and writes justification when you give inputs; but he does not see the real market, does not know the competitor's price and cannot make a price decision. The price is a very high-risk decision and belongs to the human being along with its ethical-legal limits.
Application task
Get one night of anonymous data: rooms available for sale, rooms sold, room revenue. Have AI calculate occupancy, ADR and RevPAR with formulas and verify manually. Then have the RevPAR scenario generated for three price points with the "Strong prompt" price scenario template; Tick the occupancy assumptions and write in 5 sentences which price you would choose and why.
checklist
- [ ] Have I manually verified occupancy, ADR, and RevPAR?
- [ ] I took the competitor's price from the real source, didn't I make it fit by AI?
- [ ] Have I clearly marked the AI's occupancy assumptions?
- [ ] Did I observe ethical and legal limits in the price decision?
- [ ] Did I determine the final price with my market knowledge?