There are two secrets to revenue management that fewer than 1% of revenue managers know. They use these secrets to grow revenues faster with less work. Over the next 5 years these secrets will enter the mainstream thinking and become the new revenue management foundation.
At Pace, we can't wait for the future, so I'll go ahead and share the future here. But don't worry about the general public catching on any time soon. We predict it will take at least 5 years for the trends to become common practice - which is fantastic for early adopters since it means an unfair advantage in your favour over those years.
Broadly speaking both secrets emerge from an understanding of how machine learning is changing the hotel industry. How these changes will play out can be clearly seen in other industries that are ahead of hospitality. So let's begin there.
From the time of Mad Men (the early days of the advertising industry in the 60s) until 10 years ago the advertising industry worked roughly in this way: if you needed 10,000 new customers you would go to an advertising agency. They would analyse your product and customers and categorise new potential customers into segments. Ever used the word yuppie? Comes from advertising - Young Urban Professional. DINK? Double Income No Kids etc.
Once the agency had your segments worked out, they looked at how to reach them. Targeting DINKs? Let’s do 60% TV, 40% newspapers. Of course, they also consider which TV shows, times and optimal frequency. At the end of the process you get a detailed menu of which times/programmes/pages they’ll show your ad in. Don’t forget to pay the £200,000 invoice before the campaign kicks off.
Today it works differently. In the last 15 years 60% of advertising has moved online. And nobody uses segments. The modern consumer isn't a DINK or a YUPPIE. They are: A Facebook profile who is connected with my existing customers, who has searched for one of my products in the last 30 days, who lives within the M25 of London, 25-30 years old with an income greater than £35,000.
Conclusion: Data and machine learning mean much higher accuracy because you can go deeper. Every time we go deeper we throw away previously useful tools which have now become helplessly abstract.
The traditional way of doing revenue management is the following:
Time to break some eggs
Do you have 100s of rate-codes aimed at 100s of hypothetical segments? You're wasting your time.
Over-segmentation is similar to advertisers trying to work out if people who listen to Bon Jovi like muesli or toast. Your segments need to be measurable - otherwise they serve no purpose because you don't know what your work is achieving. For example, channels are measurable so a truly valuable experiment will emerge from rate-parity being challenged.
Another example would be to add discounts to booking engines (with the goal of moving direct share of bookings from 10% to 12%). In short, most of your rate-codes have no measurable outcome, timeframe or goal. It's time to simplify: get rid of your rate-codes and start again with a smaller number that have quantifiable goals.
Let's make an omelette
Spoiler alert: Knowing how many experiments to run at the same time and what conclusions to draw are not easy tasks.
Pricing is without a doubt the single most important experiment you are running. At least I hope you're running one. If you've never moved prices up and down to understand what impact they have on bookings you're not doing revenue management.
The point of pricing experiments is to understand how price sensitive your potential guests are at different points in the booking curve. What happens if you increase prices by 10% during the final week before check-in? How about a 5% price increase?
The results of these experiments are multiple forecasts for what you occupancy, ADR and RevPAR you will achieve at different price-points. This allows you to choose the optimal price at every point in the booking-curve for every single night.
For those who cheated and skipped straight to the end here's the answer. Now scroll back up and learn why!
- Manual segments and rate-codes are holding you back - simplify
- Unconstrained forecasting will be replaced by scenario forecasting