Machine Learning and the Future of Hospitality Software

hotel data

Machine learning and Artificial Intelligence are becoming increasingly prevalent in modern society. Not so long ago, the terms were largely confined to the realms of science fiction with autonomous spaceship pilots and self-aware machines, often of malicious intent.

The introduction of digital assistants, automated appliances and intelligent homes has brought a sense of that science fiction to our daily lives but increasingly, machine learning is being applied in the operation of businesses. How they plan production, how they forecast demand and how they target their customers.

To most people and smaller businesses though, machine learning is still shrouded in mystery and the possible benefits are not well understood. So, what exactly is machine learning and how could it benefit hospitality businesses?

Programming the Unprogrammable

In standard programming, we use various operators and conditional logic to achieve a functional goal. Think of it as a digital set of instructions. If you had to explain to me how to make a cup of tea you could write down a step by step set of instructions. These instructions may have conditions too; is the water boiled? If yes, pour into cup, if not, continue to wait for the kettle and so on. I will avoid the highly sensitive issue of when to add the milk…

However, some things we know how to do but we would find almost impossible to explain in the same way. Think about an image you are given, could you explain to someone how to answer the question; ‘does this image have a face in it?’. Without contextual and experiential knowledge of what a face is, it would be impossible, there are simply too many variations in shape, size, and colour etc to define an accurate procedural process.

So, how does machine learning try to solve this problem? In basic terms, machine learning is the process of analysing large amounts of data with known results. In the example above it would be processing thousands of images with and without faces, transforming the data into numerical vectors and then ‘training’ the computer to build a predictive model based on the data. The more data that is processed, the more accurate the model will become.

There are 3 main types of machine learning: Categorisation, Clustering and Regression. Categorisation, as you may guess, is using the predictive model to guess how to categorise a piece of data. Clustering is concerned with intelligently grouping data and Regression relates to using historical data patterns to predict future ones.

How can this be Useful in Hospitality?

One of the biggest challenges in running a successful hospitality business is setting your rates and restrictions effectively to optimise occupancy and ultimately, revenue. Many businesses will use their own historic data as a guide, but this can often be a difficult and time-consuming task. In this case, a machine learning Regression model could be used to predict demand or identify optimal rates based on previous successes. Go one step further and you could use the data to predict not only the level of demand but where the demand is likely to come from, which could help refine your marketing campaigns.

As another example, categorisation could be used to profile guests and make predictions about their preferences. This could inform the right type of customer service or promotions that are offered. Certain guests may be more likely take the offer of champagne on arrival while others may go for a 2 for 1 lunch package.

These are just 2 examples but with the vast and diverse information collected in modern PMS systems, the possibilities are endless.

Machine learning is not a new concept and some companies are already offering ways to leverage it, but in hospitality, and particularly in small to medium businesses I believe we are just getting started.

At HOP, I am very excited to see how we may use machine learning in the future. Watch this space…

Author
Glyn Roberts,
Technical Architect at Hop Software

Leave a Reply

Your email address will not be published. Required fields are marked *