The people at Revinate have been experimenting with Term Frequency Analysis.
A bit of a mouthful but it simply means the company takes an online review and counts how often each word appears in it.
It also records all the metadata around the review such as review score and sentiment so the data can be further segmented by that metadata.
It all boils down to is an interesting way to gauge trends across multiple properties and markets as well as get insight into the language people use when they love a hotel versus the terminology when they have a bad experience.
The process is then repeated for millions of reviews to see how popular a word and whether it is more likely to appear in a positive or negative review.
To demonstrate, Revinate has taken 2,500 of the most popular words across five million reviews from all over the world and enabled them to be plotted on a graph.
Results are split by review score and how popular a word is for that review score so one is the most popular and 4,000 the least popular.
So, you can see that 'best' is popular across all review scores while peaceful becomes more popular as the review scores increase.
Here's another example from Revinate - four terms associated with a negative experience:
Click here to put your own words into the graph using the auto-complete feature.
Revinate is now putting the science into practice to help properties compare themselves with competitors and answer questions including whether there is a type of guest staying within its market that it is not catering for and what guests expect from its property compared to a rival hotel.
Further examples can be found on the Revinate blog as well as some observations from its director of data science, Conor Nash.
NB: Love/hate image via Shutterstock