Working in the digital travel industry for a long time, I know that there is one major issue. Each hotel and every destination would love to solve this challenge in order to distribute campaign budgets over the year: Have a good prediction of the occupancy rate of hotels within the destination. This could be used as a measurement for the necessity of starting digital campaigns to generate bookings. As this has not been done yet, we are currently working on a new kind of algorithm to find a solution for this travel industry challenge. Moreover, it looks like we are on a very good track to solve it.
You might have read my recent blog article about the digital transformation of the Swiss ski destination Saas-Fee. This is where it all began with our idea. When we set up the so-called Smart Marketing Engine for the Destination Marketing Organization (DMO) of Saastal/Saas-Fee, we had a brilliant platform but one important thing was missing. We did not know the occupancy rate of hotels for the next 4-8 weeks.
We wanted to start all future digital campaigns for the destination based on the occupancy rate of the destination’s hotels; therefore, we needed to know what would be going on the next two months. First, we thought about asking the hotels but that would not work out as you can imagine (never rely on people doing something for you repeatedly). We would have to…
talk – or at least write – to them daily to get actual data
set up a scheme where they can enter their data
receive an honest answer
hope that they have the data available and
be sure that these data are true at all
In addition, there are many more problems attached to this solution, the biggest one about how to generate all the data necessary to make good predictions.
So what else could we do? We thought about generating the data based on predictions, combining data from earlier years with weather forecasts, sold tickets and many other things. Moreover, when we looked at it in detail, we found out that there is a big chance to predict the short-term utilization of a destination at least to a relatively high degree.
What kind of algorithm are we developing in this project?
This is easy; we want to know the occupancy rate of a destination in the near future (4-8 weeks). Based on that utilization forecast we would like an algorithm to decide – when there is a lower actual occupancy rate than targeted before – if we can make up the difference target/actual by starting a digital campaign. To answer this question, the algorithm has all datasets and KPIs of our past digital campaigns integrated and by this knows exactly how much a campaign to generate a number of x leads in a certain channel y will cost. On this data, a decision can be made very quickly.
Our goal is to have the following fully automated process in place:
With the solution we are working on right now, the whole destination could pro-actively react on their occupancy rate. If this is below a certain level, the system will recognise any patterns and suggest a campaign including the channels.
Furthermore, to use this system to support the destinations service partner, they can put their budget on top of the DMO’s and by this we have made two points: Firstly, helping the partners reduce their digital marketing budgets. Secondly, by making our total digital campaign budget bigger with all partners on board we have much more efficiency within each campaign (economies-of-scale).
This project is a running for further 12 month, supported by the Hochschule Lucerne (Switzerland) and the Swiss KTI program.
If you need more information, please drop me an email and I will keep you updated.
Any feedback and comments are warmly welcome