Catala consulting teamwork

4 Critical Mistakes Most Managers Make With Data Analytics

Background: Data Analytics

What a trendy topic to talk about data analytics! Big data here, data analytics there or even more machine learning and artificial intelligence. That have been the words that every manager must have heard in the past years. As social media were a few years ago, now firms in any industries need to embrace the world of data to stay competitive. Big data has transformed the world we live in and yes anyone should take advantage of it. But not everyone knows how to use it… Indeed most of the companies are sitting on a treasure of customer data, historical performance and other worthful of insights. Most of them do not know how to use it. Here are the four mistakes that most managers make with data.

Don’t Do The Following Mistakes With Data Analytics

Mistake 1: Not understanding the issues of integration

Compatibility and integration are the main challenges for any firms and can considerably reduce the quality of the data collected. Indeed data are coming from a variety of sources and cross-referencing the different sources can sometimes be complicated. Let’s take an example in the hospitality industry. Data may come from different sources: such as the PMS, POS, Website, GDS, Channel Manager, Social Media, Benchmarking tools, and many others. Some of these systems extracts work well together and it’s relatively easy to cross-reference each of the sources, but in many cases, it’s not that simple.

Data lakes” can solve this problem. But as data are very often unstructured, it is very difficult to store it in a structured way.

Mistake 2: Not realizing the limits of unstructured data

As seen previously, managers should look into making data valuable in its unstructured nature. Advanced analysis techniques and tools are now helping managers to mine text-based data, where context and technique can lead to insights similar to that of structured data, but some other forms are not that easily analysed such as video data.

Mistake 3: Assuming correlations mean something

“Eureka! We found a correlation into the data, then it must mean something!”, well not that quickly… It is easy to jump to conclusions quickly and move on with the next problem to solve. It is also extremely difficult to draw relationships within a large pool of overlapping observational data. Sometimes with not enough observations/data to accept the correlation as valid. Think twice and test again the correlations before jumping to conclusions. Wrong observations and correlations can mislead managers in their decisions making. This may be even more damaging than not using data analytics in the first place. We can find correlations in everything and anything. It’s our job to make sure we validate the correlations before anything else.

To make big data valuable, any manager should be able to move from observational correlations to identifying what correlations indicate a causal pattern. They should form the basis for strategic action. Identifying actionable insights from data it’s a tough job. If done correctly can lead to tremendous results and more accurate decisions.

How do you know if your ad online was successful or not? Let’s say a hotel advertises to potential guests across the Internet who have previously visited its website. Raw data would suggest that guests exposed to the ads are more likely to book a room with the hotel. However, the guests who have previously visited the website have already shown their interest in the specific hotel even before viewing the ad and are more likely than the average consumer to book. Was the ad effective?  We don’t really know… It doesn’t really allow a correlation and insight on marketing efficiency.

So how do we make sure we are in front of a real correlation or not? By experimentations! Test and learn on a random set of guests, with different factors, even with A/B testing. Once again once you have found the correct correlations, the hardest job (and most important) needs to be done… Draw insights and take actions to increase performance and efficiency!

Mistake 4: Underestimating the labour skills needed

The ability to create better algorithms to deal with big data is another very valuable skills for managers. That’s where recommender systems come into play! Recommender systems are systems that rely on algorithms trained on correlational data to recommend the most relevant products to a customer. For example: when a guest is browsing through, a recommender system will present to different potential guests, different choices of hotels based on data collected online. It allows to predict customer behaviour and increase the chance of sales conversion. The quality of the system’s algorithm depends on the quality (and quantity) of the data scientists working on it. Managers who assume large volumes of data can be translated into insights without hiring employees who have the ability to trace causal effects in that data are highly mistaken…

Big data itself is unlikely to be valuable. It is only when combined with managerial, engineering and analytic skills that becoming relevant and worthful. Processing skills (understanding and drawing actionable insights) are definitely more important than the data itself. It creates value for any businesses.

Sources: (Anya Lambrecht & Catherine Tucker; 2015)

Need Help With Your Forecast And Data Analytics?

Need help with your Hotel Budget 2019 or forecasting your future revenue? Contact us today and let’s schedule a call to see how Catala Consulting can help you optimize your revenue performance.

Don’t forget to follow us on Facebook and Linkedin!

Scroll to Top


We’re sharing our checklist on Hotel Website Optimization. It’s a completely free guide which is packed with top strategies that we use to drive more direct bookings for our clients. Claim your copy today.