The Implications of a Correlation
By Damiano De Marchi, Destinations Expert, The Data Appeal Company
Over the past decade, we’ve added a new appendix to our bodies: smartphones.
Smartphones have become such an integrated aspect of our anatomical structure that there are even osteoarticular side effects as a result of overuse. These ailments include wrist and thumb tendonitis and text-neck, a painful numbness in the neck and back linked to having the wrong posture while looking at the screen.
There are also those who have gone a step further, by conducting a scientific study to prove that the excessive use of smartphones is the real cause of a bony protuberance in children at the base of their skull.
A sound mind in a healthy body – Greek Proverb
Fortunately, the pervasiveness of mobile devices in our daily lives encompasses opportunities much less dramatic than those just described. Smartphones are becoming the basis for new schemes and reasoning. A prime example is the push of internet sites, primarily Google, continually asking for opinions on places recently visited. “Describe your experience at the XYZ hotel?”, “Did you enjoy ABC restaurant?”, etc.
This phenomenon has exponentially increased the amount and quality of Digital Human Experience data, or all the contents – reviews, evaluations, comments, social posts – that collect fragments of customer experiences, purchases and relationships.
Although the process of monitoring, processing and analyzing these contents is continuously being improved and updated, there are very few companies that have the vocation ingrained in their DNA to create an instrument for interpreting reality and consequently allowing for businesses to strategize effectively and drive success.
“Today, with a plethora of options at every corner (physically and virtually), it’s crucial for businesses, of any industry, to focus on improving their perception, reputation and appeal.”
Are customer reviews a precise and reliable reflection of reality?
To answer this question I put on a white coat, protective goggles and channeled my inner chemist. I created experiments by combining digital human experience data with official visitor flow data.
We start with the region of Veneto, home to the beloved yet controversially overcrowded Venice. It’s a notorious destination and perhaps the first tourist region of Italy.
We superimpose the monthly distribution of 19.6 million official arrivals per month (the red line) with 1.8 million contents monitored by Travel Appeal (the black line). The graphic representation, though based on different scales, is noteworthy, despite the gap between the visitor’s experience and when they leave their review. The correlation is very strong with a correlation coefficient, or Pearson’s r, equal to 0.94.
If ISTAT (The Italian National Institute of Statistics) detects 7,146 official properties (of which 2,975 are hotels) and Travel Appeal monitors 7,414 properties with an active online presence (3,040 hotels, while the rest are non-hotels), how is it possible that there are more businesses listed online than those actually registered?
The Implications of a Correlation
There’s also a strong correlation in visitor origin. Travel Appeal data analysis reveals that 42.6% of online content was posted by Italians, while ISTAT reports 42.8% official Italian arrivals.
Moreover for accommodations, the comparison remains. Travel Appeal detects 1,351 active properties online while ISTAT reports 1,569 official properties.
Monterosso al Mare
Monterosso al Mare is one of the municipalities that make up the Cinque Terre. In 2018, there were just over 100 thousand total arrivals, with 94% concentrated between April and October. When comparing ISTAT data with that of Travel Appeal, the correlation is again evident and strong, with a Pearson coefficient of 0.96. There are 31 accommodation properties active online compared to 36 official ones.
Agatha Christie argued, “one coincidence is just a coincidence, two coincidences are a clue, three coincidences are a proof”.
From the first results, it’s no longer a taboo to think that online content, which transpires from the relationship between people and places, is a strong indicator of the correlation with reality. Human experience data provides insights much faster than data from national institutions.
The experiments will certainly continue, especially as destinations strategize to stay ahead of the game and understand their territory more effectively. The implications for management, marketing and forecasting is innumerable, as this data goes well beyond the borders of the tourism sector.
The Travel Barometer Index
To seamlessly monitor and evaluate the wellbeing of your destination, Data Appeal has developed the Travel Barometer. This index combines quantitative and qualitative elements of a destination’s reputation and online presence and analyzes content, sentiment scores and the digital presence of tour operators in the area. With this index, compare your performance against previous years or different time periods, against the national average or any destination of your choice.
Which signals come from traveler’s conversations about their willingness and confidence to resume travel? How is your situation compared to competing destinations?
Define and evaluate your tourism strategies with the latest data-driven insights.
Methodology & Notes
- Three (3) territorial areas have been chosen, with increasing dimensions (municipality, province, region), where Travel Appeal has monitored and analyzed the data in the reference period.
- Accommodation includes all types of hospitality accommodations excluding rental properties. For Travel Appeal, businesses active online are those that have recorded at least one content (i.e. a review) published in the time period considered.
- Seasonality refers to the comparison that takes place between the total registered tourist arrivals and the contents monitored and analyzed by Travel Appeal.
- The Pearson linear “r”, also known as the correlation coefficient, assumes values between -1 and 1. The closer the index is to 1, the more the variables are directly correlated (when r > 0.7, there is a strong correlation).
