Can the tourism industry create the perfect algorithm? One that precisely predicts where a traveler wants to go and what she wants to do when she gets there? Asked to consider this question recently on behalf of Les Entretiens de Vixouze , an annual conference on the future of tourism, my short answer is: Yes. But should it?
Predictive analytics in tourism are animated by “recommender systems” which use either the prior online behavior of an individual or a group of individuals who are like each other to ‘predict’ what we would like. Netflix, Amazon and TripAdvisor are examples of companies that use the traces of our digital behavior to predict what else we will like. Leaving aside costs and ethics for the moment, there is no theoretical barrier to more and more nuanced personalization as we all leave more and more digital traces behind. If I have already shared a preference for medium-firm pillows in my hotel room or beach vacations, a recommender system can call out more of the same.
Tourism recommender sites could help expose travelers to experiences they don’t know they want … yet
Given the clear advances in personalization, the more interesting question may be: Can tourism help potential travelers uncover or even develop desires for new experiences, places and tastes that do not emerge from previous desires and experiences?
Tourism occupies a fascinating space on the continuum of consumption and desire. Notwithstanding tourism built around needs such as medical interventions and business conferences, the ur-formula of tourism is the promise to take a traveler to the opposite shore of his current experience, as a way to help him discover a kind of true but as yet undiscovered self. Those who are working too hard seek relaxation; those whose everyday is drab seek luxury; those who feel unstimulated seek learning and adventure; those uncertain that the world is just may seek opportunities for eco-tourism or volunteerism that helps build a better one.
At its best, tourism exposes travelers to novel places, ideas and cultures and diverse experiences. This suggests a role for a few new algorithmic models. At their simplest, they can randomize traveler options, throwing out unexpected wild cards — mountain options for those who already like the beach. More complex formulations of introducing diversity are also theoretically possible.
Potential role of tourism in bursting filter bubbles
This raises the possibility that tourism, and perhaps other industries that supply us with services and experiences (such as education and entertainment), may play a civic role helping users burst their ‘filter bubbles.’
Many of us in the United States have learned about ‘filter bubbles’ as a result of discovering how the algorithms used by search engines and social media platforms—Google, Facebook and Twitter, mostly— shaped opinions leading up to the 2016 presidential election.
The more tightly that recommender systems bind us to either our own previous selections of news content (or that of others like us), the more tightly we are encased in a ‘bubble’ that isolates us in a solipsistic chambes of our own biases and preferences. (For a review of the phenomenon, Maria Popova’s interview with Eli Peliser, author of The Filter Bubble—makes quick work of the issue).
Those who study this phenomenon have proposed that media recommender systems incorporate “diversity-sensitive design” that use profiles to increase readers’ exposure to more varied information. The benefits and incentives for democratic systems to want users to have wider exposure to more information are fairly clear, even though it turns out to be difficult to arrive at consensus of what constitutes “diverse exposure” looks like.
Yet as individuals begin to live more digitally integrated lives, it may become important to explore whether there are additional industries beyond media, search engines and social media, that can or should play a role in widening consumer access to information, learning, options, and new desires and experiences. Industries such as tourism, education, entertainment and other services that seek to capture open ended desires come to mind especially.
Governments should be interested in this question on behalf of the public interest, and individuals should be interested both as citizens and consumers that our choice and autonomy is being more optimized, and less constrained.
Incentivizing diversity-sensitive design
The incentives for industries themselves may be more mixed or conflicted, as the goals of civic discourse don’t always neatly align with profit.
But at least a few options suggest themselves:
- There is an alignment between the liberal ideal of consumer and citizen, such that maximizing the choice of the consumer to select as much diversity exposure as they seek will also serve political values. As Natali Helberger, professor of Law at the Institute for Information Law at the University of Amsterdam, and her colleagues argued in a recent article in the journal Information, Communication & Society:
Exposure diversity can be defended from the perspective of the traditional liberal ideals of individual autonomy and consumer choice. From the liberal-individualist perspective, diverse exposure can be valued simply because it extends individual choice and affords individuals more opportunities to realize their interests.
From this perspective, as Helberger goes on to suggest, “The preferred level of exposure diversity is therefore likely to vary between people. Even if algorithmic power is at least partly the result of people’s own autonomous choices, there are likely to be consumers who would actually prefer a more diverse fare than what current algorithms recommend.”
Tourism recommender systems could ask travelers to indicate themselves how much they might like to be exposed to new sources of travel.
- Corporate Social Responsibility. It is not impossible to imagine that in a generation or so, the content and impact of algorithms themselves will become an issue for civil rights interests, at which point citizen-activists could make algorithmic transparency and exposure diversity matters of corporate responsibility and shaming.
- Diverse ideas and profits reside in the “long tail” of the market. Tourism providers can engineer (or co-create with users) productive convergences that bring both new experiences and new forms of knowledge to market. The Long Tail theory authored by Chris Anderson of Wired magazine a decade ago, is that “our culture and economics is increasingly shifting away from focus on a relatively small number of ‘hits’ at … the head of a demand curve and toward a huge number of niches in the tail.”
The theory was originally used to describe how, “when consumers are offered infinite choice, the true shape of demand is revealed” to be much more diverse than once thought. Online stores, such as Amazon, made it possible to offer that kind of choice.
Law professor Helberger extends the concept to theorize a “long tail” of political viewpoints, suggesting that one form of virtuous exposure diversity lies in “promoting the visibility of minority and controversial viewpoints, under the assumption that these are indeed located in the long tail of content popularity.”
The tourism industry has its own version of long tail market. Like the book business, it has huge blockbusters in the form cities and sites that are so overwhelmed by visitors that the situation creates serious spin-off problems for residents and tourists alike. Venice has been so overwhelmed in recent years that the government may limit the number of tourists who can enter the city.
But tourism’s “tail” is also long. There are many smaller and off-the-beaten-track areas that would benefit greatly from increased tourism. The need is so great that the Italian government will give you an Italian castle for free, if you agree to develop it for international tourism.
Recommender systems will have a part to play in creating exposure to the “long tail” of destinations. Along the way, they could create algorithms that do far more than customize, but that introduce travelers to new desires, opportunities, ideas and people.