I could summarize this article by saying TripAdvisor sucks and leave it at that, but I’m going to try to illustrate a broader concept with it instead.
I’ve traveled pretty extensively in various parts of the world and one thing I never do is rely on TripAdvisor to find places to eat or things to do. In case you’re not familiar, TripAdvisor has become almost the de facto standard app/website/whatever for travelers to leave reviews and find places. It’s like the Yelp of travel, and businesses have a serious love-hate relationship with it.
On the one hand lots of positive TripAdvisor reviews can cause business to boom. It’s a self-propagating cycle where the more good reviews they get the more people come and so on. But on the other hand, not having strong reviews or not having a presence at all can mean you’re relegated to struggling for business and may not stay in business long at all.
At face value this sounds fine. Good businesses get good reviews and the cream rises to the top. And in a perfectly rational world that’s what would happen and everything would be great. But that’s not what happens.
To understand what goes wrong we have to back out and look at a broader concept. In Seeing Like a State James C. Scott puts forth the argument that whenever you reduce a complex system to just a few parts, and make those parts prominently visible, or legible, you effectively eliminate the non-legible parts.
This is illustrated in his book by taking a historical look at early Prussian scientific forestry practices. The first to take a scientific or industrialized approach to forestry, the Prussians simplified forest management down to what they deduced were the vitally important parts. They switched to a monoculture of the Norway Spruce, prized for its efficiency in utility value as a commodity, planted in perfect rows, culled out underbrush and pests that didn’t contribute meaningfully to productive output, and mapped and surveyed only and exactly what was important to the forester: where the trees were and little to nothing else.
The results, like the effectiveness of letting the “wisdom of the crowds” sort winners from losers in business, were exceedingly positive in the short run. The first generation of scientifically managed forests were incredibly productive and profitable. A win for science and a boon for the economy.
Trouble was, the next generation was a disaster. Subsequent generations were so bad that the Germans had to invent a new term for the death of forests: Waldsterben.
Turns out all of those annoying factors that were simplified out of the equation actually mattered, and not in a small way. Removing underbrush, pests, grazing animals, and switching to a monocrop were just a few of the optimizations that massively disrupted the soil ecology rendering it barely able to support a forest.
Which brings us back to the legibility that TripAdvisor reviews create of businesses. Far from capturing the “wisdom of the crowd” the reality is that what is made legible what satisfies the lowest common denominator.
I won’t take restaurant or food recommendations from nine-tenths of my friends. Sorry, but most just don’t have the depth or breadth of experience with great food to be able to tell the difference between acceptable and extraordinary, and I’m only interested in eating at extraordinary restaurants (in food or experience, not in cost – there’s a difference). So to know the average of what nine-tenths of my friends think tells me nothing I want to know. What I really want to know is what that one person I really trust thinks, but the average won’t tell me that.
What is made legible instead is a mix of what the predominant audience thinks is important. In Italy for example, Germany makes up the biggest group of tourists followed by Americans. Americans are notorious for complaining about having to pay to use a restroom (which is actually less common in Italy than it was a couple decades ago, but that’s neither here nor there). So there’s a high likelihood that one of the factors that is going to influence a review of a place is whether or not the bathroom is free. This kind of information might be useful to the ignorant looking for a free bathroom, but it’s useless to me looking for the best experiences and happy to pay a few cents to use a bathroom that will inevitably be many times cleaner than a free one.
Once this kind of information is the only thing that matters to the gatekeepers like TripAdvisor, they become the only thing that a business focuses on. People give high reviews for menus in English, so the restaurateur has to spend their energy creating menu translations that could have gone into creative dishes or spending more time picking what was fresh at the market. Hyper-local traditional menu items give way to more familiar favorites. And so on.
Instead of the cream rising to the top, the map becomes the territory and the territory becomes homogenized.
Instead, just as in all the other domains where selective legibility from high above a complex system is woefully ineffective, I rely on local knowledge.
I’ll find a local resident, preferably one not directly in the hospitality industry who may have vested interests in steering people a certain way (such as the hotel desk or concierge) and ask them where they would go, or take their family to eat for a good meal. Not where they think tourists should go, but where they’d go.
And while it’s not a perfect system, the results are far better than I could ever get from relying on average ratings from the masses. Asking someone local surfaces all kinds of complex knowledge that would be impossible to capture in an app rating system. Maybe the best restaurant around has better food but the owner is a total dick and a cancer in the community – I don’t want to support that, and a local is probably going to take that into account without even thinking twice about it.
In short, TripAdvisor sucks. But reducing complex systems to over-simplified parts to make them legible not only sucks, but it destroys the very systems they make legible.
It’s your turn to think about this. In what ways have you seen something complex and nuanced reduced to an oversimplification, and what harm has that caused?