Crowdsourcing Doesn’t Work for Travel Writing

Via Metafilter, there’s an interesting travel piece in Slate. Tim Wu attempts to used WikiTravel to choose a rock-climbing location (and associated accommodation) in Thailand. He fails miserably:

My plan for this trip to Thailand was based on a somewhat corny vision: I wanted to climb a limestone cliff with my bare hands and look out at the ocean below, while pursuing a mind empty of stray thoughts. That meant I had three basic questions: Where can I find rock climbing in Thailand? What kind of people go there? And where can I stay? On all three questions, Wikitravel failed to deliver—in part because it’s still new but also because, ironically, Wikitravel fails to capitalize on the full potential of the Web.

Things go wrong because some of the central tenants of Wikipedia (and associated sites like WikiTravel) are neutrality and fairness. That’s exactly what I don’t want in my guide book or travel writer. I want to read an opinion on the best destinations, attractions and restaurants, not a consensus. WikiTravel apparently has nice things to say about all destinations, in the interest of fairness and neutrality.

This is why it’s such a joy to read travel writers like Bill Bryson of Bruce Chatwin, and why National Geographic Traveller magazine has pretty pictures, but is totally soulless.

In fact, Wu’s serious problems appear when he tries to choose a place to stay. Apparently, “it’s no fun visiting and ranking dozens of grimy hostels and boring hotels, especially when you’re on vacation.” I just checked the WikiTravel entry for Malta, and there are no accommodations listed at all. Clearly, WikiTravel needs more attention.

Thanks to Trey Ratcliff for the photo.

UPDATE: As per the comments, the accommodation listings aren’t associated with the main country page, but are linked off of each individual city page. For example, here’s where you would find lodgings in Malta’s main city of Valletta (listings are pretty slim there, anyway). That’s partially my bad, but seeing as both Tim Wu and I made the same mistake, there may be a usability issue there as well.


  1. I think you have to take into account that any one source can’t serve all purposes for all people. Wikitravel might be a good overview site, say for choosing Vietnam or Cambodia. When you actually get off the plane/boat, you then need a recent guide that tells you about lodging and dining in terms you can relate to (ie you pick a guide that fits your travel style).

    Wikitravel seems to lack several things to achieve the latter: critical mass of reviewers (nobody wants to add the first hotel for Bangkok because it’s a drop in the bucket) and opinions. I think you’re right that a plurality of opinions is better than neutrality. I would add that it would be good if they were qualified opinions (eg “quaint bungalows, but I don’t mind dirt floors and pit toilets”). It seems like the other Travelfish site that the author mentions has achieved both those goals for a smaller geographic area, mainly through paid content (is that bad?).

    What is needed are local experts, but then the question is whether they would be personally better off to contribute to the commons or make their own website/guidebook/whatever (as I have done). I’ve often wondered if I should duplicate my content to Wikitravel, but then I’d have to rewrite it and maintain it, and I don’t have that much free time.

    So what does it mean to say WikiTravel needs more attention? Seems like I’ve argued that in the market of online travel information, it might be a limited model, and that others might be more deserving. Not that I want to see it fail, but it seem like they haven’t fulfilled the needs of the travelling public.

  2. Just a quick correction: Wikitravel has extensive hotel listings for Bangkok– they’re just in the district articles! For example lists about 20 places.

    We also recognize that there’s more to travel than consensus guides, that’s why we’ve just launched Wikitravel extra ( We plan on having blogs, reviews, forums, the whole works, linked to Wikitravel guides.

    I hope you consider contributing to Wikitravel– feel free to add your info and then let the community edit and maintain it.

    Let me know if you have any questions.

    Co-founder, Wikitravel

  3. Michele: Thanks for coming by. As Andy observed, I do think it’s partially a question of critical mass. I also do think there’s something to individual voices and travel. I’ll check out

  4. Hey, Darren. So, firstly, I think it’s great that the current critiques of Wikitravel is that we’re not perfect… yet. A few years ago, people said it would never work. Now they’re impatient to see it all perfect! That’s a lot of progress in people’s expectations.

    I think there are some regrettable errors in Tim Wu’s article. As Maj mentioned, we have many, many hotel listings for Bangkok — Tim just didn’t look in the right place.

    And he also mixed up our “Be Fair” policy with Wikipedia’s “Neutral Point of View”, which was a really big mistake. We explicitly do _not_ require a neutral point of view; we want to have information that’s slanted towards the needs of travellers.

    We also encourage lively, descriptive writing that conveys an accurate description in an entertaining way. We know that “nice” descriptions don’t get us to our goals.

    One place I do agree we have a shortfall is in accommodation listings. Our main source of information is the first-hand knowledge a) residents in a city and b) visitors. Residents often go to local restaurants, bars, museums, etc., but they rarely stay in a hotel in their own home town.

    Similarly, a visitor on a 3-day trip to a city will go to 6-8 restaurants, a similar number of tourist sights, but will probably stay in only 1 hotel. So our growth in accommodations listings falls behind other kinds of listing.

    But we _do_ have accommodations listings, and lots of them. You don’t see listings on the Malta page, though, because we put accommodations listings in city articles; for countries, we just have an overview of the kinds of sleeping arrangements you can find (hotels, guesthouses, pensions, etc.).

    On the positive side, Wikitravel was just nominated for a Webby Award for Best Travel Site:

    I guess we’re not doing all that bad!

    One great thing about the Tim Wu article that Tim told me in private email is that when he got back home, he added what he had learned to Wikitravel pages, and that he plans to continue contributing. I think that’s the real vote of confidence.



  5. Just a quick note in relation to Andy K’s comment about “paid content” on Travelfish. All the content on the site is written by our team of paid researchers — it’s 100% original non-syndicated content. One of the issues with the wiki model (as mentioned by Evan above) is that volunteers may not be all that inclinded to go review 40 sets of beach bungalows during their trip — we get around that problem by paying our researchers to do it — and readers like Wu benefit from it.

  6. I’m late to this conversation, but anyhow.

    In general I support wikitravel, and the one idea I don’t like is that in Slate I was just beating up on wiki models or web travel sites as some kind of reactionary.

    Wikitravel, as it is, is pretty good, if not good enough to be a primary source in my opinion.

    Maybe wikitravel will, on its on, eventually crack the nut and reach critical mass.

    But my suspicion is that it needs somehow to get more opinionated, less like an encylopedia — for example, reverse the policy of telling people to leave things out if they just have something bad to say.

    I think travel writing needs to be lively to work, and the wiki format right now isn’t getting there quite. But it could I hope it will.

    Anyhow all of this is in the spirit of improvement, and I plan to continue contributing to wikitravel in the future.

  7. The problem with Wikitravel is that the wikimedia engine is the wrong model for a travel website. The wiki concept is great for organising unstructured data. Wikipedia proves that.

    Other Wikimedia Foundation projects, such as the Wiktionary, have used structured models and been successful. That’s because the data itself, words and their definitions, have structure. So far, the Wikimedia Foundation has wisely limited its projects to areas that it can support.

    Wikitravel doesn’t work because two kinds of data, structured and unstructured, must be integrated. The Wikimedia engine lacks structure, and doesn’t even support some of the data formats that would be needed, especially for time based and geographical data.

Comments are closed.

%d bloggers like this: