Sunday, March 08, 2009

Travel Discovery, Triporati and the music business

Through post commentators and email exchanges with readers I have been thinking about whether or not we should be classifying all the travel content, planning, community and search sites that have emerged in the last year. Classification will help to identify competitor sets, clarify business models and help with predictions as to who will be the winners and losers. Also we need something to differentiate all the companies that have launched since 2006 with the word 'trip' in their name.

The first category I have decided to turn my attention to is "Travel Discovery & Inspiration". These are companies that help with the very first part of trip planning - coming up with the inspiration for where you want to go and what you want to do. That help the potential traveller narrow down a world of opportunities and possibilities into a basket of ideas to be explored and researched further. Another reason I want to start with this category is that I have been thinking about the general area of web supported discovery for some time now.

At WebItTravel 2008 in Singapore last October, Ram
Badrinathan of PhoCusWright asked me to name my three favourite start ups. One of those I highlighted was not a travel company (and is not even a start-up any more). I talked to him about music social network and discovery site last.fm. Last.fm is the best product I know for discovering music. It tracks the music you listen to, then looks around for other last.fm users that listen to the same music. Then it recommends tracks to you that people listen to who like the music you like. In effect it crowd sources music recommendations based on the similarity of your music tastes with others in the network. A great manifestation of this is your ability to listen to a 'neighbours' radio station. A neighbour being someone with similar tastes to yours and their radio station being a collection of their favourite songs.

Last.fm's chief rival, Pandora, has the same aim - helping you discover new music- but instead of using crowd recommendations like last.fm Pandora has teams devoted to the genomics of music (Music Genome Project). That is breaking down a song or artist into the elements or themes ('genes') and matching to artists or songs with similar genomics.

While they approach it in different ways the concept is the same - bringing to the web and technology the power of word of mouth and trusted advice as a tool in pre-purchase discovery.

The applications to travel are clear. Helping consumers to answer questions of "where to go next?" and "help me find somewhere to go" through networking with other consumers or expert fed technology based query engines.

Triporati is a company that has really impressed me in their efforts to undertake a Travel Genome Project and build a query engine for recommending travel destinations. I first came across Triporati at PhoCusWright 2008 in LA where they participated in the Travel Innovation Summit. They made the short list of six (out of thirty two) at that conference as well as being one of my picks for a top six spot.

Triporati was launched by online travel industry founding fathers/mothers Jim Hornthal (Chairman) and Sharlene Wang (Chief Product Officer). I call them that as they were the builders of Preview Travel, who's sale to Travelocity in March 2000 (announced in Oct 1999, closed in Mar 200o) marked the beginning of online travel as a serious economic force (and temporarily consolidated Travelocity's early lead in online travel). Like Pandora did with music Hornthal and Wang have drawn from travel writers and experts to identify 62 elements of choosing a destination. A user selects (and ranks) up to ten of the elements that interest them
and some other data (like home airport and number of travellers). Triporati recommends destination options. For example I chose a number of beach, swimming and snorkelling themes. Recommended for AsiaPac were Fiji, Tahiti and Queensland. For Europe Gran Canaria, Catalonia and the Italian Lakes Region. None of this is surprising but then I know the areas well and generating recommendations on sea, sun and sand is not that challenging. But in regions and search combinations that I am less familiar with I was presented with destinations and travel ideas that were new to me and intriguing. For example, selecting "Wine Tasting", "Zoo" and "Foreign Languages" I was presented with the Cuyo region in Argentina - near the border with Chile - which sounds amazing.

I have been trying to find others in the content/planning model that have followed this Travel Discovery & Inspiration path in using destination idea generation as the means for taking travellers down the trip planning (and therefore eyeball monetisation) path. There are plenty of sites using combinations of editorial and user generated content to provide advice and recommendations on what to do in a (known) destination but I have yet to come across another like Triporati which recommends destinations based on broad traveller . I did come across want2bethere.com in an email exchange last year and in 2007. They claimed to be working on technology that allowed a customer to outline the requirements they were looking for in a trip (through drag and drop), which would then be matched to recommended destinations. Unfortunately their website now seems to be down.

What do you think of my first efforts at classification? Do you know of other companies building discovery engines like Triporati (and last.fm/Pandora in the music world)?

FYI is an interview with Triporati Chairman
Jim Hornthal at PhoCusWright last November.


18 comments:

Anonymous said...

I agree that "Discovery and Inspiration" is the first step when planning a trip, I would like to suggest though that recommendation engines in the last.fm sense are only one way to address this phase. I think there are plenty of other ways how people get inspired, ie reading travel articles and blogs, looking at pictures, talking to friends.

Basically all websites that focus on these activities would fit in your category. HOWEVER I dont think this would include the hundreds of "user generated travel guide littered with ads" websites as often they completely lack emotional appeal. And in my view this is a key aspect of anything that wants to inspire.

So I would argue, for "Discovery and Inspiration" it is essential that the site has a certain emotional component which gets me excited about travelling. Example: lonelyplanet.com does not have a recommendation engine but still it is a very inspiring website, and in my view still #1 in this area.

Anonymous said...

A site I like that offers "Discovery and Inspiration" with a added travel community is a site called www.nozio.com

Anonymous said...

to break the trend of trip named sites - with a bit of tongue in cheek, ive just registered twip.com.au (might be angle for twitter related trip site ;-)

Tim Hughes said...

@Markus - good point. I agree that everything from a newspaper article to a postcard to a lonely planet website is a cause for discovery and inspiration. I agree with your comment about an emotional component. I would also submit that it needs matching process. This is very hard. If I ask my best friend of 25 years, he knows what I like and want I dont like and can match his recommendations to that experience. Replicating this in technology - even UGC supported technology is very challenging (as I am sure you know).

@Chris - thanks will check it out. At first glance it has interesting content and information but only if I select the destination I am interested in. What I am hoping to explore is websites that recommend the destination/activity that a consumer should do.

@Steve S - nice. Twery twery nice.

Anonymous said...

I wonder if there's a small step beforehand, what 'triggers' the desire to actually travel, an article, a photo, a flash of a doco, a brag from a friend?

I'm fascinated by this and we're currently exploring.

When it comes to 'Discovery and Inspiration' I think as time goes on, content ( and tagged content) will become more readily available to sites, therefore making the challenge to 'engage through design' even more of a priority.

What user experience will bring you back as a traveller and what will accommodate your changing travel desires over time?

Anonymous said...

Great topic! Can I put ourselves in the "travel-website-without-'trip'-in-our-name" category? :p

At GeckoGo.com, we definitely recognise that there are a number of inspiration trigger points for travel planning including stories, images and videos. These can evoke strong emotional reactions and interest to a place.

I believe though that interest-based recommendations formed by other people like you is one of the most powerful travel discovery and planning starting points, and that tips, stories and images are a great way to supplement and reinforce trip ideas planted. We've gone and put our recommendation tool front and centre on our homepage. You can get recommendations of the top places to visit from our travellers based on your interests and the time of year. These recommendations are created by our members (as is all our content right now), and our plan is to further personalise recommendations for you over time based on your travel personality and profile. Hopefully we'll get there! :)

From what I've seen, Travel Muse and Nile Guide also seem travel inspiration based. :)

Anonymous said...

@Tim, completely agree there is a matching component. The question is, how much of this matching can we automate? Basically, as you say, it is going to be very hard unless there is huge amount of data feeding the algorithms (as in case of last.fm).
On the other hand, humans have excellent "pattern-matching" capability, so as long as the algorithms dont work, I think its perfectly fine to simply present well-structured information and give the user good tools to properly filter it.

@Pokin, I really like the time component in your recommendations. Maybe missing a few more categories, ie what about "Family-friendly" or "Budget". Are these suggestions all user-generated? Impressive

Anonymous said...

Hi Tim,

Good post.
I agree partially with your analogy to Last.fm (which I love and use also)
Unlike travel, music can be evaluated by a participant in a matter of minutes.

I wanted to bring Tripbase.com to your attention.

Tripbase launched its travel recommendation engine in May 2007 and has since been a leader in this area.

We actually hold a registered trademark on the phrase "Travel Genome Project" and as the project matures and develops, more results will become publicly available at travelgenomeproject.com. We currently have over 20,000 destinations in our genome with more being added daily. Data points per destination range from around 50 to just under 500 depending on the destination.

Our approach differs from Triporati's in several ways.

Our methodology to create the Travel Genome Project includes the following:
1) Data mining and natural language processing technology to process hundreds of thousands of pages of content.
2) Experts that manually validate and confirm the automated results and add their own input.
3) Analysis of travel communities to process travel trends and interests.

Most of the data required to give good recommendations already exists on the internet. Our main goal is to process it and create a comprehensive genome.

My background is in computational biology, BioInformatics, biological networks and artificial intelligence.

A lot can be learned from BioInformatics, in which large amounts of existing data are processed to create a clear understanding of biological structures and interactions. This field has become known as BioInformatics or Computational Biology. As time goes on, I am sure we will see many more players in the TravelInfromatics, or shall we say WebInformatics, space.

Tripbase.com is still in Beta and only utilizes a small fraction of the data we currently have.

Sincerely,
Reuven Levitt
CEO Tripbase.com

Tim Hughes said...

@worldnomads agreed. Good point

@Pokin - I have been thinking about travelmuse and NileGuide. Could be that I am creating distinctions where there should be none but I see those sites as still requiring the customer to know where they want to go. Having already decided on the city they are going to. Will definitely form part of the second category.

@Pokin and @Reuven - will give GoGecko and Tripbase a look

Anonymous said...

Great post. I think it is worth classifying different points in the purchase funnel where inspiration plays its part.

(1) 'blank page' inspiration (what you are concentrating on) is probably mostly picked up sub-consciously from browsing travel press, reading books, watching films and TV and talking to friends. I would argue that people are rarely actively in this space, they will usually have some broad pre-existing categories in their head - a country or broad activity they are interested in.

(2) Inspiration when you know you are planning a trip to a rough area: Once you are at this stage, god travel writing, UGC, images and videos can all play a part as mentioned above come into play.

(3) Inspiration when you are looking for specific products. When looking for particular hotels, hostels or tours, inspiration is still a critical tool. At this stage quality of media (pictures, descriptions) becomes critical to stand out and grab the readers attention.

It strikes me that the majority of effort and online innovation at the moment is in the second category and there is surprisingly little focus on the third category which tends to be commodotised to a greater extent online. This is where we are trying to focus with Tourdust. I have posted twice recently on my blog about this topic (http://trailbeater.blogspot.com/2009/03/travel-inspiration-seeing-watcing-or.html) and would welcome any comments. I know your focus is on the first, and it would be fantastic if it could be cracked, but at the moment I would argue it is still at the marketing gimmick stage. i.e. I'll try it once, think "oo isn't that clever" and then not be compelled to return.

Tim Hughes said...

@Ben C thanks for the comment - will give tourdust a look also. I think you have a very good point about the longevity/stickiness of the "blank page" inspiration sites. With an OTA is makes sense for a customer to come back again and again as they buy more and more travel. Same with a content or community site where information and friendship connection keeps a user coming back. But with a Triporati style Discovery & Inspiration site, they have a dual marketing challenging - getting the customer to find and use the site in the first place, then getting them to come back and use it more than once.

Anonymous said...

I always felt that I should be paid for contributing to these trip planning sites. Some do offer rewards and therefore in my opinion are better.

Interesting to note the TOS flare-up with Facebook when many UGC content sites are basically doing the same thing with info that can be quite personal as well...to wit the content TOS for a recently viewed trip planning site (name removed):

By posting Content on the x Site, you agree to and hereby do grant, and you represent and warrant that you have the right to grant, x, its contractors, and the users of the x Site an irrevocable, perpetual, non-exclusive (including any moral rights), royalty-free, fully sublicensable, fully paid up, worldwide license to use, reproduce, modify, delete in its entirety, adapt, publish, translate, create derivative works from and/or sell and/or distribute such materials and/or incorporate such materials into any form, medium or technology without compensation to you, as outlined in x's Copyright Agreement. You also permit anyone to access, display, view, store and reproduce such Content for personal use.

Anonymous said...

Tim,

Excellent discussion on the topic of the 95% that happens in travel before the 5% of the actual booking which until recently has been the near 100% focus of the industry and its pundits.

The Dream - Learn - Plan - Go process that anyone goes through in leisure travel has for many years been the focus of DMOs as their main objective is to get on the travelers radar screen as early as possible in the thought process and stay there. For the online travel players this was not nearly as important, as long as they could complete the transaction. Now, this seems to change.

It will be interesting to follow the development of the new entrants in this area of the industry and the solutions they offer to the frustration of online travelers with the transaction focused approach still evident on the major sites.

The idea of an intelligent agent presenting relevant choices based on personal preferences will be a compelling one and one that will transform the industry in this next phase of development. Of interest as well will be the role the semantic web will play in this process.

A key to success will be the integration of the path from beginning to booking and post travel to offer a truly user friendly overall experience rather than a multiple hand-off among partner sites with the frustration inherent in that model.

Anonymous said...

Hi all,
great post and discussion.

The problem is that if a site focuses on destination inspiration without any connection to product at some stage a user has to leave, change sites and repeat much of the search process over again...frustrating?

Tourabout (a non trip name :-) started at the product end in 2006 developing systems to mine partners sites, load and process product and index attributes to make product searchable (structured and natural lang). An benefit we found in starting with product is you also end up with most of the dataset of the attributes for the destinations relating to those products. At 30,000+ trips to 20,000+ destinations, currently it's a simple Qualifying Engine for travellers and new business referral/enquiry gathering operation for partners.

We see the key to success in the next cycle being vested in a combination of Sramana Mitra’s definition (Web 3.0 = (4C + P + VS) + Place), a semantic application with product, destination and crowd sourced content and an open Marketplace connecting travellers, agents and suppliers for an integrated inspire, plan and book experience.

Aside: Another benefit of a focus on product and search functionality is you can deliver B2B systems which benefit revenue(now licensed by over 4000 travel agents).

Anonymous said...

Great post Tim. I have been looking forward to this (and the ones to follow) since https://www.blogger.com/comment.g?blogID=29875741&postID=4691306197902263531.

I think you are spot on with the first classification.

Before reading your post, I would have placed travel sites into the following classifications:
1) Research and inspiration (decide where you want to go and what sort of things you want to do when you get there)
2) Planning (create a details travel itinerary with prices, maps, directions etc)
3) Booking (Actually book the travel components online)
4) Social networking (Keeping a journal, interacting with other members etc)

But I think you are absolutely correct that the first classification can be further broken down into:
a) Travel Discovery (Actually find a number of destinations which seem appealing and suit your requirements)
b) Research (Research a destination to see what it has to offer, with guides, photos, videos, UGC etc)

Obviously some websites will fall under multiple classifications (e.g. ePlanner.com currently falls under 2,3 above and will soon also fall under 4 and 1b, with 1a following at a later date to provide a total one-stop travel site).

Tim Hughes said...

@Anon - in the early days it was enough to compensate UGC reviewers with kudos and profile ranking. I agree that this will not be enough going forward. The really good contributors are going to need more than that. Maybe it is rev share on PVs, maybe it is outright payment, maybe a discount on travel.

@Joe and @ CarlJ - good point on the potential need for integration between the dream/discovery part and the booking part. Raises the very interesting question as to whether or not the dream/discovery people can be stand alone or have to be part of a booking business.

@Jonathan - thanks and snaps for planting the seed on this area. More to come!

@Martinha - thanks for the (almost) spam

Jonas said...

To bring up another point, what personally do you guys think is worse, false positives (as happen with a lot of the fully automatic semantic search engines) or not enough coverage?

At our site: PlanetAbroad through our technological approach we try hard to avoid false positives (e.g. recommending a beach when there is no beach because some text somewhere said "don't go there for the BEACH") but currently are still in our infancy in terms of coverage due to that.

What do you guys think?

Tim Hughes said...

@Jonas - thanks for comment. Very good question. Here are my thoughts

- I think consumers come at search (whatever they are looking for) hoping for the "right answer". Not thinking about the balance btw false positives and depth. If there is a false positive in spot one but the right answer in spot two then they will be forgiving of the false positive. Similarly if the are only 3 responses but 2 of them are spot on then consumers will forgive the limited results. I remember reading alot of the commentary about Cuil's failed launch was around Cuil not producing the "right answer" not around balance between false and depth.

If I had to choose - getting rid of false would get you closer to the right answer than depth