Kayak CEO Steve Hafner Interview: keep it simple, focus on search, stay out of Asia..oh...and get back into offline marketing

Enjoyed a chance to catch up with Kayak CEO Steve Hafner last week. I had planned the call to be about the challenges of a traffic arbitrage business model. I had hoped to draw out of Hafner that there was pressure in the Kayak model fuelled by rising paid search costs, being late to the review game with the Travelpost revamp and tremendous marketing and product pressure from the drop/elimination of booking fees by the big online travel agents (OTAs). Instead Hafner was relaxed, confident and ready to push ahead with millions of dollars in offline (yes offline) advertising planned.

We touched on two main areas. His focus and plans for the next twelve months (including plans for Travelpost and why the OTA fee cuts don't phase him) and his thoughts on expansion outside America (not in Asia and measured in Europe). On the former there is a lot to be worked on for Travelpost to catch TripAdvisor but there are plenty of flaws with the TripAdvisor product and an accompanying disquiet amongst users. On the latter, the potential risk I can see is that they may be under estimating the challenges of growing in Europe where they have more competitors and less compliant suppliers.

Here is our exchange in detail.

On plans for the next twelve months

Hafner says that Kayak is exclusively focused on three things:

1. More focus on core search: The measures here are speed, accuracy and simplicity. Hafner is measuring his world in terms of milliseconds in response time. I asked if we was worried about price accuracy, database loads, hotel and switch look to book issues but none of these concerned him. For Kayak the true cost per query is falling to near zero through caching and the costs of bandwidth. This allows him to focus on the speed of search and the comprehensiveness of the results. His goals are big but simple - that the submit button results in a search in 15 milliseconds, that the results contain every bookable option and that the filtering and customer profiling gives the client the results they want. While this sounds obvious it was the simplicity and aggression in his focus that impressed me;
2. Driving awareness: Hafner believes it is the perfect time to get back into offline marketing to take brand awareness to the next level and compete with the OTAs. He believes that Kayak is "fully penetrated online" and that the costs of offline has "come down by about a third". Critically he does not want to leave the offline channel as the exclusive domain of the big spending OTAs, especially because (as he puts it) "the fee cut [by the big OTAs] takes margin away from their P&L and out of their marketing budgets". New CMO Robert Birge has a $100mm to spend on marketing and a CEO keen to see the brand in lights on TV (example below of their "trip idea" commercials from back in 2006). Right now Hafner is claiming that 8% of online shoppers have heard of Kayak (cf he claims Orbtiz number is 60%). In two years he wants the number to be 20%. ; and
3. Making Travelpost a viable competitor to TripAdvisor and Travelzoo: Kayak has followed the much smaller Uptake into the review meta-search model through a revamp of Travelpost. Prior to the revamp Travelpost (acquired by Sidestep) was a user generated hotel review site much like TripAdvisor. Now post revamp it aggregates reviews from around the web as well as allowing direct posting and commentating. He plans to go after both TripAdvisor and Travelzoo with this new product. He hopes within two years for Travelpost to be generating about half the revenue that Travelzoo is making from deals and to be 15% of the size of TripAdvisor's media revenue (up from 1% now).
   

On Expansion plans

An interview with Hafner is famously free from PR generated answer obfuscation. I asked a detailed question about the Asian market that started with a lead in on the challenges in the market, the earlier successes of Qunar in China and Wego in Singapore and Australia. Even made a reference to Sprice and Cheapflights. "So Steve," I concluded, "do you have your eyes on Asia too?". Two word answer - "absolutely not". In short he thinks the market is too small (in terms of search volume) and not mature enough (in terms of online advertising).

Europe is another matter. He admits that the change in MD "reflects a disconnect in aggressiveness" which I read to mean that the outgoing MD had a more aggressive plan than Hafner did (see the Travolution post on this for more details). This does not mean they are pulling out of Europe and he rejected any suggestion that Kayak had made a "false start" there. Instead they will keep on with the general three strategies above run by the two people in the London office. He conceded that there are product gaps in Europe (no Rayanair and some other low cost carriers) as well as higher costs from online marketing as Google is so much stronger but he is there for the long term even with no plans to replace the Managing Director role.

The competitors are coming fast at Kayak with big marketing budgets and constant model changes. Kayak's response is keep doing what we are doing only better and now on TV. What do you think. Good plan?

---------------------
Example of earlier offline advertising efforts by Kayak.

Content, content everywhere but who is getting the traffic

Content, content, content. Everywhere we look there is content. Back in August I posted that the next phase of online travel will involve consumers being bombarded with "too much information". Having mastered the art of booking travel online, consumers now found themselves overwhelmed with choice and information - looking for help. Recently I co-blog-erated with Alex Bainbridge on some rules for success for start-up content companies. With this background I was looking through my inbox at all of the different types of content companies that were writing to me. I found a spectrum of approaches to capturing long tail traffic with review and content sites.



Kango is targeting the tail through meta-search for review content. As per my Interview with founder Yen Lee here Kango have built a review meta-search business. Modifying business models of the meta-search booking sites like the combined Kayak/Sidestep, Kango has built an aggregation engine for reviews from a multitude of sites. It takes time to put together (hence the beta is only on two destinations - California and Hawaii) but Yen has been able to raise a lot of money and is working on balance between building search architecture and attracting content. UPDATE - Kango now called UpTake.


While Kango is looking to index content on a grand scale, I am constantly coming across or being approached by deeply targeted local content players. 71 miles covers weekend trips in a small part of Northern California. Volcanoetna.com covers - as the name implies - tours and travels around Sicily's famous Volcano Etna. While covering completely different parts of the world and travel experiences, Adam Rugel's 71 miles and Enrico Forte's Volcanoetna are at the longest point of the long tail - focused on detailed and informed content in a very specific location. The pluses for them is that the long tail loves targeted and specific. Nothing brings long tail traffic better than detailed, focused and informed content. The problem is scale - a highly targeted site can become the number one place for a specific piece of long tail traffic but you still need a minimum of traffic to generate advertiser interest. Rugel of 71miles admitted to me that traffic was not a problem but monetisation was proving difficult. Local advertisers who should have been interested didn't get it (online that is) and are spending their money in the local newspaper instead (my theory). Better to be part of a network - Rugel said.

So targeted brings the traffic but not the money. Want proof? Four years and GBP7.4 million pounds later and VisitScotland has shut its doors. Hotelmarketing.com has the story - but the short version is not enough traffic to justify and not able to monetise.



This is where Gawker Media's Gridskipper has an advantage of the hyper long tail sites. It has been around long enough and writes on enough material that it generates the long tail traffic. At the same time it has the network effect of the Gawker family (including Gizmodo, Lifehacker, Consumerist and the NSFW Fleshbot). UPDATE - Gridskipper no longer owned by Gawker.


Video is also playing a role in this race for content. Geobeats are trying to build a business around user gen travel videos. Players are racing into this space as well - driven by the phenomenon of youtube. Pixsy is in this space too (including a relationship with Lastminute). Both are following Travelistic (who I first covered here) who (as far as I know, tried to do it first). The advantage of video in this race is that video generates a greater engagement than a text site. They may generate less page views than text but good video sites keep people on the site longer and interested longer - all elements that advertisers love. As GigOM said "Page Views are dead, Engagement named as Heir".

The final style of player is the now ubiquitous social media, social networking and user generated player. There are a lot to chose from here in this post:

• Where Are You Now (WAYN) - the well funded (but not owned by AOL) travel social network. Millions of travellers talking to each other - telling people where they are, what they are doing and what they want to do. An advertisers paradise;
• Guidespot - targeting DIY local guides. Contributors generate traffic and "kudos" (maybe one day revenue sharing) from producing and publishing their own destination guides;
• Localguides.com (profiled here);
  
• The Sidestep (now Kayak) owned Travelpost - specialists in hotel review infrastructure and traffic. Sidestep bought it and Kayak doent seem to want it but the traffic is flowing; and
  
• Travelpod - started travel specific blogging and had the hottest Facebook widget (Traveller IQ test) for some time. Again - plenty of traffic off the back of a flood of interactive and interested consumers. Now part of the ever expanding TripAdvisor family.

New players are rushing into this space:

• Vailoma - still in beta- trying to build a community around content and guides UPDATE - Vailoma now called TripSay;
• SimpatiGo - leading with the "classic" map mashup around activity. Trying to drive search and "engagement" through map and neighbourhood based information and search.
  

As I said in one of my early posts - the difference between a online content company vs a retailer is that it is significantly easier to launch a content company (no need to build extensive supplier relationships) but it is harder to achieve and maintain scale. This means that content companies have to innovate sooner, faster and more often to keep the traffic flowing. So after an analysis of a dozen companies I feel even more certain that my four rules of success apply.

1. Content - lots of it
2. Index - a fantastic Google friendly index and expertise in search optimisation
3. Access methods - varying ways and means for consumers to access the content
4. Patience - time (and money) to wait for the traffic to build.

What do you think? Who have I missed?

UGC vs Editorial. What's better? What's the balance? What's more 2.0?

A debate is going on about content and Travel 2.0/Web 2.0. What is better at providing recommendations and information - the crowd or the specialist? TripAdvisor has proven that that crowd has power. Millions of people with millions of points of view producing a searchable and understandable review mechanism. But Web 2.0 is also pushing back on the size and difficulty of listening to the crowd. This has produced plenty of sites that believe that editorial is now critical - for exampl human search start up Mahalo is producing dedicated and editorially driven guides.

The crowds eliminate the errors of a editorial statistical sample of one but there is a point at which UGC is too hard to navigate. Ever tried surfing YouTube? Just going to the YouTube homepage with no plan in mind and looking around for something to watch rarely produces entertainment. Compare that experience to going to Break.com or Digg Videos where an editor or technology backed by the community has produced an easy list to tell you what they think you should watch. There are less videos on Digg and Break, and you may not like the ones you see, but you can be certain that you will find easily something to watch that does not involve a cat or an idiot miming Justin Timberlake.

Travel reviews and content are similar. As much as I shift from giving Lonely Planet a good and a bad time for their online activities there is a sense of comfort and certainty you get from reading a Lonely Planet story or recommendation. This comfort comes from their brand, their readership but mainly from their editorial rigour. They send someone to everywhere they review and they send them back again and again. This is the pure editorial model. The positive of it is that trust is already established and time and energy is spent maintaining that trust. The downside is that it is less reactive. The editorial based product sometimes misses updates because its research or publishing cycle is...well.. by definition a cycle and not immediate.

In the wisdom of the crowds model you throw as many human beings as you can at a problem/issue/thing/wall and hope that the accumulation of knowledge produces the right answer and that the right answer can be filtered by those looking for it. This is what TripAdvisor does and there is a point at which for every hotel they have their is a terrible review (room sucks, staff awful, never again), a brilliant review (love it...oh my God I love it) and a whatever review (it was ok, kinda, sorta, you know). I discussed this my "Who you callin' ugly post" about Travelpost.com.

As with all these things it is about balance. You need the crowds to give you the immediacy of updates and the statistical spread but you need the editors or trusted uber user to help you sift and order the information. This has problems of scale (how do you balance controlling quality across a huge range of destinations and information) and openness (to be an editor involves cutting out some of the crowd - a crowd that can get ugly when you cut them out).

My thoughts on how to balance this are the same as for a transactional site/intermediary that has thousands of hotel/supplier options. In the early days of hotel/air intermediaries there was discussion that as some point a limit to the number of choices would arise. How could a consumer possibly look through 10,000 hotels, then 20,000 hotels, then 30,000? Now people like HRS and hotel.de are claiming 200,000+ hotels. The answer is more than better search technologies. It is combing technology, merchandising/rev management smarts and user reviews to generate sort order biasing. This is the approach that content companies need to follow. Match UGC with editorial biasing.

This thinking came to me in an email exchange with a startup pitching itself as the review/information site that has found the balance. The email exchange was with Jim Johnson Vice President of Product Management of LocalGuides.com - itself a spin off from a local search company called LocalMatters. The LocalGuides site is a site that wants to provide...well...local guides. Detailed information on destinations about activities, restaurants, hotels etc. Jim and I had a very interesting mail exchange on the how they would find the balance. Here is what he had to say...

"Our approach to Local Guides has been to seed the site with the key editorial content to provide users a starting point to iterate based on their own local expertise. The interesting transformation is the community has used the tools to create a personal publishing platform (think blogging on steroids) in ways beyond what we originally thought. Guides have been created around making sushi at home, best concert venues and how to get out of the doghouse with your wife. We expected more top 10 restaurants’, hotels and bars to be the norm. The distinguishing focus on Local Guides is in its “personal local publishing” orientation; rather than aggregating all user’s content into one summary view of a business or place, it supports each user’s unique views as “guides”, and builds community around exploring, sharing and relating to “individual voices”. While sites like TripAdvisor, while highly useful for opinion aggregation, we aim to retain the opinion and viewpoint of each individual. Somewhat more of a “reporter’s view”, if you will."

Their view is that the editors set up the framework, provide the basic facts and guidance and then let the crowd fill in the detail with more freedom than you see on TripAdvisor. It is hard to provide much commentary on their approach and whether it will work as the site is in beta - though I would call it very early beta maybe even alpha as there is a lot more content that they need (editorial or otherwise). I like the idea but not sure if that will provide the balance I have discussed. If you give the crowd too much control you risk looking like the comment list in a popular video on YouTube.

Been a bit of a long post but would be interested to hear your thoughts on getting the balance between the crowd and the editor.

Sidestep can't stay out of the news - what will be next

Sidestep seemed very comfortable for a long time being a pure meta-search company (is arguable that they invented the model). The economics were simple - buy traffic from GYM (Google, Yahoo, MSN) cheaper than you can sell it to advertisers who will come to the site because you can deliver highly qualified traffic. However somewhere something started to break in the model, most likely this was that PPC costs for the best of the travel keywords were becoming too expensive. Traffic had to be acquired in cheaper ways. So Sidestep has gone on and acquisition and content deal blitz-a-thon.

The latest is today's announcement of the acquisition of travel social network TripUp (here is the paidContent story). TripUp has all the usual features of a social network around travel - reviews, videos, photos, community, interest matching etc. What it does not have is a way to search and book travel.

Rolling up all Sidestep's recent activity now they:

• own a social network;
• own a blogging/review platform (Travelpost);
• have a content deal with Frommers;
• have a blog content feed from RealTravel; and
  
• another round of financing banked.

That means content, distribution, meta-search and money. There are also very strong confirmed rumours of profitability.

So what is next for them? Not sure if it is time yet for a sale. To extract the most from their valuation they probably want to have a longer period of earnings growth. But I expect them to continue with content and distribution acquisition through either buying more companies or further licence and affiliate deals. On the product side there is more international expansion to do. I am also sure they are contemplating more in the area of video and photos (Hotelsbycity for example). To round off the speculation, what about a technology buy. Something that would improve their ability to push contextual deals in automated forms - think Google Adsense but for travel deals.