This article originally appeared on Phocuswire.
Last year I wrote a piece entitled Three Reasons Why I Don’t Use Recent Travel Startups, casting a rather downbeat eye over a number of travel services trying to make their way in the market.
The piece was a bit controversial, I expressed some views about how I didn’t find many social travel or inspiration and discovery apps to be useful, and said that Hipmunk wasn’t a big change from Kayak.
When Tnooz (now Phocuswire) suggested I write an update on my views, I ran it by a friend and she jokingly asked: “Who are you going to insult this time?”
First, an update on the views I expressed a year ago, and then some new opinions.
1) Travel inspiration/discovery need to think like Pandora
Right now these companies are either dead or treading water. Trippy pivoted to become“Pinterest for travel”, WanderFlygot acqui-hired by TripAdvisor, GTrot pivoted from travel discovery to local discovery and then completely out of the business altogether and into a social gifting startup called Boomerang.
Gogobot is still plugging away, but I tried to use them for a recent trip and the lady who responded to my requests was a Gogobot employee, not a user, so perhaps not a good sign?
I stand by my previous opinion that Google Images, Wikitravel and asking friends on Facebookare all good enough, and that this is still a big data problem.
Take Pandora, the “music discovery” engine came out of years of music theory research and fine tuning algorithms – a great example, so perhaps these companies need to approach travel discovery like Pandora approached music discovery.
I don’t need one of these apps to tell me that Rome is some place I should probably go. That’s the equivalent of Pandora telling me that I should probably check out the Beatles.
I want them to know that I like Eze, France, and suggest Saint Paul de Vence. For these startups to be useful they need to be so comprehensive that they allow me to discover little gems all around the world, the travel equivalent of the indie punk rock band.
But even once someone nails the algorithm, the current business models won’t work: upselling hotels and flights won’t work as the “inspiration and discovery phase” is too far up the funnel for most people to convert.
Again, the best hope is to model something off of Pandora or Stumbleupon, who have been able to monetize discovery.
2) Hipmunk seems to be starting to act like the rest of the OTAs
My previous opinion was that while Hipmunk was an improvement, it wasn’t different enough to inspire enough people to transfer away from Kayak or Expedia naturally. I still stand by that.
I interpreted the $15 million Hipmunk raised last summer as the brand’s realization that, as a travel search engine that offered flights and hotels, like hundreds of others out there, they needed to pour more money into marketing.
I started seeing Hipmunk advertisements on MUNI buses in San Francisco, and their monthly uniques went up by 2-3x.
Now, some new views and predictions on where the real problems reside where startups can make a difference:
1) Distribution is key because eyeballs are so expensive
If a person only travels once a year, then it is going to be very difficult to reach them directly. There is a reason Expedia is still the industry leader despite little innovation over the past decade, and it’s because they’ve been around the longest.
If they already go to Expedia, the easiest way to reach a traveler is to piggyback off of Expedia’s traffic.
B2B travel companies, or consumer facing companies that have the opportunity for a lot of partnerships and referral traffic, are where the most potential is at. It costs so much to acquire an eyeball in travel, that the easiest way is often to cut in the guys who already have them with some type of revshare model, or focus on an area that they are lacking in.
That means any new consumer-facing travel app is facing an uphill battle acquiring customers, and should think long and hard how they fit into the travel industry ecosystem.
Limos.com partners with a lot of OTAs and airlines. Startups focused on taking on hotel channel management, like Dashbell, can have instant revenue once they lock in a few major customers.
Viator and Grayline, though not exactly startups anymore, partner with major airlines and OTAs for tours and activities. Wanderu is trying to be the ITA for buses and trains; SilverRail, an ITA for rail.
If you don’t have any clever distribution channels, you better have some massive financial backing and clout. Room77 was able to get a $30 million dollar round from Expedia and Hipmunk’s $15 million round is another example.
But if you are a mere mortal, think of what place you can fill in the massive puzzle that is the travel industry, and use that as a launch point.
2) Multi-modal’s time is approaching but not quite there yet
Until you can get results on a multimodal search engine that are at least 90% as comprehensive as doing a search on all the individual transportation sites, all-in-one transportation sites will be useless.
The problem with current alternatives is that for a San Francisco to Los Angeles search, one of them suggests BART as the only way to get to the airport, but as a local I know there are at least 20 other options including more than a dozen shuttles and airporters, Uber, Sidecar and Lyft, so I immediately don’t trust them anymore.
Right now multimodal search gives mediocre results for most forms of transportation and legs of a journey, and savvy travelers figure this out pretty quickly.
Mozio’s initial goal was to built a multimodal, point-to-point search engine, until we realized the timing wasn’t right: other forms of transportation need to aggregate themselves to make it possible.
Wanderu and SilverRail need to succeed in aggregating buses and trains. Each one of these industries is so massive and so fragmented that any startup that thinks they are going to effectively aggregate the world’s transportation through pure hustle is naive.
The smartest companies in this sector have started with a similar focus, and Waymate is one of my favorites. They have also honed in on a niche use case: the Deutsche Bahn in Germany, and used that as a launching point for their entire engine.
They are using a specific use case to build a user base while they work on making Waymate more comprehensive. KelBillet has seen good user growth in France and GoEurojust raised a lot of money to take a crack at the same problem. Focus on a sector, a country or region, and let the industry catch up to you.
3) The focus shouldn’t be “multi-modal” but “point-to-point”
Once the various forms of transportation do aggregate, it will be easy for Orbitz and Expedia to add buses and trains as well. Then they will be multimodal, and have all the brand recognition compared to the startups, who will be fighting an uphill battle and won’t even be unique anymore.
So the smartest way to get in on the “multi-modal” revolution is to go one of two ways, either be a Wanderu, SilverRail, Rome2Rio (since it’s now pivoted to B2B) et al, and be the resource the big guys will use for their buses, trains, ferries and transfers, or focus your resources on a different type of booking experience.
Orbitz and Expedia won’t be changing their booking flows anytime soon, they have all been too thoroughly A/B tested and they are afraid of conversions dropping.
The real value add of the new travel startups won’t be the fact that they have buses and trains, it will be that a user could go to them and better compare the ramifications of taking a bus that arrives in the center of a city vs. a flight that arrives at an airport an hour and a half away.
They will be able to figure out that while the flight may be 2 hours shorter, after getting to and from the airport and security they might as well just have taken the bus.
Additionally, they will be able to figure out that they are being flown into a remote airport and it will cost them as much as their plane ticket just to get a taxi to their hotel.
If you don’t get in on the data side of the multimodal revolution, you better be ready to present a more natural travel search than the big guys, because in the end they will probably have the same data you do, but 20 years more of name recognition.
