NB: This is a guest post by Alex Kremer, co-founder at Flextrip.
Jeff Lawson, CEO of app software firm Twilio, gave a great talk at last week’s Gluecon conference about enabling so-called “DOers” in large organizations.
His primary thesis is that easy-to-use, self-service APIs that are clearly documented, clearly priced and allow instant sign-up enable innovation actually work and benefit industries.
By this standard, is the travel industry focused on enabling innovation? Not really.
There are dozens of APIs and data feeds offered by large and small travel organizations sitting on mountains of data, but many of these APIs share significant problems.
The most common are:
- No self-service signup — must talk to a sales/BD person
- Data constraints — only serving part of a larger dataset. Frequently the most useful/valuable data is locked away.
- Lock-in — the sole purpose of the API is to generate transactions and customer ownership for the API provider.
- Pricing — frequently a huge problem for startups, especially in Air (see Ashley Raiteri’s recent Tnooz post on Why is the travel industry so unfriendly to startups?)
All of the above represents myopic thinking and severely limits innovation.
Existing gatekeepers in the travel industry love it, of course: barriers to entry are kept high, and attempts to revolutionize inefficient markets from within are next to impossible. The sleeping at the wheel can continue.
What needs to happen?
The practical case for open, easily accessible APIs is clear. The most popular APIs have spurred not only innovation, but entire ecosystems (eg. Twitter).
However, these cases didn’t have the entrenched business model issues plaguing travel. Lets tackle the above points individually:
1. Self-service signup
This should be the easiest issue to solve. Enabling self-service doesn’t threaten business models.
Enabling a DOer to sign up and start hacking away makes sense. Putting a layer of sales-person in between the DOer makes him a lot less likely to DO.
Shaival Shah wrote a great blog post, Cannibalize Business Development by Popularizing your API, which expands on this topic.
2. Data constraints
The openness and availability of data is the single biggest enabler of innovation. Big companies that hoard data are doing themselves a disservice in the long run.
If the data is essential to opening up a market, an innovative startup will eventually find a way to gather it without said big company and render the company irrelevant.
A great example are APIs such as SimpleGeo, Foursquare Places, Google Places, and Facebook Places.
Does a bootstrapping startup need a $20,000+ business listings database anymore? Nope.
The big players who for years priced startups out of the market thinking their data acquisition efforts would pay off in eternity are busy refocusing their models right now.
In travel, provider-specific databases (think air pricing data) are less open to being crowd-sourced, but this problem will also solve itself over time: all it takes is for one airline to take a leap of faith in opening this data up to foster more direct-connect relationships.
When (not if) this happens, expect the days of extravagantly expensive airfare feeds to quickly fade. The mantra is simple: be open yourself, or someone will eventually do it for you.
3. Lock-in
There are tons of examples of creative uses of data opening up entire markets. Locking users of an API into a single, specific use is short-sighted and detrimental to the core existence of said API.
Yes, many APIs are built in furtherance of a business case but it’s frequently the non-traditional uses of an API that create great products.
Being as liberal as possible with usage restrictions has clear value in the long run. Allowing DOers to DO more with API data also enables the API provider to monetize more of said data.
4. Pricing
It’s pretty obvious that $50,000 API costs stifle innovation. Many data-dependent startups won’t make that in their entire first year of operation.
APIs don’t need to be free, but they need to be reasonable. There are startups in the API market that allow companies to monetize their API efficiently and on a per-request basis.
While this model isn’t universally applicable, it makes a ton of sense in travel.
What enables innovation more? A $50,000 up-front cost? Or $0.01 per call to allow a small, scrappy startup to invent a great product that will eventually use millions of API calls per year?
The answer, I hope, is obvious.
Conclusion
Opening up data is scary for anyone. But the thing to remember is that enablers of innovation do not automatically enable their own competition.
The focus for current gatekeepers should be to find new ways to monetize the ecosystem by allowing DOers to DO. This will undoubtedly create hundreds of new, amazing services and products that sell more travel.
What are you waiting for? Let’s get started.
NB: This is a guest post by Alex Kremer, co-founder at Flextrip.
NB2: TLabs Showcase – Flextrip.
Related posts:
About Special Nodes
One of the very few best guest posts I have seen for ages. Good piece of information worth reading. Thanks for the good post.
Good job Alex. This is a very practical article that hopefully will open up the eyes of the “old guard”.
Innovation has always threatened the status quo. But hang in there! Innovation will prevail.
People inside the industry have been saying this for years, but like you say – until someone takes the plunge, it’s difficult.
Kayak had an open API for a while but the data was low value. Full value API access is sort-of available on some affiliate programs, but the most hardcore API datasets (full pricing for everything) allow a new company to set up a virtual distribution channel. Travel is possibly not ready for that, because – let’s remember this – margins are shockingly low, especially for transport (rail, air, ferries…)
You’re right about the margins. This is exactly why I hope gatekeepers will see APIs as a new revenue channel on a per-call access basis.
Alex, if we add a few more voices we might make up a chorus calling for change.
Great piece
Great post Alex, couldn’t agree more. Whenever I talk to travel companies about this, I hear a load of excuses ranging from data ownership issues to security concerns to the complexity of transactions in travel.
I believe those excuses will catalyze the demise of the big industry players. They must learn that being open is the best way to solidify their role as the true data owner and maintain a place in a changing industry, not by being overprotective.
As you say, innovation is good for all (yes, scary at times for some people, but still good!).
There was a time – way back when – I was a gatekeeper. I took my role very seriously. But I learned that it was also foolhardy to perpetuate this. So developed a process for getting around the various roadblocks inside my own GDS (Worldspan). In return Worldspan became the GDS of choice for developers. Including arguably the biggest API user of all time – Expedia.
The recent THacks while rudimentary in their use and application have demonstrated that there is a large and eager community to open up the gates… indeed Mr G – OK the three of you – (and I don’t mean Gorbachev) Tear down THIS wall.
However there are still other significant barriers to entry. These are commercial and legal. The current crop of GDS APIs come with such an appalling amount of legal mumbo jumbo and restraint as to be unusable. Further the full commercial cost is not designed to protect the technology platforms of the GDS (fragile though they are) but rather to prevent any development that would ultimately come and create a viable competitor to these legacy GDS entities. Witness the DoJ investigation into Sabre for doing just that to Farelogix.
We have a long way to go. But as we all know – right doesn’t beat might. But if right acquires enough might then it can defeat the forces of evil…. know what I mean?
Cheers
Second post in a row ranting against gatekeepers and old guards, and still no names mentioned?
Daniele, Timothy was clear that Worldspan [proxy for the GDS as a whole] was one of the original gatekeepers.
I agree with Timothy that the barriers to entry today are largely not the technical ones, but are business models and contracts. The original three GDS New Entrants had the technology to marry buyers and sellers, but were not able to aggregate the buyers. A better mousetrap was not enough.
Now the circumstances are slightly different. We’ll see how it turns out this time.
Another key point Alex is IP restrictions / code sharing restrictions. For example if someone writes some code using an API that is hidden under layers of NDAs – then that code can’t be shared between companies – or put online – because that act of publishing sample code will expose the structure and nature of the API (which the NDA will restrict you from doing). Hence there ends up with a lot of wheel reinvention going on – and very little standing on the shoulders of giants that have gone before.
A counter example – e.g. Alex’s company (article author) has, as a result of us opening our API, built a code library that he has open sourced and let other developers use…. That to me is a classic benefit (to us) of going 100% open with our API.
A theory….
In tours/activities you can’t do price comparison as everyone has different products so the premise doesn’t conceptually exist. That is probably one of the reasons that the “old school” data in flights / hotels is so locked up – because you could take data sources from 10 places (if they were open) and only 1 would ever get the transaction….. the rest would just be used for price comparison – or even price setting
Perhaps thats why us tours/activities people can completely open our data while others don’t
We are currently building a system to analyse the rental holiday market and relative attraction between listing sites like Holidaylettings and Homeaway etc. It will work on the premise of data inserted by individual owners and then supply them with aggregated data to show who achieves the most bookings in a given area against given metrics. We hope this will support the smaller sites as they achieve bookings but are not able to promote themselves in the massive ways the big sites do.
We then also use the data we collect to market our affiliation network to bring onboard those smaller sites in our network.
You are right. Big is not necessarily beautiful!
Alex, being a young travel technologies company ourselves, i can partially agree with your view. However you need to keep several crucial points in mind:
1) Support for open access costs a lot – If GTA or Amadeus had to manage thousands of developers, 95% of which would never deliver complete products.
2) What would happen to travel tech market would be similar to website development markets, creating a virtual pricing without correlation to quality of products. This will also hurt small medium dev. companies, whom large players can adapt and outbid in any case.
3) The GDSs are already restraining most participant from running multi GDS queries and demand search to book ratios, partially due to enormous server loads. What if those loads would double or triple in one year?
4) Most of the existing travel inventory can’t be booked without direct commercial relations with suppliers or appropriate licenses/memberships. Thus unless i can afford the operational costs, what will i do with all the data?
I think these are the reasons why meta-search and affiliate models became so popular, giving the start ups access to data. But essentially not having full control over it, what they really are creating are nice looking interfaces.
I am a Meiya travel company staff, API integration technology will bring wireless business opportunities, better services for tourists.