Support Skift’s Independent JournalismMake a Contribution Now
Editor’s Note: This year we expanded our coverage of the technology companies that do the behind-the-scenes work of powering the systems of the world’s major travel companies.
This is the final Q&A in a series of conversations with industry leaders.
Our Travel Tech CEO Listening Series has showcased the industry’s top minds revealing where business-to-business travel tech is heading.
On any given day, the typical hotel sees between a third and half of its rooms filled by guests who booked their stays via third-party sites such as Booking.com, Expedia, and Ctrip.
But the rise of digital distribution didn’t happen overnight, of course, and many companies powered the trend. One of the most representative and significant of these companies is DerbySoft, a Shanghai-based vendor of connectivity solutions to hotels.
DerbySoft is best known for building custom software for the world’s largest hotel chains, including Hilton, Marriott, and InterContinental Hotels Group (IHG). All of the top 10 largest global hotel groups and all of the biggest online travel companies use its connectivity tools.
To be clear, though the company is Shanghai-based, 97 percent of its revenues come from outside of China, and about 80 employees work abroad in a dozen countries.
When chief executive Ted Zhang helped to launch DerbySoft in December 2002, his team got their start by building customized enterprise software, which mainly connected into hotel reservation systems and distributed inventory to online travel agencies and other channels.
Back then, this was untilled soil. Most hotel rooms that weren’t booked directly were booked via travel agencies, who mostly used a handful of technology middlemen, such as Amadeus, Sabre, and Travelport, to process their bookings.
During a recent interview, Zhang recalled the moment. He said his team was betting that Internet-based agencies would displace traditional travel agencies and technology middlemen in the distribution chain.
Roughly 15 years on, DerbySoft’s engineers have been proven partly right, as the rise of giants like Priceline Group shows. That said, technology middlemen still have tremendous power.
Looking ahead, DerbySoft is betting that consumer platforms, such as Facebook and LinkedIn, will become as important to hotel distribution as today’s travel-specific brands like Expedia are now.
“There’s no reason why LinkedIn, Amazon, eBay, Alibaba, and so forth won’t want to sell travel product either like online travel agencies that actually handle the bookings or via an advertisement model like Google uses today,” Zhang said.
He added: “We believe we need to prepare for that day, so that when the big guys like LinkedIn, eBay, and Amazon come to us and say, ‘Hey, we want hotels,’ we will be ready.”
Last December, DerbySoft launched its first custom connection with Facebook that takes advantage of the social network’s Dynamic Ads for Travel program. (For context, see Skift’s earlier articles about Facebook’s effort.)
Marriott was a launch partner. It has been testing selling hotel rooms via Facebook using DerbySoft’s skill at making sure that rates and availability are displayed quickly and accurately. Zhang claims that a couple of other hotel groups are in the pipeline.
DerbySoft is also placing another bet on the future. It believes that independent hotels and small hotel groups will need software to connect with third parties like Ctrip and Trivago and efficiently manage their mix of third-party distribution.
Last November, DerbySoft began offering software aimed at this market. Its hotel connectivity product One is its first attempt at offering software-as-a-service to thousands of small, independent properties — which form the so-called “long tail” of the industry, accounting for roughly 70 percent of hotels worldwide.
Among other things, this software lets a hotel manager offer to Booking.com customers, for example, five king-size suites for specific dates, but only offer Expedia customers, say, three double rooms on the same dates.
Targeting this market poses a tall challenge. DerbySoft is, at its heart, an engineering company focused on giant corporate accounts. It needs a different mindset for its new efforts. It must find its sales-and-marketing feet as a software company operating at scale with small providers.
Skift is not yet hearing anecdotally that there’s been an active reception for the product in the first target market of Europe, though it is early days.
We caught up with Zhang to find out more about DerbySoft. Our lively conversation covers why he turned down Priceline Group’s proposed investment in his company, how he rates Google’s recent change to fees in its metasearch, and why he dislikes the push by some industry forces to create standards for hotel content.
We edited the interview for brevity.
Skift: We’ve recently interviewed the CEOs of the major travel technology distribution middlemen: Amadeus, Sabre, and Travelport. What do you make of those companies?
Ted Zhang: Amadeus, Sabre, and Travelport use a hub-and-spoke model. Given new programming interfaces for the Internet, that model is outdated.
We believe every hotel’s e-commerce system should be a node in the network. We don’t need those hubs to connect hotels with the entire digital marketplace of demand. There are smart people at those companies, but they’re on the wrong side of history.
Skift: How does a small software company in China become the most-widely used connectivity solution between global hotels and online travel agencies?
Zhang: We had small ambitions at first. We nearly failed at the start.
We tried to launch a business called MarketDerby in the U.S., but not long after, September 11th happened. A horrendous tragedy for the country.
We, the co-founders, left for China because we are Chinese and we could wait out the recession that followed here.
Immediately we figure out that China doesn’t even have a system for connecting hotels to anything.
So our first venture is, essentially, to provide the network to empower the independent hotel to have its own website for direct bookings and to be able to process bookings from outside travel agencies and partners and corporate managers.
Skift: What happened then?
Zhang: After we had been the first to develop such software for small hotel companies in China, we figured out that travel agencies are going to need a better system, too. We developed a tool for that.
Next, we connected the hotel systems with small online travel partners for distribution to customers worldwide.
We’re pretty much the first outside tech vendor to work with IHG and Hilton for distribution to small online travel agencies.
Then IHG and Hilton then gave us a shot at connecting them with the big demand partners, meaning Booking.com, Expedia, and so forth.
The other global hotel chains followed.
Skift: Something’s missing from the story. How does a tiny company in China in the mid-2000’s get “in the same room,” so to speak, with a foreign giant like Priceline.com?
Zhang: In mid-2005, The Priceline Group buys European hotel booking website Booking.com. I saw early on how big Booking.com could be.
By that time, we were helping many hotels connect with small online travel agencies. Our reputation was getting around the industry, including at trade conferences.
In mid-2007, I meet with Jeff Boyd, the CEO of Priceline Group. I say, ‘Hey, Jeff: You’re always telling hotel chains they should connect with your extranet.’
[Boyd’s response was]: ‘Well, from your standpoint, making hotels use Priceline’s extranet is great. But for hotels, it’s tough. It’s tough on their IT departments to figure out how to connect with it. And it’s tough for hotels to manage a different extranet for every OTA [online travel agency] on the planet. How can they maintain the multiple extranets? It’s almost impossible. You should provide an alternative system.’
Long story short, Jeff says, ‘Ted, are you guys profitable?’
I say, ‘No, we are not.’
And he says, ‘Well, you know, Priceline Group could send 10,000 bookings a day through your system. But if you crash as a company, we’d be in trouble. Can you handle that volume? And will you be profitable and still around years from now?’
I say, ‘Hey Jeff, listen: This is a scale game, and once we have your volume, we’ll be break-even. If you don’t work with us, it’s really not your business what happens to us.’
Jeff says, ‘That’s a good argument.’
Then he asks me another question: ‘Do you guys have the technology to support our huge volume of transactions?’
I say, ‘Jeff, honestly, if I say yes, will you believe me? It’s better to do a test with our technology and persuade yourself.’
As I remember it, Boyd and Glenn Fogel [then Priceline Group’s head of strategy and executive vice president of corporate development, now its CEO] had the CTO of Priceline plus the director of information technology for IHG coordinate a test.
Skift: And you passed the test?
Zhang: In the test, they raised one issue: You guys are not PCI compliant. [The service didn’t meet the Payment Card Industry Data Security Standard (PCI DSS), which secures cardholder transaction and personal data.]
Skift: So what did you do?
Zhang: We spent almost a year and a half, from pilot to full launch, to make all the connections in a PCI-compliant way and to handle the volume.
Skift: And you got there.
Zhang: We successfully connected IHG and Hilton with Booking.com. Since that, we’ve built up a diverse portfolio of clients.
Skift: If you’re so great, why didn’t Priceline buy you?
Zhang: Glenn Fogel said, ‘Maybe we can invest in your company.’
Also one of the major global hotel chains has come to us many, many times, saying, ‘Hey, we want to put money in your company. You guys do a great job.’
We still say no to them. Because we want to maintain the status of a Switzerland and be a neutral player in the industry. We want to make sure hotel companies should trust us with their data.
Skift: And to address Jeff’s question: Are you profitable now?
Zhang: We’ve been profitable for many years. We have deep pockets. Deeper pockets than our business-to-business tech competitors.
Skift: So a few years pass and, next thing you know, DerbySoft is building customized software and services for Marriott, Hilton, IHG, and other global hotel chains. How did that come about?
Zhang: It’s not by choice. The industry is too complicated.
Each hotel company has different inventory coded in different ways and different business goals. They may want some information released to one travel agency site but not another, or a travel agency’s site in one country but not in another country. There’s a lot of computing involved.
It would be simpler to understand if there were a switch.
Think of power plugs for electricity. Think of a universal electric current adaptor. Imagine that for hotel distribution. You can’t. The work that needs to be done to get all the systems talking nicely to each other is too complicated.
We had to keep doing custom work for each provider to meet their distinctive needs, rather than force them to fit some generic template, so to speak, that we might try to impose artificially.
We do have a common, standard interface for small companies who don’t care that much and who have simple needs. But most hotel groups need one-on-one service.
Skift: Several of these major hotel groups, like Hilton and Marriott, have in-house engineers and teams and budgets. It’s quite unusual for them to outsource major technical work to external providers. Why are they making an exception for DerbySoft?
Zhang: It would cost them more to try to reproduce our solution internally.
If every hotel group connects to every demand partner, how many connections do they need to maintain every single day? That would be a nightmare for a lot of non-tech companies.
Right now, IHG and Hilton, as examples, only need to maintain one pipe, so to speak, from their central reservation systems to DerbySoft.
We maintain the pipe from DerbySoft to Booking.com, the pipe to Ctrip, the pipe to a global distribution system like Amadeus, the one to MakeMyTrip, the one to a corporate travel management company.
Let’s say hotels start offering virtual reality imagery of their properties. And let’s say the different online travel players have different ways of accessing it and displaying it. We could engineer the solution to that problem.
We also specialize in the skill of caching data.
Skift: What does “caching” mean? And why do that?:
Zhang: Some hotels can’t handle all of the requests for data that come in around the world in real-time. By caching, we protect their systems.
By “caching,” I mean storing a copy of the hotel’s data on our systems and answering a portion of the queries from online travel companies with the copies.
We are specialists in caching, so we can do it faster than the hotels themselves.
If we can quickly deliver the data, then we make the data more accurate. Because we need less time to synchronize with the systems and less chance that rates or other details will have changed in the interval.
Skift: There are companies out there that offer some services that overlap with or compete with what DerbySoft offers. There’s DHISCO, Amadeus Hospitality, Shiji Hospitality Solutions, and so on. How can you keep your competitive advantage?
Zhang: I don’t know why they cannot steal business from us, despite trying.
But I do believe that if we know the customer better than them and if we know where the Internet is headed and we adapt in advance, it should be very hard for them to steal business from us.
I focus on the customer’s requirement. I focus on tomorrow’s technology. It’s those two things and two things only that drive me.
Skift: What do you see on the horizon?
Zhang: Every consumer marketplace company will sell travel eventually.
Skift: What makes you so confident about that? When Amazon tried to sell some travel, it struggled and then stopped the program early. Is your idea just that platforms like Facebook and LinkedIn just add links to Expedia Affiliate Network, or something similar?
Zhang: What I mean is, consumer marketplaces will either sell travel products as if they’re online travel agencies or post advertisements of travel deals with others fulfilling the service.
Google, as an obvious example, does not sell travel. They advertise real-time bookable travel listings.
Yes, Amazon gave up. They gave up too early.
They didn’t find the right people to run the business, in my opinion. Otherwise, I’m a fan of their company.
Skift: Why are you so confident about non-travel platforms like LinkedIn starting to sell travel?
Zhang: Because I’m an engineer. I can see the development of the future.
In four or five years, five years maximum, Internet-based technology will allow the typical hotel chain to add a new distribution channel at a little incremental cost of operation, whereas today it costs them a non-negligible amount.
Skift: And why would DerbySoft lead in developing those solutions?
Zhang: We are different from all the existing companies. We are run by engineers. We’re not marketing-first. Engineers focus on solving problems first. We should create a solution much better than others.
Skift: That sounds like bullshit, with all due respect.
Zhang: That sounds like bullshit, but that’s deep from my heart. That is really what I think. I sit and work with engineers every day. Its the lens through which we view everything.
Skift: So you think Expedia may disappear within a decade?
Zhang: No. I didn’t say the LinkedIns of the world are not going to use solutions provided by the existing online travel players to power their travel businesses.
What I said is all the major Internet companies such as LinkedIn, Amazon, eBay, etc., will sell travel.
Of course, there will be multiple ways to sell and support a travel product, not just with white-label services.
Some companies, like Costco, for example, may have special requirements. It’s hard for OTAs as currently constituted to support their business model. Now they may even use a wholesaler or just go directly to hotels to get room supply.
Skift: You seem to be betting big on third-party distribution staying strong. But hotel groups seem eager to have more direct distribution and avoid what they say are extortionate commissions.
Zhang: Well, we never ever encourage hotels to shift from OTAs and wholesalers. Our mission is not to shift hotel business from that kind of entity. Our mission is to empower hotels to be able to effectively work with all their demand partners to the extent that they choose to.
Skift: But guest acquisition costs eat up anywhere from 10 to 20 percent of room revenue at upscale and luxury hotels in North America, according to research by Kalibri Labs. Don’t you think hotels should work to tame third-party distribution?
Zhang: Nobody should dominate the partnership and force another side to do something they don’t want to do.
That said, OTAs provide huge value to the hotel companies. They will exist forever.
Skift: Little of your revenue comes from China, but China is the future. What’s a trend in China that Western hotels ought to take advantage of?
Zhang: Chinese consumers are used to paying for everything with WeChat [an instant messaging platform from Tencent used by more than 700 million people] and communicating with brands via WeChat .
There’s no one use of WeChat. We’ve seen totally different models. But hotels need to set up mobile sites in China and integrate to make the most of WeChat functionality.
Skift: There are no universal standards for hotel content. One association is promoting one set of standards. Amadeus is promoting its own. And so on. How do your engineers cope with that?
Zhang Honestly, I’m not a big fan of standards. A standard means you ask the customer to do something your way. It’s better to create the capability to work with each customer’s requirements.
For example, we connect Marriott content with Ctrip. There’s no standard. They do have some guidelines. But it’s the job of engineers to smell the customer’s requirement, so to speak.
Skift: But having a standard would make your life better?
Zhang: I don’t think we should have standards. No.
Skift: Let’s talk Google. Two summers ago, the search giant made its big change to its hotel price-comparison search. It began to move away from a click-based model to a commission-per-booking model. DerbySoft was right there at the start, helping to connect hotels into Google Hotels’ metasearch using the new model. Is DerbySoft part of some of those elite Google teams that test products in advance and help it with innovation?
Zhang: Yes. Google had a problem. Its metasearch linked to a lot of small and medium-sized hotels that have websites, particularly mobile websites, that offer a poor experience. Google would hand off the customer, but the customer didn’t end up booking.
I suspect that is why Google is pushing instant booking, where you store and use your payment details at Google and remain within its user interface.
I may be wrong, but it seems like right now the hotel has to choose to either do instant book or to handle the transaction on their own. Why not both? Google should support both.
I don’t know, but I suspect they want instant bookings as the default for any hotel company who doesn’t have a sophisticated website. The goal must be more conversions.
Skift: Is Google’s volume of transactions processed growing fast? Does DerbySoft see a lot of volume going through Google?:
Zhang: I don’t know. We help hotels distribute via Google. As of today, Google still does not generate a significant revenue for us yet. Honestly, that’s only a very small portion of our business, so far. But that wouldn’t necessarily say anything about how well Google’s effort is going.
As far as DerbySoft goes, we’ve learned a lot, and we think we have data to show we’ve helped deliver on our promises to Google.
Skift: Let’s end by going back to your philosophy of focusing on the customer’s requirements and on anticipating the demands that will be created by tomorrow’s technology. What other changes do you, as an engineer, see coming ahead?
New technologies will cause hotel back-end systems to merge. Right now there are these siloed systems. They will merge.
The wholesaler business — that entire market is starting to shrink, and that will continue. On balance, the wholesale model is broken and will continue to get even worse.
The Airbnb-type of short-term and vacation rental model will continue to grow extraordinarily fast. It hasn’t peaked yet. Airbnb right now is, room-number-wise, probably the second-largest lodging company in the world, second after Marriott.
Hotel companies need to innovate and redefine themselves as lodging companies more broadly.
Skift: Are you predicting that hotel chains will create vacation rental brands?
Zhang: I do believe the innovation will happen, but I don’t know what shape it will take.
I do know that the cost of guest acquisition will come down. New technologies will help brands gain customers much quicker and more easily.
That doesn’t contradict what I said earlier. If guest acquisition costs go into reverse and start declining as a percentage of room revenue, that should not affect the business of demand partners like the giant online travel agencies. Generating direct business will become more efficient, but so will the OTA model, too. Both will remain important.
Check out other articles in Skift’s Travel Tech CEO Series, including interviews with the CEOs of Amadeus, Sabre, Travelport, SITA, Points, DHISCO, and Farelogix.