Late last year, Kayak co-founder and CTO Paul English stepped down to pursue another venture. The news was significant, as he was a key driver to the brand's success over the past decade.
English was succeeded by Giorgos Zacharia, previously the chief product officer and a five-year Kayak veteran.
As English said in a recent Tnooz interview, he's "thrilled" with the guy that took his place.
It's no wonder why: Zacharia has an intimidating background that includes a stint as a sergeant in the military and more degrees than one hand has fingers.
Zacharia's academic bonafides highlight a deep understanding of machine learning, having earned several degrees from MIT culminating in a PhD in Electrical Engineering and Computer Science.
Zacharia also co-founded two companies that used proprietary trading algorithms in the financial markets, bringing some startup knowledge into a company that prides itself on maintaining a scrappy internal culture.
So what does Zacahria have in store for Kayak's future? It seems that there will be much staying of the course, as English left to pursue another venture rather than because the ship was drifting. The course has been well-charted and navigated, and much of the product coordinates were put in place by Zacharia in his prior role as Chief Product Officer.
Iterating current products via live testing will continue to receive the primary focus, with data and analytics maintaining the core of the Kayak technical process.
And with stats like 35 million app downloads, 73% of mobile users are iOS, 54% of mobile bookings are same day and 50% of mobile flight bookings are one-way, mobile-optimized experiences will continue to be heavily prioritized under Zacharia.
Tnooz spoke with Zacharia from the company's Berlin offices, where Zacharia is on a multi-country tour of introduction to staff and media.
These are definitely some big shoes to fill. We just did an interview with Paul on Tnooz, where he said, “I'm thrilled with the guy who took over my job.”What are you bringing to the table compared with Paul?
I have actually been with Kayak for more than five years. A lot of my responsibility I was already doing over the years, as part of the longer-term. I worked with projects very very closely. What I bring to the table is a heavy data focus. My academic background is in machine learning, and I started a couple of companies using machine learning.
Kayak has a lot of data. Over the years, we have been optimizing the website for the best user experience by running lots of A/B tests. Every day, 30-60% of the users are part of some experiment we are running. We may be evaluating a new machine learning algorithm for our personalization of fares, or a complete redesign. We’ve redesigned the website about once a year for the past few years, and the same goes for our mobile app.
The way we figure out how good our ideas are - we have very talented designers - [is that] our users tell us what works with their clicks. Incrementally, we optimize the website.
Part of the mobile tradition for Kayak is that our designs are increasingly more mobile inspired. The most recent design has actually been driven by the mobile design team as well.
You’ve got that deep background in maths and sciences. Do you see Big Data as the most pressing issue in your new role? What do you think is the first opportunity to address?
We’ve been very big on Big Data at Kayak for a long time actually. We have our own internal analytics tools that we’ve developed over the past 5 years, and a few of our recent features on Kayak, such as the Price Forecasting, rely on Big Data and machine learning.
My big focus is to improve the user experience, particularly the booking experience, and especially on mobile. One recent experiment we’ve run [mockup below] is a booking path experience where we allow the user to stay within the Kayak website or mobile app and execute the booking from one of our partners within the Kayak experience.
Traditionally, we chose the cheapest price and presented that one option. This experiment showed all partners, and this had a dramatic impact on conversion rate. So we will be transitioning the mobile experience to show all the possible brands that Kayak has integrations with in the native app.
We also recently released a touch-optimized version of the website for tablets. It’s only Kayak US right now, and we’re working on translating for the rest of our locales. We’re in the process of taking that design and applying it to the native application for iPads.
You’ve been the Chief Product Officer already in charge of product development. Tell me more about how you’ll be driving new products, and the continued evolution of the on-site experience.
Kayak Trips is a free itinerary management solution. A vast majority of the itinerary is sent to us, and the user gets a nice, clean, free flight status notification on their mobile phone. It’s particularly popular with our frequent users on Kayak.
The booking path was partially motivated by our users who complained about the experience they got after they click out to other websites. And that’s an indication of how we handle our user feedback. We have no customer service function, so if a user sends a complaint, it’s part of our process to aggregate user feedback to see what we should be doing next.
We use some text management techniques to automatically categorize the issues, as we get a lot of feedback. We have to automatically analyze some of it to figure out where you should be spending your time. Also, there’s nothing like exposing the same engineers to the same problem a few times a week so they are more motivated to fix it.
Mobile continues to dominate the allocation of resources in many travel companies. How important is the evolution of Kayak’s mobile product? Is the web still vital, or is it moving towards a mobile-first experience with the web seen as complementary?
The web is still the majority, but we can see the future in several of our countries, where mobile will be the majority by the end of the year. At least from a design perspective, we are definitely mobile-first. From a user experience perspective, many of the problems we are tackling are mobile-first.
But desktop is still a big deal for Kayak. As far as the brand identity of our designs, we are definitely mobile-first. You will see increasing translations of the desktop experience to mobile.
Let’s talk about flight search. Do you have any plans on improving this product and getting it ironed out?
One side effect of being a mature company and a mature product is that we keep adding all of these features and it can make the website look busy. That was part of the motivation of the mobile design as well. Let’s only keep features that are used by a big number of people, not edge cases, while still allowing users to get to the edge cases.
So our flight UI will be revamped significantly.
I encourage you to test our iPad experience - I think it gives you direction.
How do these products play into your strategy for overseas expansion, especially looking to Europe and growing outside of North America?
We launched in 15 new countries last year, and we’re in 51 countries now. We have a few more in the works, and that was actually part of the motivation of the Priceline Group acquiring Kayak: to help us expand internationally.
Initially, we only launched the hotels and the light version of our flights, with the focus on making those complete products. Flights will again be the lead vertical in these additional countries. But we’re looking to make these complete products - including mobile.
As Expedia-owned Trivago pushes into North America, do you think this consolidation make your job harder or are you well-positioned to keep this pace of expansion?
Our Big Data expertise allows us to have a better experience than others. You will not find duplicate hotels on Kayak. We have the most success of hoteliers that are coming into Kayak, and we match and merge hotel records. Most of this is done automatically using machine learning.
You will find the cleanest data, the cleanest directory, on Kayak. We have more than half a million unique hotels on Kayak today, and we expect to have at least 600,000 by the end of the year, given our growth.
In the US, with Trivago coming in and a very active Google, do you still have the same advantages? How do you maintain your position there?
Our biggest difference in the user experience is that we focus on conversions, while others focus on click outs as they get paid per click. So I think we have a significantly better user experience. And again, the cleanliness of the data. Some of the names you have mentioned have a challenge in maintaining non-duplicated data.
Google is a great source of traffic for us. Hotel Finder and other approaches that Google has tried haven’t really affected us.
There hasn’t been much talk about integrations with any other Priceline Group companies. The brands remain competitive, like at Expedia. Is there any sort of synergy there?
Priceline is run as a holding company, so there are no integrations. There are lessons exchanged, so we have an SEO share or mobile share, but we are truly run as a separate company. Booking.com has great expertise in making inroads into new markets, and marketing expertise, and there are lessons there that we are putting into practice. But no specific integrations whatsoever.
There are exchanges of ideas, but the resources are our resources. The code is our code. Each company is responsible for their own destiny.
Paul talks about a rule within the engineering department about not having people that have worked in travel. How are you going to build your team? Are you going to stay the course?
Definitely follow the course there. We actually tried hiring a couple of travel people that didn’t quite work out for Kayak, in terms of speed, creativity and cultural fit.
Do you find that it helps to look outside of travel to get the best engineers, especially when competing internationally for the best talent?
Hiring great engineers has always been a challenge for us. That is also part of the reason why we organically grew a couple of engineering teams outside of the US. One of them came through the acquisition with Swoodoo, so we have a great team in Lithuania. And recently we launched an engineering office in Berlin.
How do the international teams contribute their own perspectives to the product?
The Kayak package tour vertical was led by our European team. [They contribute] from a product management perspective, and even a development perspective. Now that product has made it into the US. We have placed that code there, and that was developed in Europe.
Recently in Europe we launched a credit card fee calculator, a use-specific case for Europe, as different folks have different fees.
Product usage related to mobile is different country-to-country, and have different distributions of iOS versus Android, so we have to pay attention to both. And we see different usages of mobile web versus native applications.
This means we have to pay attention to both native apps in iOS and Android, and patch the UI on mobile browsers.
Traditionally, we’ve paid more attention to native apps, and we’ve realized we need to pay more attention to the Web version on mobile - which is why we did the whole tablet web redesign. [The iPad experience] has been one of the most successful experiments this year - conversions, repeat rates and revenue are all up.
What do you think when you look 2 years ahead? What are some of your specific challenges and opportunities? What’s the specific growth target from your perspective?
The growth target is definitely more mobile and definitely more international.
You recently decided to stick with ITA. How much inventory is coming from Amadeus versus ITA, and will distribution change with you?
We don’t disclose this anymore as we’re no longer a public company, but we still have Amadeus, ITA and TravelFusion. We have a very tight relationship with Amadeus.
Amadeus is the best internationally. ITA is still better for certain domestic type of queries, but increasingly a lot of our domestic inventory comes from direct integrations with major airlines in the US - and also Amadeus.
So as you simplify your website and get rid of things that no one uses, who else are you looking to as far as the web experience?
I think Airbnb stands alone as a unique example, as far as simplicity and the smooth and modern UI.
Is it possible to deliver such a complex product, such as travel with flights, hotels, cars, and more, so simply?
For a user who goes down our unique booking path experience, the answer is yes. But we have our own challenges, especially for users that choose to go down the click-out path. There’s not much we can do there in terms of wherever they end up with the click.
NB: Path image courtesy Shutterstock.