How to become Google and Amazon. Richard Harris, CEO Intent Media talk summary at TTCR 2018
Richard Harris, CEO Intent Media Inc, came from New York to Moscow to talk about discoveries and ideas that they did in Intent Media while working with global travel brands. What happens with travel retail websites in the era, when Google becomes a hotel and flight search engine?
We publish the summary of Richard's talk. You can watch the full talk on YouTube channel Travel Tech Conference Russia. Subscribe:)
We publish the summary of Richard's talk. You can watch the full talk on YouTube channel Travel Tech Conference Russia. Subscribe:)
By the way, f you want to get more notifications about new articles on Travel Tech Conference Russia, subscribe to our newsletter.
We are a machine learning company for the travel industry. We use machine learning to predict user behavior to generate more profit and loyalty.
I want to share some big trends. I don't pretend to be an expert in Russia, but from the global market we see that the giant trend is convergence. Specifically between commerce and search. And I also want so share some micro data about user behavior that you can predict drawn from billion user searches.
The ability to predict what your user want is one of your main defences if you are in the transaction data.
Here's the trend of convergence. Travel and search were divided for years. Transaction business was concentrated on such metrics as transaction values, conversion rates etc. On the other side are search engines who care about completely different metrics.
In past, especially in travel, those were separate worlds. The DNA of people and companies was different. But now we begin to see the blurring between that. The headline from this year is that Google is shifting to the commerce and Amazon goes into search.
One thing to note, Amazon today is going to make 5 billion dollars in revenue in advertising. Now 54% of commerce searches start on Amazon instead of Google. Here's a look on the Amazon page, and it looks a lot like Yandex or Google. Amazon a becoming a retail search engine. Google, the ultimate metasearch, is getting into the booking game.
I want to share some big trends. I don't pretend to be an expert in Russia, but from the global market we see that the giant trend is convergence. Specifically between commerce and search. And I also want so share some micro data about user behavior that you can predict drawn from billion user searches.
The ability to predict what your user want is one of your main defences if you are in the transaction data.
Here's the trend of convergence. Travel and search were divided for years. Transaction business was concentrated on such metrics as transaction values, conversion rates etc. On the other side are search engines who care about completely different metrics.
In past, especially in travel, those were separate worlds. The DNA of people and companies was different. But now we begin to see the blurring between that. The headline from this year is that Google is shifting to the commerce and Amazon goes into search.
One thing to note, Amazon today is going to make 5 billion dollars in revenue in advertising. Now 54% of commerce searches start on Amazon instead of Google. Here's a look on the Amazon page, and it looks a lot like Yandex or Google. Amazon a becoming a retail search engine. Google, the ultimate metasearch, is getting into the booking game.
So, search engines become retailers, and marketplaces become search engines.
My message to you is what we are observing that the difference between those businesses will disappear.
Why this convergence is happening? It' is all is driven by the user. Users don't care what is their search engine or the transaction engine. They want information and they don't care how we make money. They just want answers to their questions.
Before, the user will show up in Google, search for hotels in London. Maybe visit flight retailer first, few people will convert, then go back to Google. So basically all users always had to go through the gatekeeper, Google. That's how Google made billions of dollars. Even today, major players such as Booking Holdings and Expedia, spend every year 5 billion dollars each in Google.
Not a great structure of our industry. So, what we can do?
I think that the answer is that you can let Google or Yandex get between you and the customer, and suck most of the profits from our industry, or you can become a Google yourself. That sounds quite crazy, how you can become Google or Amazon?
It is about using the data that you have about your customers. If you use your data correctly, you can use this data and deliver the best experience. To understand when it is best to be Google, and when you have to be Amazon. With the data you can make this decisions in real time, so you can vary between being Amazon model and Google model, and deliver the most relevant experience.
If you are willing to cooperate with your competitors, which is hard, but if you want to sell them a user from your commerce property to their property instead of going to Google, that would feel terrible. But you realize that the only alternative is letting your users to find your competitor through a search engine, which is worse.
Becoming Google is really about data, because you have to understand your customers to tailor the best experience. Data is power and the fundamental asset. When you understand that, you may end up behaving very differently. Once you are using you data, use to know your customer, to predict what they want to do, to tailor their experience, and on the spectrum between Google to Amazon that kind of tailoring can make your experience much more relevant.
I'm going to share some predictions that are possible. On this graph you see all of your traffic, from 0 to 100%. This traffic is sorted by Intent Media predictive algorithm based on the average from the billions of searches that we see each month. On the bottom is our prediction and on top are actual observations.
So the average unsorted user will be a flat line with a line of 5% on average. It would mean that in each of these user groups are just randomly put together. That's a model that doesn't work, and yet most of us thing of our businesses this way. If you are in the conversion business, and someone ask you what is your conversion rate, you'll probably say 5% or 7% or 9%. You never say "it depends".
My message to you is what we are observing that the difference between those businesses will disappear.
Why this convergence is happening? It' is all is driven by the user. Users don't care what is their search engine or the transaction engine. They want information and they don't care how we make money. They just want answers to their questions.
Before, the user will show up in Google, search for hotels in London. Maybe visit flight retailer first, few people will convert, then go back to Google. So basically all users always had to go through the gatekeeper, Google. That's how Google made billions of dollars. Even today, major players such as Booking Holdings and Expedia, spend every year 5 billion dollars each in Google.
Not a great structure of our industry. So, what we can do?
I think that the answer is that you can let Google or Yandex get between you and the customer, and suck most of the profits from our industry, or you can become a Google yourself. That sounds quite crazy, how you can become Google or Amazon?
It is about using the data that you have about your customers. If you use your data correctly, you can use this data and deliver the best experience. To understand when it is best to be Google, and when you have to be Amazon. With the data you can make this decisions in real time, so you can vary between being Amazon model and Google model, and deliver the most relevant experience.
If you are willing to cooperate with your competitors, which is hard, but if you want to sell them a user from your commerce property to their property instead of going to Google, that would feel terrible. But you realize that the only alternative is letting your users to find your competitor through a search engine, which is worse.
Becoming Google is really about data, because you have to understand your customers to tailor the best experience. Data is power and the fundamental asset. When you understand that, you may end up behaving very differently. Once you are using you data, use to know your customer, to predict what they want to do, to tailor their experience, and on the spectrum between Google to Amazon that kind of tailoring can make your experience much more relevant.
I'm going to share some predictions that are possible. On this graph you see all of your traffic, from 0 to 100%. This traffic is sorted by Intent Media predictive algorithm based on the average from the billions of searches that we see each month. On the bottom is our prediction and on top are actual observations.
So the average unsorted user will be a flat line with a line of 5% on average. It would mean that in each of these user groups are just randomly put together. That's a model that doesn't work, and yet most of us thing of our businesses this way. If you are in the conversion business, and someone ask you what is your conversion rate, you'll probably say 5% or 7% or 9%. You never say "it depends".
But in fact, it very much depends on what kind of user we are talking about.
Now let's look at the same users, de-averaged. This is the predicted model used our model. Your top-20 or 30% of users are the most valuable with the highest conversion rate. In the same time, lower part shows users with lowest chances to convert.
Now let's look at the same users, de-averaged. This is the predicted model used our model. Your top-20 or 30% of users are the most valuable with the highest conversion rate. In the same time, lower part shows users with lowest chances to convert.
Think about it as a call to action. If I would know it in advance, would I do anything differently? Would I change my experience for users? And the answer is yes.
This is the bounce rate. People that show up on your property and don't engage at all. 80% will probably bounce, and the conversion rate will be much lower. What interesting is this can be predicted in advance. When you understand that these people are arriving, and 81% of them would just leave, would you do something differently?
This is the bounce rate. People that show up on your property and don't engage at all. 80% will probably bounce, and the conversion rate will be much lower. What interesting is this can be predicted in advance. When you understand that these people are arriving, and 81% of them would just leave, would you do something differently?
Because we are the network, we know that in lower segments 75% of people will show up at your competitor's site at the same day using Google. So if you know that they are going to compare searches, would you do something differently?
In the same time, top segments are very loyal, 72% aren't going to other websites, would you do something differently for them, from the average?
This is the paid traffic share. Based on this traffic groups, who are the people that arrive through the paid traffic? This is very discouraging, but those users who show up and bounce most probably come from paid channels. So you pay to attract people who are least likely to buy something. So there's a very high chance of negative value traffic. So when you know that your bottom 10% of traffic is almost five times more likely to come from the paid source, would you change your marketing strategy and rethink the channels that you use?
In the same time, top segments are very loyal, 72% aren't going to other websites, would you do something differently for them, from the average?
This is the paid traffic share. Based on this traffic groups, who are the people that arrive through the paid traffic? This is very discouraging, but those users who show up and bounce most probably come from paid channels. So you pay to attract people who are least likely to buy something. So there's a very high chance of negative value traffic. So when you know that your bottom 10% of traffic is almost five times more likely to come from the paid source, would you change your marketing strategy and rethink the channels that you use?
When you start to think about aggregating these kinds of insights, and there are many many more of them, you would think to provide very different experience for users. For these users who are the most loyal, if you use the ad product, it is a big mistake. For those bookers, you'll probably want to get out of the way and let them finish the transaction. In the same time, for people in the bottom who won't convert and are very likely to come from the paid source, you might want to help them to get where they are going next. If they go to the competitor, they can go through Google, or they can do it through you.
That's what Expedia does with us, and that sounds a bit crazy. When you go to flight search on Expedia, you can see the Priceline ad. Priceline are their rivals, Booking Holdings. They spend days trying to understand how to kill each other. And yet, here's the Priceline's brand here. All because of Intent Media predictive model, because the user will go there anyway, and that makes a very significant portion of profitability for Expedia globally.
That's what Expedia does with us, and that sounds a bit crazy. When you go to flight search on Expedia, you can see the Priceline ad. Priceline are their rivals, Booking Holdings. They spend days trying to understand how to kill each other. And yet, here's the Priceline's brand here. All because of Intent Media predictive model, because the user will go there anyway, and that makes a very significant portion of profitability for Expedia globally.
The CEO of Expedia said that more and more revenues look like media revenues. CEO of Lastminute said that they consider themselves as a media company.
Important thing is that you need to get the data to run the predictions to figure out how your experiences should look like. That's hard. Couple of things to look for if you want to partner to do that. Make sure that your partner is good at machine learning and that you can make the predictions in real time, because the users that you will gather are constantly changing, you need to make the right decisions. And you have to focus on experience, so in Intent we have the strong belief that money follows the best experience, and if you get the experience right, all else will follow.
If you want to get more notifications about new articles on Travel Tech Conference Russia, subscribe to our newsletter.
Important thing is that you need to get the data to run the predictions to figure out how your experiences should look like. That's hard. Couple of things to look for if you want to partner to do that. Make sure that your partner is good at machine learning and that you can make the predictions in real time, because the users that you will gather are constantly changing, you need to make the right decisions. And you have to focus on experience, so in Intent we have the strong belief that money follows the best experience, and if you get the experience right, all else will follow.
If you want to get more notifications about new articles on Travel Tech Conference Russia, subscribe to our newsletter.
© 2018 Travel Tech Conference Russia