Machine Learning: How Google is Helping Hoteliers Get Smarter with Search
Last week, we were joined by Jennifer Wesley, Industry Director for Travel at Google, to discuss Smart SEO Strategies to Drive More Direct Bookings. She provided a fascinating insight into how the online travel shopping journey is evolving, how Google is adapting, and how hoteliers can capitalize on all the amazing technological advancements now on offer to get in front of the right people, at the right time, with the right message.
Below is just a snippet of Jennifer’s presentation, when she discussed changing consumer demands, the power of machine learning, and how hoteliers can capitalize on it all to deliver a superior brand experience.
“As marketers, we need to think differently about how we show up [for consumers]. This isn’t just about how consumers use devices. We’re really talking about how we develop experiences, and we have to make sure that every experience we develop is assistive to consumers.
[As consumers], we are looking for ways to get relevant, frictionless and useful experiences, whenever and wherever we are. As consumers we know this, but as marketers we haven’t been so good at delivering it.
By and large, if I talk to 50% of major hoteliers in the US, I’d say 10% of them are meeting consumer expectations. Very few are exceeding. When it comes to what brands are doing in the mobile space, we still have a lot of work to do to improve those experiences.
The good news is, we’re closer today than we were 2 years ago, or even 1 year, in being able to fully harness all of the information and data that is being extracted from our mobile usage.
Today, a person’s journey to book a travel reservation includes over 400 touch points, and 80% of those are on mobile devices. So it’s very hard to get a hold of where [a traveler] is, when [they] want something, and what message to deliver to [them].
We have made some incredible strides in machine learning that are allowing us to take advantage of this data at scale without necessarily having to intervene as humans. What we’re doing is taking all this rich intent data – which is people starting their travel searches on Google, surfing the web for information while they’re dreaming and planning their vacation – and combining that with the data we have around users. If you are signed in [to Google], we understand who you are, where you are, what’s in your Gmail. We can then combine that with incredible knowledge in machine learning to deliver assistive recommendations to consumers, and to marketers.
That’s probably the biggest change we’re seeing in the search space; the ability for marketers to let the machines do some of the work. You can put in your business goals, tell us what you want to achieve (profitability, number of bookings, etc.) and let us help you optimize your budgets against that using data about intent, the user, their behavior, their needs and their context, and then give you the right consumer at the right time with the right message.
We are very eager to help marketers use that data in the right ways. If you have first-party data, like a database of your most loyal consumers, we are working actively to help you match that up with our users in a very safe way [that protects consumer data]. We’re currently doing that with a lot of brands, where we allow them to upload their audience, combine that with our audience, and go out and find people who they know, or people who look like people they know, that they want to attract to their property.
An Example of Machine Learning
Google’s photos app is a great example of machine learning.
There are 1+ billion photos uploaded every day to Google Photos. In that app, there is something called Assistant. And Assistant will allow you to sort those photos, filter them, create an album, a video or just a collage of photos entirely based on the criteria in the image. So you could pick someone’s face and curate all the photos of that person. You could pick a location and quickly find all the photos you took there. But you could also put in something more generic, like an orchid and find all the pictures you’ve ever taken that had an orchid in it.
You can see the power of machine learning, even with your own small collection of photos. That’s the same technology that Google is using in its advertising algorithms today.
When you use AdWords or run a YouTube campaign, we’re using that same machine learning to optimize your campaigns; to understand who your users are so that we can predictively find the most valuable customers for you.
What Does Machine Learning Mean for Digital Marketing?
But to take full advantage of all these possibilities, we still have a lot of work to do. Consumers want brands that are helpful and stay one step ahead of them, but only 1/3 of brand experiences are described as helpful. To stay ahead of consumers, we need to really understand them and do some predictive analysis. Data and machine learning are part of it, but it’s not enough.
The power of context has changed marketing. If someone is close to a hotel, or even standing in it, we can now tell that. That means, if someone has found you in Google search on their phone, but they don’t click on your ad, but then show up in your hotel, we’re able to report on that. And that’s really a big change.
We’ve used this a lot in the retail space, but not a lot in the lodging space. But this technology is something absolutely critical for hotels that thrive on the last minute booking, where you often see the booking come through on a smartphone or a walk-in.
How Can Hotels Improve Their “Brand Experience”?
The brand experience is increasingly being influenced by personalization. When we think about personalization for hotels, we often think about the onsite experience. Most brands do really well at this. What we have to work on is the digital experience that’s happening before, during and after. Today there are 10x more digital touchpoints than there are on-property. But [hoteliers] spend the vast majority of their time focusing on the on-site experience.
Improving your brand experience starts with acknowledging that mobile is paramount. We’ve learned that the drop-off of users between 1 and 7 seconds of a mobile website loading is double. Mere seconds make a difference in whether someone stays on your mobile website or not. At Google, we’ve made some changes to how we rank our listings based on how ready they are for mobile.
How Can Hotels Take Advantage of Data to Improve the Travel Shopping Experience?
Joseph (our first traveler) is searching for the best rates on hotels and wants to compare the lowest prices from a trusted source. His search term is actually pretty close to the search term of his two cohorts (see image below). But because of the rich intent data that we have, we can actually tell that he wants to compare rates. And so you’ll see that the Priceline ad and the language of that ad is different. It says “compare cheap hotels and save up to 60%.”
Carey (our second traveler) is searching for the cheapest hotel. Like Joseph, she wants to book a hotel, but she’s also trying to understand how to save the most that she can. She’s not looking for a comparison; she’s just wants the cheapest. So when the ad shows up, we’re enabling advertisers to show exactly the right message. In this ad, the call to action is just the great prices Priceline has.
Aditi (our third traveler) is also searching for cheap hotels – using the same words as Carey. She too wants to save money, but she is also loyal to Priceline. We know, because of our ability to integrate first-party data from Priceline, that she’s a loyalist, so we’re able to deliver a specific message to her around best rate guarantee.
So, we have 3 people using the same language to search for hotels, but having a very different intent. But using data and machine learning, we can deliver a personalized advertising experience to drive up the conversion rate. We take intent, context, and identity and mix them together to deliver the right message to the right person at the right time.