Generate super cool landing pages with an image recognition API? (For OTA’s and Meta search sites)
Lower your cost of acquisition.
In this post I am going to show you how to make quality landing pages based on image features in and around the hotel that will rank and convert.
First of all we have to make some data preparations.
- Your images. Do you have all images in one database
- Do all your images in that database relate to the hotel?
If the above is a yes you’re good to go.
Before we are going to use the API straight away lets play around with the demo a bit. You can go here for the fancy one or here for the fast one. Doesn’t matter which one they both have the same output.
We trained a ton of hotel specific labels. In 2012 I lived in Bangkok for 6 months and on thing I loved there are the rooftop bars. Think about Sirocco, famous from the Hangover movie, or Above 11 in Sukhumvit…. you guess it, 11. So we have the Rooftop bar label, just as the Rooftop pool label and we will use this in the example below.
You want to see all labels? Check them out here. But you can also run a batch of random hotel images and see what labels are popping up to give an idea.
Time to make some landing pages! And let’s start with Rooftop Bars & Pools!
Now we go to the Cliqorange API at dev.cliqorange.com and sign up for the free tier. Get your API key and we are almost ready to start the magic.
OKay, let’s say you are an Online Travel Agency (OTA) Prepare your batch of images from one city or country with the Python code on Github [link] Run the “experiences” label because that’s where the rooftops are.
Yeah, let our GPU cluster do some hardcore work out! And…. here are the results. Rank ‘m from high to low and there you have your rooftops. No go to your copy writer and ask him to write an awesome text about the rooftops in that city you just run. Woohoo! You just created an awesome rooftop landing page of all rooftops in City X.
That means traffic. Because now you can show your users where they can find the best and most awesome rooftops! Do I hear more traffic? Okay let’s do “spacious rooms” The same process but then with our image label “spacious rooms” In all locations! More traffic, more bookings, what a great tool!
x thousands of labels to go, ready to do the rest of the world? Get your free API key here.
its one strategy to have every landing page generates traffic above a certain ratio of search volume of keywords in the page. And if not the page has to be deleted.
The other strategy is to have as many landing pages as possible and scrape as much traffic as you can. But there are loads of arguments against this last one as the danger of generating duplicate not unique content.