Leveraging Booking Intent & Search Data to Understand Your Customers
By Marc Young
Account Manager
Meet Marc, one of our Marketing Account Managers here at Umi Digital. Marc has a passion for web development and design, taking pride in our clients websites and SEO work to ensure performance is top tier. With a huge amount of experience and interest in data, using it to make educated decisions and further develop strategies. Fun fact, Marc loves Jujitsu - playing and watching!!
If you want your hotel’s marketing campaigns to land at exactly the right moment, you’ve got to understand how and when your guests are searching. It’s not just about knowing who they are, it’s about knowing when they’re ready to book so you can align marketing and advertising efforts.
A bit about me If you haven’t met me yet, my role at Umi is to bridge the gap between the marketing and development teams as a technical account manager. I joined Umi following a career in software development but wanted to become more client-facing and more deeply involved in the marketing. Alongside the management of the more technical accounts, I work to create internal tools to make our lives and our value proposition even better!
At Umi, we derive insight from a variety of different places, but a recent activity of mine has been to dive deeply into booking intent by looking at user searches. There are typically ten times more searches than actual bookings so, even the smaller websites can generate meaningful data sets for us to work with.
By analysing search behaviour, broken down by day, time, week, and month, we can identify those peak moments when guests are most active online. This means our advertising campaigns can run smarter: showing ads at the right time, massively reducing wasted spend, and maximising CPA and ROAS.
This ties directly into the first type of analysis being lead-time analysis, which helps us understand how far in advance people are booking. Perhaps your guests start searching eight weeks before Christmas? If your campaign launches with 6 weeks to go, that’s two weeks of potential revenue lost. These insights make all the difference when planning seasonal activity or pacing out your ad spend across the year.
Turning this data into strategy
In real terms, and to make this actionable, we track key intent-based data points such as:
search time, arrival date, departure date, property_id, client_id (if logged in). This gives us a clear view of how guests move from dreaming to booking, allowing us to build campaigns that meet them at just the right stage.
When you combine those signals with hotel performance data and availability gaps, you start to see the full picture. Marketing is not just about who your guests are, but how they think and when they act. That’s the sweet spot for both advertising optimisation and conversion growth.
How we began automating the process
In the early days, Umi did this manually on our individual devices - We logged search and booking data into a database, exported it as CSVs, and then ran a series of Python scripts (specifically Pandas) for data cleaning and Matplotlib for visualisation to make sense of it all.
Once cleaned, the data was grouped by time, day, and month to reveal patterns in booking behaviour. From there, we could create heatmaps showing when guests were most likely to search, essentially a visual pulse of guest intent across the year.
These insights became gold dust and then fed into the advertising and marketing teams. It helped our hotels understand their customers better, and also allowed us to spot industry-wide patterns across multiple properties to give us a wider lens on how travel intent shifts over time.
From manual to automated (with a few tech hiccups along the way)
Manual analysis worked, but it was time-consuming. So, true to Umi form, we looked for a smarter, more automated way to do it.
The first step was moving the whole process to a hosted infrastructure so anyone on the team could upload data and instantly get insights. The idea was simple - make the tool our V1 tool accessible, fast, collaborative and secure.
But of course, there were a variety of of hurdles to overcome and as we started working in more of a hosted, JavaScript-based environment we needed to swap out certain python elements for tools like Chart.js and Danfo.js (the JavaScript equivalents of Matplotlib and Pandas), but the results weren’t quite as consistent or flexible as our Python setup.
Enter Docker and AWS Lambda
After a bit of brainstorming, we decided to Dockerise our Python scripts, containerising them so they run in the exact same environment every time, no matter who’s using them.
Next, we’re integrating the setup with AWS Lambda, allowing the scripts to run in a fully serverless environment. This means anyone at Umi can trigger the process without worrying about local dependencies, setup headaches, or compatibility issues.
In short: it’s consistent & scalable.
What this means for our clients
Once the system is fully up and running, any member of the Umi team will be able to securely import a booking tracker data and instantly receive beautiful, interactive visualisations of key booking intent data.
That means faster insights, smarter decisions, and ultimately, better results for our hotels.
Hotels will be able to see, at a glance, when and how their guests are searching, what patterns are emerging, and where they can focus marketing effort for maximum impact.
In a world where timing is everything, this kind of clarity gives hotels a huge competitive edge. It’s data-driven strategy made simple, and exactly the kind of thing we love building at Umi.
Final thought
We’re still refining the system, but even at this stage, it’s proving to be one of the most exciting internal tools we’ve built. It’s a great example of how data, automation, and a bit of creative problem-solving can come together to make hotel marketing more intelligent and efficient.
At the end of the day, it’s not just about collecting data, anyone can do that, it’s about turning it into insight and actions.