Data foundations for effective travel marketing in the age of AI

By Caroline Routledge, Associate Director and Travel Acquisition Specialist, PDV

Travel marketing is getting harder. Rising acquisition costs, more complex customer journeys and the rapid arrival of AI tools are putting pressure on marketing teams at every level. The good news is that most of the problems are solvable and the solutions tend to start in the same place – better use of the data you already have.

At the 2026 ABTA Travel Marketing Conference, my colleague, Scott Logie, and I explored three areas where we see travel businesses consistently leaving value on the table. None of them require enormous investment. All of them require an honest look at how data is currently being used.

Understanding your travellers

Most travel marketing is still built on demographics. Age, household income, postcode – these are the variables that underpin the majority of targeting decisions in the sector. The problem is that demographics explain very little about why someone books, or when, or what will tip them from browsing into committing.

The data that actually predicts booking behaviour is behavioural – search patterns, site navigation, email engagement, previous booking history. A 58-year-old and 34-year-old can display identical booking intent signals. Treating them differently based purely on age means you are optimising for the wrong variable.

Better segmentation combines three things:

1. Behavioural signals – what someone is actively searching for and engaging with right now. 
2. Journey stage – whether someone is in the dreaming phase, actively planning, ready to book or a returning customer who already knows you.
3. Value-based segmentation – that looks at future revenue potential rather than past spend alone. A first-time booker with strong behavioural signals can be worth more over time than a lapsed customer who has a single high-value transaction in their history.

Also, true personalisation goes further than a first name in an email subject line. It means changing the message, timing and channel based on how someone is really behaving. A traveller who has visited a cruise destination page three times in a week is telling you something. The question is whether your segmentation is set up to hear it.

The payoff from getting this right is tangible – higher conversion rates, lower acquisition costs and a better experience for the customer. These outcomes are not a trade-off. They follow from the same underlying discipline.

Proving your marketing works

Last-click attribution remains the default measurement approach for most travel businesses. It is the simplest option, but it only tells you part of the story.

When a customer sees a direct mail piece, searches for your brand a week later, opens an email and then books via a paid search ad, last-click tells you the paid search ad did the work. It does not tell you that the direct mail prompted the initial search or that the email kept the brand front of mind. Each channel claims the same booking; the actual contribution of each one is invisible.

This matters because it drives poor budget decisions. Channels that do the groundwork but rarely convert at the last step get cut. Channels that are good at closing but depend on other channels to generate intent get over-invested in. Over time, efficiency falls and acquisition costs rise.

The solution is not necessarily a full econometric modelling project, though media mix modelling has its place. A good starting point is to group channels by the role they play, whether that is building awareness, driving consideration or converting intent, and measure them accordingly – multi-touch attribution. Even simpler, run a hold-out test, otherwise known as incrementality testing. Take a comparable segment of your audience, withhold one channel or campaign from them and compare outcomes. It is one of the clearest ways to understand what is genuinely driving bookings versus what is just present when bookings happen.

Multi-touch attribution and incrementality testing are not exotic analytics capabilities. They are practical tools that give marketing teams the evidence they need to allocate budgets more confidently and defend spend to finance teams.
 

Getting ready for AI

AI has been generating significant interest across the travel sector for a while now. Predictive targeting, personalised messaging at scale, customer lifetime value (CLV) modelling, smarter media optimisation – the use cases are real and the results are already being demonstrated by early movers.

But AI is only as good as the data behind it. Feed a predictive model fragmented, inconsistent or duplicate-riddled data and it will produce unreliable outputs, with a great deal of confidence. The technology does not fix a data problem, it amplifies it.

The most common data issues we see in travel businesses are: 

• Siloed systems that do not communicate with each other.
• Duplicate customer records created across booking platforms and communication tools.
• Inconsistent data structures that make it impossible to build a coherent view of the same individual across touchpoints.

An AI-ready data foundation does not have to mean a full CRM rebuild. A useful and often overlooked starting point is a tactical single customer view – a unified record built for a specific purpose, such as identifying lapsed high-value customers or scoring booking intent across your active database. It does not need to be comprehensive to be useful. It needs to be clean, consistent and connected to the right data sources. 

Once that foundation is in place, even relatively straightforward AI applications, such as propensity models that identify who is most likely to book next or CLV models that prioritise acquisition spend on the most valuable prospects, can deliver meaningful improvements in efficiency and revenue.

The broader risk of standing still is worth naming plainly. As AI becomes more central to how competitors target, personalise and optimise, the gap between businesses with clean data and those without will widen. The cost of inaction is not staying where you are, it is falling behind at an accelerating rate.

Where to start

If any of these three areas resonates, the practical question is where to begin. Our suggestion is to start with an honest audit of your current data – what you have, where it lives and how reliably it reflects the same customer across different systems. That audit will tell you more about where your highest-value opportunity lies than any external benchmark.

The businesses making the most progress are not necessarily those with the biggest budgets. They are the ones asking the right questions of the data they already have.

Caroline Routledge is Associate Director and Travel Acquisition Specialist at PDV. For further reading, visit the PDV blog.