I’ve been working with lots of hotel clients during my time in digital agencies. I’ve seen home grown hotel booking forms, 3rd party booking engines (aka OTAs) like Synxis, and I myself even built one web booking engines which directly interfaced with some backbone hotel reservation systems. In this post I want to share my thoughts on what kind of measurement is useful to understand so that we can optimize the booking engine performance.
An ordinary flow of booking engines
Regardless of all fancy designs, optimized placements of elements or steps, almost all booking engines I’ve seen include the following steps:
Availability Search > Guest information > Confirmation
For example, Four Seasons:
Peninsula Hotels:
In the first step, or sometimes, broken into multiple steps, guests would be allowed to select their desired period of stay (or select desired destination / hotel for hotel chains) and search for available rooms and packages. In this step there can be filters applied to help guest further fine tune their search results, for example, accessible room filter or promotional rates filter.
Once guests selected their desired dates, rooms and packages, they proceed to enter their guest information. That includes their contacts (e.g. names, mobile) and payment information. Sometimes you can specify additional requirements like early check in, airport pickup within this step or in another extra step.
Finally, once guests reviewed their selection and information, they can confirm the booking and arrive the confirmation screen. Usually it shows guests’ selection and price information, together with a confirmation number for guests to inquire the booking in the future.
What we have been tracking
Like when we’re analyzing shopping carts, we love to see the ‘funnel’ of the booking process that highlights where people drop out:
This is cool, but as a second thought:
- It’s actually not difficult to figure out the availability search step would be the step where most people drop out, even without data
- People usually would do price check to compare before they buy, not necessary mean that’s something wrong with your design
- Even if the funnel is telling you 80% of people drop out from the availability search, you just don’t know WHY
- So what we have here alone cannot lead us to something ‘actionable’ – that we can make a change to improve our performance
My recommended approach
Our goal for tracking is always to find learnings that lead us to further actions. For an online property like a booking engine, whether or not the guest converts based on lots of factors, including your UX, your acquisition channels, your wordings, your room prices and inclusions, or even the economy.
My recommendations here focused on 1 thing: optimize the process and remove as many barriers as possible.
Availability Search -are we delivering suitable search results to guest?
In this step what we can possibly help guest to proceed is by providing them with relevant search results based on their searching criteria. Just like when you’re using Google – if the search result is relevant to what you’re looking for, you are more likely to click on the links to further explore isn’t it?
To summarize, metrics that I care in this step:
- Time for search results to return
- Number of search results for rooms or packages, segmented by occupancy
- Number of search error or no availability scenarios
- Click through rate of search results
Based on the learning, actions that you can possibly take:
- Optimize search results return speed
- Can we optimize our search query? I have no benchmark here, but just imagine if you’re the guest, would you expect your search results to be returned in 20 s? Benchmark yourself with your competitors and see if you’re doing better or worse?
- Can we cache some of our search results?
- Make search results match guests’ expectation
- Do we return the right room types to them? Are we returning rooms based on right occupancy / promotions? Can we make more relevant promotions for majority of our guests (e.g. > 75% of your room searches are with occupancies > 2 people, maybe you will start thinking to add more family offers?)
- What type of errors are we always returning? If it’s room availability related, can we allow more rooms to be sold online? Or if it’s related to minimum number of stay nights, can we adjust the stay requirement? (although that has to align with your hotel room inventory policy)
- Do we return enough and comparable choices for our guests? Remembered I read the book Predicatively Irrational by Dan Ariely: people can make decisions more easily based on comparison, especially when the comparison is an easy one. If I just return 1 search result, you will find it hard to decide as you cannot value your selection; if I give you 100 results, you will probably feel overwhelmed; but if I can offer you several rooms or packages, with 1 of them obviously cheaper or includes something exclusively better than the others, you would probably be able to make a decision
Our key metrics here is the click through rate of the search results, we will now do whatever we found from above data or test whatever we believe to move the needle of the click through rate to our desired target.
Guest Information -are we stopping our guest from submitting the booking?
Similarly, we’re not doing anything particularly fancy on guest information step; our goal here is to allow guest to complete personal and payment information easily to avoid them from dropping out from the checkout process.
To summarize, metrics that I care in this step:
- Number of errors that blocks form submission; better by error type
- If you have a multiple-steps checkout process; the drop out rate of each sub-steps
Based on the data, actions that we can possibly take:
- Update UX of the step
- If there’s a sub-step that people drop out vigorously, can we drop the step if it’s not compulsory?
- If I found guests keep getting an error very frequently, can we actually skip that field?
- If we cannot skip a field, can we make amendments on that field (e.g. can we automatically detect credit card type?), or adding some validation logic to avoid guests from putting wrong information?
- Can we consider move the field to somewhere after the booking is made as supplementary field?
- Or can we allow guests to login so to avoid them from entering data to avoid errors?
- Update wordings of the step
- Again, if I found guests keep getting an error very frequently, will it be my instructions are not clear? Will it be the labels or hint is not obvious so that our guests missed it?
Our key metrics here is the task completion rate. We optimize the step by removing barriers like the various types of error to facilitate the guest to confirm the booking.
Of course following this step (or maybe before) you may have steps that allow guests to personalize their stay. You can for sure track the activities there but I would focus on the key steps that lead guests to make a booking as my top priority.
Summary
These are my thoughts on the things that we should be looking at for hotel booking engines. Of course there’re a lot more to look at: for example, which acquisition channels are bringing more conversions in the booking engine? And if I can segment my guests demographically, what are their age / gender / occupation / type of travel, etc, lots of segmentation that you can do to learn more about you guests. Again, what I’ve put here are more on removing barriers for booking completion, but really lots of factors can impact your number of bookings. Knowing your guests better should be the key factor to get more bookings. For example, maybe a majority of your guests at a certain age drop out, you may want to adjust wordings to suit their age?
Please note my above recommendations are not bounded by any analytics or tracking tool; you can do it in Google Analytcs, Omniture, Mixpanel, Heap Analytics, or even by studying the booking data you collected in the backbone hotel system.
So dear hoteliers, now it’s your turn. Do you think what I’ve recommended for tracking is good for you to understand your booking engine better? What’s your favourite metric that you’re using today? Is there any parameters / segments that you think is very important for you to improve your sales? I would really love to hear from all of you to make this piece more close to hotelier’s mind. Looking forward for your comments!