The past few weeks on LinkedIn have seen some stimulating
debate among us travel nerds on the topic of personalization and why it’s
such a perplexing challenge in our industry.
For a market with so many user profile, purchase and
behavior attributes to mine, the issue certainly isn’t lack of data. Particularly for the
larger, established online travel agencies, airlines and accommodation providers, their data lakes
are incredibly well‐stocked.
With the inexpensive computing power and artificial intelligence/machine learning tools
available today, providers can fish for customers in those data lakes far more
efficiently than ever before. So it’s not a process issue.
Why, then, am I getting Motel 6 ads in my Hotels.com
results? And why does another traveler get an email titled “Travel recommendations just for you” featuring
airfares to places he’s never expressed interest in from an airline he hasn’t
flown with in five years?
Subscribe to our newsletter below
Why does it take me at least three to four clicks to filter
results on every search I do on Kayak or Expedia to curate what I want to see
(4‐5 star, no apartments/motels/B&Bs/hostels/capsule hotels, etc.) –
despite having a decades‐long profile of shopping and booking with both sites?
They should know me well enough by now. Instead, they’re like that guy from the
movie "Memento" who can’t remember anything.
Every time I search, I’m like a brand-new person to them.
The truth is these sites have made an intentional business
decision to NOT use their talent and resources to provide a more personalized
experience for their users. They have all the brainpower, data and technology
to do so, but they choose not to.
The only logical explanation for this is that
these travel providers believe they make more money being UN‐personalized than
they would if they invested the resources to make a truly curated experience
for their users. Expedia explicitly acknowledges this:
That seems really bizarre to me in this day and age,
especially considering every other service provider I use has aggressively
moved to a more personalized experience.
Amazon knows what I need before I do. Netflix and Spotify
are pitch‐perfect with both their default recommendations (what they know I
like) and their “discovery” suggestions (new stuff they suspect I will like).
StitchFix has done an incredible job leveraging data science
that is layered on top of purchase, feedback and gamification inputs to create
a hyper‐personalized personal stylist experience so that I never had to set
foot in a mall again (check out their Algorithms Tour!). Virtually everything
they present to me is stuff I consider “buyable.” I can’t say the same for the
The end result is that I am an incredibly sticky customer
for these providers. In the case of Amazon, Netflix and Spotify, they bill me
every month and I pay without thinking about it. With StitchFix, they are the
first (and usually last) place I go when I need something new to wear, so their
share of my wallet is close to 100%.
Meanwhile, the OTAs and other travel suppliers continue to
serve up an undifferentiated, mass‐market, one‐size‐fits‐all user experience,
expecting us users to do all the manual labor to sort and filter and click our
way into “personalized” results. That’s a lot of unpaid labor! Maybe we should
How it SHOULD work
A better approach makes the product do the work – not the
Investing in a product experience that makes use of all the
data customers provide, either explicitly via profiles and preferences, or
implicitly based on their purchases and actions, leads to a far more seamless and
efficient shopping experience for users, and a shorter purchase funnel for
- Let users define their own default for how results are
filtered for display (e.g., star‐rating, property type, non‐stops,
- Give users the option to save different searches for
one‐click shopping (like Zillow does).
- Enable multiple personas that curate results accordingly
with a single click, based on what the user defines for that persona (e.g.,
traveling on business, with kids, with my significant other, with my friends).
- Create a much richer customer profile by asking users about
their preferences (like StitchFix’s Style Shuffle). It doesn’t have to be a
gating factor to sign up or an IQ test to use; when done right, it builds the
profile bit by bit over time, using gamification or other quick prompts.
Investing in the product experience to make it much stickier
and more engaging for users makes for a much more loyal user base in the long
run. Anyone who has spent time crafting their playlists on Spotify knows what
I’m talking about … zero chance I’ll go through that again with someone else!
I’m sure I’m not alone in wanting to see someone finally
tackle this personalization challenge in a truly customer‐first way, using the
modern technology now available. With all the brainpower in our industry, from
the OTAs to travel suppliers to the startups looking for a legitimately
disruptive business idea, it shouldn’t take long.