I'm forever looking for that much-anticipated point where a "data-driven culture" turns into a business result. Fortunately, when it comes to ROI from a personalized customer experience, the proof points are a bit easier to find.
Even so, it's still pretty rare to be approached by a customer with "CX" project results to talk about. Adobe caught my eye with this one:
Holland America Line realized they needed to drive more traffic to their deposited bookings. The result? Using Adobe Target and Adobe Analytics Improved retention for online bookings by 5%, resulting in seven-figure ROI.
Then there was audience targeting:
Using Adobe Audience Manager, Holland suppressed ads for people who had already booked a cruise so they could focus solely on new customers. The result? 24 percent savings in one of their largest display advertising campaigns.
Good stuff - but the danger of CX is that you succeed in one project, but miss the need for a more comprehensive change. The cruise industry is another where complacency is high risk:
Customer experience expectations are an all-time high across industries, but this is especially true for the cruise industry, where an open bar and friendly room service will no longer cut it.
"Digital analytics is about serving our guests better"
So how do you get ahead of that? To find out, I hopped on the phone with Aaron Fossum, Director, Digital Analytics with Holland America Line. When I asked Fossum what drives the Holland America Line vision, he didn't talk room service and open bars:
Holland America exists to make connections with people, to open minds, build connections and inspire shared humanity. It's the overall vision that our president, Orlando Ashford, has cast.
How does Fossum's analytics team support that?
Digital analytics is really all about knowing our guests better, so we can serve them better. All the signals that a guest is giving us; are they having a good time when they cruise with us? What are they looking for when they want to cruise? How are we doing in terms of preparing them for the cruise?
I usually see a disconnect between corporate mission and departmental functions. But Fossum ties it together:
We see at scale all the signals that our customers are giving us. Our job, then, is to put that together, so that we can open minds and inspire shared humanity, like we're trying to do. And if we do a good job of that, we'll sell more cruises.
Makes sense - but how does that translate into day-to-day? As Digital Analytics Lead for Holland America Lines, Fossum sits in the integrated marketing team. His team is responsible for reporting out on their website's performance. But the tougher job is making the right recommendations on how to improve results.
Fossum didn't make the Adobe decision - that choice was already made when he joined Holland America Line in December 2016. But Fossum's past experience with Adobe products was a big reason why he came on board. Soon he was knee deep in the Adobe CX stack, from Experience Manager to Audience Manager to Adobe Target. The goal? Take that integrated view of their guests via the Adobe Experience Cloud, and prove that it can make a difference for customers.
Fossum's purview included another key job: modernize the web experience. When you consider that a web site now influences seventy percent of cruise decisions, that's pretty high stakes. That doesn't necessarily mean travel agents are now out of the loop. The best mix is helping all the channels work together. For Fossum, this means taking a more "programmatic, scientific, and disciplined approach to digital marketing."
Yes, that means breaking down the customer data silos:
I didn't know much about what pre-existed, but tools were a little less integrated with each other, and a little bit more siloed off. There was a desire to bring the data together, and put it into a more cohesive story that we could optimize against.
The current analytics state - on pops, drops, and course corrections
Holland America Line has made big progress on Adobe since they went live in 2016. Fossum shared their "present state."
We have a weekly dashboard that we produce where we bring together all the signals. And it's not just the digital signals that come from Adobe. We also bring in some CRM data and some call center data. We review that on a weekly basis, and look for insights. And primarily we're looking for what I call pops and drops.
Pops and drops? Those are data tidbits that lead to actions. Some examples:
- Was there a particular marketing channel that had an increase in traffic, or a particular page that was associated with more bookings than it usually is?
- Or was there something that started to not work?
- Are we seeing any notable visitor trends?
That analysis gives Fossum's team their week-to-week priorities.
Based on the pops and drops analysis that happens at the beginning of the week, we'll start to dig in, and answer the questions that were raised by that evaluation.
The weekly analysis functions as a kind of course correction - and balance to the longer term priorities.
We have a programmatic approach to optimization that's a little more long wave, and not quite as tactical, where we've identified parts of our site that maybe guests are struggling with, or that we feel could be optimized better. So we'll bring together a team, and undertake a series of activities that start with research and analysis, and then move on to designing enhancements that we can make, and then test those enhancements.
Fossum's team has one more main role: supporting ad hoc analytics.
The third big piece we spend a lot of time on is just ad hoc questions. Somebody has a question about something they're trying to do, and they'll come in and ask us to support that decision. So between the ad hoc questions, and the weekly, more tactical view of how the website is performing in course corrections, and then the long term optimization wishes, that's really how we spend our time.
Acting on Adobe data discoveries
As for those weekly dashboard sessions, what kinds of actionable discoveries have they made? Fossum told me about a new site Holland America launched in June of 2018:
One emphasis of that website was to improve our e-commerce outcomes, make it a little bit easier to do business with us, in addition to finding the information about cruises that you want.
Soon after launch, they noticed a decrease in "courtesy holds."
We saw a lot of the online behavior shift from what we call a courtesy hold, which is where you can say, "Hey, I'm interested in taking a cruise, please hold a room for me for five days."
That meant more booking deposits:
We saw shifts in that courtesy hold activity to deposited bookings, which are good, because they retain better. And so we saw an improvement to website performance as a result.
The bump in deposited bookings was nice, but Fossum's team really wanted more deposits and more courtesy holds. What to do?
So, using Adobe's platform, Holland America ran A/B tests to target an increase in courtesy holds. One A/B test was a simple change of text from a heavily-commercialized "Book Now" button, to a "continue this conversation about having a cruise" approach. That test did the trick:
As a result, we saw a 56 percent increase in guests going from that cruise details page to the next step of the booking process. And at the end of the day, we saw a 42 percent increase in the number of people who had actually made that courtesy hold, after viewing the cruise detail page, which is a massive lift. And we did that without having any detrimental impact to our paid bookings. So we were pretty pleased with it.
There are big projects still ahead for Fossum and team. One biggie? A/B testing is powerful, but it's not the same as true personalization. As Fossum said to me:
We're also looking to Adobe to help us with the site personalization. The A/B that I just talked about is a change that we make for everybody, but it doesn't get us to that dream of personalization.
Now we're going into Adobe Audience Manager combined with Target, to power some personalization. So that when somebody comes, we think, "Hey, this person's a good fit for Alaska." We show them Alaska. And we think this person's a good fit for the Caribbean, so we show them the Caribbean.
I'm a big believer in "right time" data. If weekly analysis gets the job done, then the distractions of hourly or daily updates might be just that - a distraction. But isn't Fossum's team thinking about adding in more real-time alerts, and daily course corrections?
Fossum says it won't replace the weekly discipline, but they are now building and setting up alerts for key activities:
I just actually finished an analysis of what I call visit intent segments. So, by looking at what kinds of content people are viewing on the site, we can predict why they've come to our site. So if they're looking at a lot of, cruise search, they're searching cruises over and over again, we can determine that they're pretty seriously shopping. But they haven't entered our booking process yet, so maybe they're not ready to buy, right?
And so we can target a certain kind of messaging to those individuals... We've established some segments in the analytics, and they're behaving as we would expect. We're continuing to refine those segments.
At this point, readers are probably expecting me to extol the virtues of AI. But my takeaway from this use case is: once you trust the data and see the customer picture, now you can dig. Now you can test, investigate, and make changes. Sometimes that data investigation leads you to counter-intuitive places.
Fossum described one situation where they found, through testing, that visitors who reached a certain point and saw their itinerary, might drop out of a shopping process. But those who continued on were more likely to convert:
It was a very counter-intuitive indication, where we actually reduced engagement to the next step, but we actually improved the overall result.
That wasn't about "AI," but good old-fashioned testing and problem solving.
We ran the test in Adobe Target. But when we had the counter-intuitive result, we dug deeper by going back into Adobe Analytics, and looking at some of the different segments.
Now if I can just get Holland America Line to stop tempting me with pictures of Alaska when Las Vegas dominates my calendar...