Adobe EMEA’s Head of eCommerce candidly speaks about the 300+ optimisation experiments his team have run across the firm’s digital properties—two-thirds of which fail.
At the recent Adobe Customer Experience Forum in London, Adobe’s Head of eCommerce – EMEA, Petar Karafezov, pulled back the curtain on the trials and tribulations of optimising Adobe.com, detailed below, as well as the firm’s own five-year digital transformation (which our sister site Technology For Marketing has covered).
Most of Karafezov’s team’s optimisation experiments fail—possibly not enough of them, he ponders—and they have had to learn the hard way that there are many ways to accidentally kill a customer journey.
Focusing on micro conversions
Karafezov admits with somewhat of a weary sigh: “Looking back, it feels like it was easy [to reorganise the business around data-driven insights] but it wasn’t easy at all.
“Because we were a dominant market player in the creative space [with tools like InDesign and Photoshop], web customers were visiting Adobe.com every couple of years to get a new version. Our product portfolio was relatively simple and we would present, for instance, a new upgrade price—that was it.
“Things are different now because of the move to being a subscription business, which means a higher frequency of customer interactions. It means we have to focus more on ‘micro conversions’ and our success as a business is no longer based on just acquisition, but acquisition and retention of customers.”
“We’ve had to develop a cross-customer lifecycle programme of micro conversions with a focus on active use.” An example of a micro conversion in this context would be a customer that has signed up for a free Creative Cloud trial but not downloaded the software. “We broke it down deliberately, what is the next step? What is the next thing after someone downloads? We want to make [the customer] successful. We want to make sure that we give them specific content that enables them to achieve the objective of the trial.”
The team introduced an experience specific to those customers, aimed at giving them a reason to migrate. Going one level deeper, they realised they also didn’t want active customers, returning to Adobe.com, to see the same content over and over. “It looks stupid if you see same message again and again, so we started to evolve these micro conversion steps into a conversation. Instead of sending the same message, we deliver a sequence of messages that feed off of each other—hoping the customer is seeing it, of course!”
The retention game
This micro conversion approach makes a difference, according to Karafezov, for example in migrating customers in Germany.
“Germany is a challenging market for us in moving towards subscriptions. We had a specific ‘migration journey’, where we had self-selection touchpoints for them.” In practice, a site visitor could identify themselves as a customer of a particular configuration of software, and see specifically designed web pages in accordance. “That worked great but we noticed that they were coming back and hadn’t converted yet.” Taking a closer look, it turned out these customers didn’t want to go through the self-selection process again, necessitating a different set of content for return visitors: “When you’re coming for the third time, to the home page or specific product pages, [we’ll show content that talks] about specific product benefits.
“It was a challenge. We were not used to working with people who [we knew] were paying customers. They didn’t subscribe to us before yet now we have 10 million subscribers—three million are in EMEA. And we have to keep them because my team is compensated—and company growth is based on—net new revenue, which means that you’ve got to acquire and retain customers. [Returning, current customers] used to be a sliver of our visits. Now they represent 20% of all visits to Adobe.com. We decided to focus on retention and learn how to do this.”
Ecommerce case study: Assumptions about existing customers can lead to revenue loss
“We created an audience segment that was based on our paying customers. We [did it in this way] because we wanted to make sure that customers were addressable not only when they’re logged in to the site—i.e. when we explicitly know that they have paid membership—but also when they’re logged out, when they go outside of Adobe properties.”
The team also took a hard look at the question of why people were cancelling. “They were saying that the products are too difficult, it takes a while to learn, they’re too expensive.” This led to the decision to create a high quality ‘onboarding experience’, to ease customers into using various Adobe products.
Using insights to feed into the project, Karafezov and team looked at where visitors tended to visit, designing journeys for them: “If you come to the homepage, we have a specific entry point; if you’re coming from the [internal] Creative Cloud overview, we’re going to replace the whole Creative Cloud Adobe.com page content with an onboarding experience.” The assumption was that “there’s no need for us to tell you what the Creative Cloud is—you’re already a customer, so obviously we need to help you learn about this.”
The new setup was launched and left for a month and half. “It was great. All of our predicted scores went up. Retention would have improved by about 15% which was multi-million dollar revenue.”
As it turned out, the assumption about the need for onboarding led to an own-goal: “But we were losing money. How could we lose money from people that are paying customers and are not cancelling?” It turns out that only 35% of these paid members were people who needed onboarding. “The other 65% were people who were near the end of the experience and were there to buy a second application, for example, they had Photoshop and were thinking of buying InDesign.”
Karafezov is contrite:
“Because we didn’t think through carefully through the customer experience and customer segments, we didn’t go back. We didn’t follow our own process by deciding exactly what the business objectives were. We designed a customer journey and in the process, killed another one.”
“It was a lesson learned for us and now we’re refining it but I’m sure as you’re working through and implementing these customer journeys, you’re going to be dealing with your own set of challenges. My note would be, think carefully through these audience definitions and try to understand what is it that they want to do. Look at detailed analysis so that when you create your segmentation scores, that they are truly reflective of where the customers are in their lifecycle, in their interaction with the product and the brand.”
Poorly segmented apples and oranges
Prefacing his analysis of Adobe’s improved ability to define audiences, Karafezov points out that: “Customers will always take their own route that is the shortest way to where they need to be.” He treats the audience to a picture of Tottenham Green, where a perfectly respectable pavement has been ignored in favour of pedestrians cutting across the grass to reach a bus stop:
“People leave a trail—in our case, they leave a digital trail. When we look at customer insights, we’re using them not only to obtain a final analysis what happened [throughout their journey], but also to understand where customers went. This lets us continuously optimise the journey and improve micro conversions.”
Looking at the sorry state of audience definitions prior to Adobe’s internal data adjustments from 2012 to today, he points out: “Before we used [Adobe’s Audience Manager], it was a little bit apples and oranges. Every [team dealing with an individual channel] had their own journeys mapped out: the search marketing team were using a mapping page; the display advertising team was using [landing pages not part of] Adobe.com. There was really no conjoined profile.”
The teams moved to a ‘single source of truth’ for incoming customer data and outgoing messaging. “Audience Manager has allowed us to do this continuously, it hasn’t betrayed us.”
Customer insights don’t have to be big
The ideal, says Karafezov, is to use insights to be able to identify opportunities to makes customers’ lives better and thus make more money.
A ‘customer insight’ can be a very small nugget, he points out. “When we were dealing with some of the challenges in terms of helping some of our German customers migrate, we saw that they were a lot more likely, in comparison to the UK, to seek self-selection. They really didn’t want to be guided into the next step—they wanted to choose for themselves.
“We did a super simple test with some self-selection options on the lobby page and improved revenue by $1 million”:
Boasts Karafezov: “A million here, a million there, you gain credibility in terms of actionable insights, building the business case, bringing it to market and iterating on it. This is key in being able to source funds for your next insight, your next evolution. That insight could be as small as ‘people like self-selection’ in a particular market. It doesn’t have to be this big thing.”
Adobe now has targeting across all of its home pages which are collectively a valuable asset, contributing to a third of orders:
Balancing targeting with corporate strategy
This thorough targeting strategy needs to be balanced against corporate priorities, Karafezov counsels. The team got greedy at first: “We wanted to target everything—first and second visits, student and business segments etc. This introduced the need for a lot content and our creative team was starting to push back.
“When you have a campaign, you need to produce X number of additional images, you have to localise them in X number of languages and finding the balance between how much you target versus the cost of implementation—well, that’s still an ongoing thing for us. Being able to understand which of the targeted home pages is worthwhile keeping has [required prioritisation of specific markets] in EMEA.” For example, certain ‘Tier 1’ markets enjoy comprehensive targeting, while others are stripped back, seeing a different level of targeting where the cost of implementation doesn’t stack up against the potentially lower revenue impact.
Karafezov stresses that it they had to understand the challenges that each new insight and personalisation opportunity could create—and calculate whether the rest of the organisation, including creative teams working across multiple touchpoints in multiple markets, could cope.
Ecommerce case study: Targeting a segment of business customers across different markets
Karafezov explains that the team took a specific business segment and rewired touchpoints including Adobe.com. Instead of delivering the generic ‘Make it. Creative Cloud’ messaging, visitors were delivered business audience-specific messages across the site:
“This brought us an estimated $3m in additional revenue but it didn’t work as well everywhere—in fact, it didn’t work at all
in Germany! We used the same audience segment, the same approach. We defined the business objectives, then the audience. We deployed it and we just didn’t see the same results.
“Why? With [audience segmentation], there are a number of attributes that make up the definition. In the German market, ‘unternehmen’ could mean a small business or it could mean a large business, so some of these customers were not equally qualified in comparison with the UK when there was a very clear definition between ‘SMB’ and ‘enterprise’. We had to redefine all of our segments and have a regional adjustment, to achieve better segmentation.
“As you’re preparing for deploying a single customer view and using it, especially if it’s done across different markets, my advice would be to be ready and think about how you can augment these segments. Because, when you put it into practice, it ends up being slightly different across different markets.”
If you get addicted to testing, get comfy with failure
Some of Adobe’s strategic areas of focus when it comes to optimisation
Enthuses Karafezov: “The more insights you generate, the more you want to try them out, take them to the market and explore these opportunities.” He claims that they have run 300 tests on Adobe.com and across digital properties, generating an estimated $40 million boost in revenue from these optimisation activities. “And we’re just getting started. There are a lot more opportunities to build on what we’ve done so far.
“A third of [our ecommerce website] tests are successful—conversely, two thirds of them fail. It’s a dirty business, a business of failure, so be ready for that. You generate an insight and then you have to put it in practice because if you don’t test it, it doesn’t matter.
“Sometimes I ask myself whether we challenge ourselves enough. Should that ratio [one third of tests are successful] be one quarter? Would other businesses see that as a good ratio? Maybe we’re not as smart as we think and we we need to just create even better experiences!”
Check out Petar Karafezov’s explanation of Adobe’s digital transformation from 2012 to 2017 over at Technology For Marketing.
Three final tips: Risk, transformation and creativity
Karafezov sums up with three pieces of advice for marketers working around ecommerce websites:
Not taking a risk is a risk in itself
“It’s much easier to take a risk these days because risks are measurable and you’re much more in control—it’s not a leap of faith. We understand what actually happens across the channels and if we have the discipline to analyse customer interactions with proper organisational support, processes and technology, then it’s not as scary. But not doing it is worse than actually failing and learning from it.”
Organisations have to transform
“We now feedback customer-centric insights, generated by the digital marketing team, to our creative teams.” What they get back is pleas for help: “Everyday we hear ‘tell me which one works better.’”
“Let’s redefine our creative process so that when we say it’s a ‘design organisation’, it’s not only creative design but customer-centric design. Creatives and marketers have never been closer than before in Adobe. We have to be ready to enable these organisational challenges, interactions between [teams], getting more… data insights.”
We invest a lot in technology but we’re not robots
“Technology will never be able to solve our challenges. It comes down to people—change comes from within. It takes you, I and everybody, our teams, to be galvanised to actually make a difference. Creativity is still king. Focus on the customers—that’s still the most important thing, regardless of the tools and processes. That’s what’s been powering our success—the people within Adobe.”
The post ‘Two-thirds of our optimisation tests fail, it’s a dirty business’ – Adobe EMEA’s Head of eCommerce appeared first on eCommerce Insights.