Building an Optimization and Personalization Program as a Leader in Financial Services
I recently had the opportunity to interview Mike Girard, Senior Manager of Testing and Optimization at a leading Fortune 100 financial services institution, about his experiences in that role, and for insights and advice that others in optimization and personalization could apply in their work. With a brand that provides retirement expertise and services a multitude of personas with individualized financial needs, personalization is critical to growing and retaining happy customers. Here are just a few highlights of that interview.
What’s your role at the company?
I am the Senior Manager of Testing and Optimization. I serve as the testing and personalization subject matter expert.
Can you give me a brief history of the brand’s use of optimization and personalization?
Our first A/B test happened in 2016. We launched 35 tests that year, 75 in 2017, and 120 in 2018. The complexity of activities progressed, from basic A/B tests of titles and subject lines to testing the product flow, basic personalization for campaigns, multivariate testing on a key product page with 36 variations and using Automated Personalization on the homepage.
What’s the general structure of the optimization program?
I am responsible for test setup, execution, and analysis. However, I work with all the different teams throughout the organization to develop a testing strategy for each group. I help them to clarify their test ideas, so their experiments answer the question they need answered. We also have a homepage personalization team and while I will work with everyone on their personalization strategy, we have only done very basic things with personalization so far. That is changing in 2019!
What are some of the significant activities or milestones for growing the program?
The biggest one was getting approval for using Adobe Target on the website in August of 2015. Then it was getting the first simple test live to show how Adobe Target works, and similarly, doing a first test in our secure “My Account” environment. We established a monthly A/B testing review for 20 people—now my invitation list is over 50! We also ran an Automated Personalization activity and benchmarked against rules-based personalization strategy in the hero banner. Structure-wise, we’re taking over the personalization piece in 2019, which will require an additional head count.
Can you describe the current culture around optimization and personalization?
It is always a work in progress. Some groups test everything, and others are not quite as mature yet. The organization believes experimentation is a critical activity we should be doing, which is great. Personalization is a key 2019 priority, and the teams are building out their personalization strategies as we speak.
How did you get executive support and buy-in?
Wow! I know everyone says you need executive buy-in to start a program, but I actually say you just need someone higher on the food chain that believes in it. It’s much easier fighting the battles when it is not you against the world. We started small and tested only little things. Once we got a few wins, we wouldn’t stop presenting the results in monthly review presentations to whoever would listen. It sort of snowballed from there, with a few VPs beginning to question why their teams were not doing this. The CMO also agreed to the Automated Personalization on the homepage and highlighted the results in his year-end presentation.
How does the program regularly socialize its value to the organization?
We give those monthly presentations to whomever will listen. It’s great seeing how many times the invitation is forwarded. We also occasionally have Adobe experts present new features or industry trends to keep it fresh.
[Editor’s Note: Check out Adobe Experience League for links to resources, value realization presentation templates, videos, and more!]
Tell me more about the recent activity in which you moved from random and rules-based personalization to using Automated Personalization.
We wanted to increase the clickthrough rate on our homepage hero banners. We had tried rules-based and randomly served content without very good results, so we wanted to try machine learning with Adobe Sensei. In addition, we had worked hard to get offline data flowing into Adobe Target, so we wanted to take advantage of all that wonderful data. We import many data points from our CMS system. We chose Automated Personalization because we wanted to let the machine do the work using some of the offline data and we wanted to see if we could find key segments we had not focused on previously. Additionally, we had over 15 variations of banners, and really, we were throwing up banners to see what stuck.
How easy was it to set up and maintain the activity?
Setup was easy. Overall, it is easy to maintain, include new copy, and pull poorly performing banners.
Are you seeing a significant impact from the Automated Personalization activity on the hero banner?
The best way to put it is, we are looking to use Auto Personalization on other pages and sections to see if we can duplicate the success.
What’s different about financial services compared to other industries?
What’s good is that we know a lot about our customers. What’s bad is that we have so many regulations that it takes a long time to get activities through Legal. Retail has a much cleaner result—the customer purchases or adds an item to a cart. In Financial Services, we have so many end results, so it’s harder to create a test that has a deeper funnel impact.
What advice do you have for others considering moving into using automation and AI-driven personalization capabilities in Adobe Target?
Got for it! What do you have to lose? The benefits of machine-learning are evident, and we should think of more ways to utilize the functionality. Just like any other testing that you do, determine the goal ahead of time, and make sure you can measure it with your Automated Personalization activity. Check your traffic and conversion rates to make sure you have enough traffic and conversions so that the machine can learn and give you the results.
Another way to influence things using AI that I want to try next is something that will help if you are hesitating to use it because you have already identified your important personas or segments. Build those segments in Adobe Audience Manager or some other solution and run the AI activity. Then you can see if those segments are actually the biggest influencers. You get the benefit of AI, and you can confirm if the key segments truly are the key segments.
It’s impressive how much Mike’s lean optimization team has accomplished in just three years. We’re excited to see the successes as that team expands and begins taking on personalization for the organization. We also hope he shares his results from the tactic he suggested of using AI to verify that important segments are indeed important. Many thanks to Mike for sharing his experiences, insights, and advice with us.
Looking to become a personalization pro like Mike? Download our personalization playbook today and get the play-by-play for moving to deeper personalization.