Insight/Action/Outcome: The IBM Cloud product design team pulled data from Amplitude and saw that their What’s Next notification received little to no clicks—despite also seeing users search for documentation on appropriate next steps. The team redesigned the notification to make the next steps more visible. This notification now has 8X more unique users.
In 2022, I joined the product design team of IBM Cloud. I am a passionate individual with a keen interest in startups and a drive to move fast. I enjoy collecting insights from various sources to create an action plan for progress. Fortunately, my management team supported my eagerness to learn different areas and my dedication to data and research. The company has provided me with the freedom to learn about analytics and apply it to user experience design, which is why I see myself developing a long career here. Even though learning about data is not a traditional UX path, I believe that integrating data into design can help my team create better products.
Bringing data into design helps teams create better products.
Better autonomy for the design team
I had two goals as a designer:
- To help product designers achieve data autonomy. Being able to create charts and analyze them ourselves, instead of waiting for someone to pull the data for us, would mean we could turn those insights into action much faster. We could include more data in our daily processes, which would also serve the goal of accelerating workflows.
- To foster increased collaboration between teams. Nothing exists in a vacuum, and data is always relevant to more than one area. We are all interconnected, so we needed a way to share and communicate our analysis.
Before working at IBM, I had experience with Pendo and Looker. In both cases, I had to rely on others to pull the data. IBM used Amplitude, which gave me the opportunity to understand instrumentation and access data myself.
Compared to the analysis I had received from other tools, Amplitude is more robust, with more features to analyze and visualize data. I especially liked that I could build a funnel and return to that same funnel to gather completely different data.
I took an asynchronous, self-guided approach to training, using IBM tools like Cloud Analytics: Amplitude 101 for data. I was determined to develop data analysis as my superpower and share that superpower with everyone else.
Evangelizing for analytics in design
After learning some basics, I asked myself how I could show other designers the importance of data and Amplitude’s ease of use. I began by fulfilling requests from my team: If they needed any data to make more informed product decisions, they could make me a ticket or simply slack me, and I would pull that data for them.
From there, I held a working session to show design teams the value and power of Amplitude and how easy it was for them to explore themselves. A recent working session had seven participants who recently received an Amplitude license: some design architects, some UX designers, a manager, and a team lead. During the session, I walked the group through a high-level overview of Amplitude Analytics and how events and segment tracking work. Then, we built a funnel chart together to analyze product usage.
We have also shared our experiences using data to increase awareness of the platform and its capabilities in our Public Cloud Design All Hands and IBM’s Spark Design Festival — where we presented an Amplitude Ramp Up topic to a community of roughly 3,000 professional designers.
Two surprising insights turned into action
Not long ago, we had an aha moment around our recently redesigned Classic Infrastructure Overview page. The team asked me to check the page’s performance to see how much time users spent on it and which tiles they clicked. When I pulled the data and saw a 1,000% monthly growth in unique visitors, I didn’t believe it. I thought something must be wrong with the data—historically, this overview page wasn’t as popular as our provisioning page, for example.
But living IBM’s value of “stay curious” and believing in the data, I decided to investigate the reason behind this figure. I pulled the referring locations for all Classic Overview unique visitors and noticed the top reference was IBMCloud.com. Collaborating with the marketing team, we discovered the cause was the Walk Me Smart guided tour, which directed users to the Classic Overview page.
This exercise demonstrated the importance of following the data because the answer may be an opportunity for cross-functional sharing and collaboration. In this case, an insight told us not what to change but what to retain. The Walk Me Smart tour was under review since we were unsure whether it brought enough value. But Amplitude showed us how much users value the feature, so much that IBM decided to keep the guided tour and develop it further.
It’s important to follow the data, because the answer may be an opportunity for cross-functional sharing and collaboration.
Another example of how designers can use data effectively is our What’s Next notification. This notification displayed as a dropdown on some provisioning pages and informed users about the next steps post-provisioning. The designer from my team was considering redesigning this notification, so I pulled data to determine whether a redesign would be worth the resources: we figured that in some product areas, no one clicked on this notification for as long as a month. This insight was especially surprising because I could also see users searching for documentation on appropriate next steps.
From this analysis, our designer decided to redesign the notification, eliminating the dropdown and putting the next steps in front of the user immediately so they don’t have to look for it. This notification now has 8X more unique users.
You don’t always get what you want (or expect) when you follow the data. But Amplitude always delivers the insights we need to formulate an action plan.
By the numbers: improving conversions with UX
In the six months after starting the initiative to bring Amplitude into our design process, I developed a plan tracking 11 areas within Virtual Private Cloud (VPC) with more than 70 charts. Generally, we are interested in unique visitors, conversion rate, button-click events, and referring pages, but that varies based on the area and the page type. Our ultimate goal is to increase provisioning, which increases revenue. We do this by improving the customer experience and helping users through the steps toward deciding to provision a new server.
We always ask what interests our users most and what info they still need to make an informed choice. Do we need to restructure a page or add more documentation and notifications? Are there any hiccups that we can address to help the user move through the provisioning pages faster and hit the create button? Gathering data in Amplitude, we can answer these questions.
As a result of making more data-informed decisions over these six months, our SAP Wayfinder unique views grew by 418%. Meanwhile, our Virtual Server Instance Creation time on task and median time to convert dropped by 20% in 2023. In other words, users took 20% less time creating the instance and deciding to purchase. For Bare Metal for Virtual Private Cloud Creation, time on task and median time to convert dropped by 81.5% in 2023.
Small expenses can quickly add up at a company of IBM’s scale. For my team alone, eliminating unnecessary instrumentation reduced our data costs only at one page of the flow by up to $1,000 annually.
In addition to increasing conversion and revenue, we also aim to reduce costs. One way we can do this with Amplitude is by reducing unnecessary instrumentation. I noticed in Amplitude’s event tracker that some events triggered multiple actions mostly because of event duplication in the code.
Small expenses can quickly add up at a company of IBM’s scale. For my team alone, eliminating unnecessary instrumentation reduced our data costs only at one page of the flow by up to $1,000 annually. We identified that opportunity through Amplitude, and we've since taken a more holistic, cross-platform approach to event tracking.
Better product knowledge makes us a better team
IBM is a big enterprise, and our processes can be lengthy. But now, instead of waiting on the product and analytics team to pull data for us, designers have gained data autonomy and can self-serve the information we need—within an hour.
Amplitude usage has increased by 980%, and we're a better team for it. Designers use Amplitude daily and have made Analytics part of every process, every redesign, and every new feature development. With charts giving us every angle on the product, we understand the product better than ever, and data gives us a stronger voice to influence the product roadmap.
As UX evolves, Amplitude is becoming an important tool for product designers, making us a better team.
Amplitude may have been originally intended for product teams, but as UX evolves, Amplitude is becoming an important tool for product designers, too.