Omnichannel
In Klaviyo, users create and send emails, SMS, and push notifications to their customers, but a user can only send a message on one channel at a time. Customers needed to more easily orchestrate marketing across multiple channels to save time and optimize ROI.
Working across 3 product teams, I designed the end-to-end solution and worked with R&D leadership to break down, plan, and allocate work.
Discover
At the outset of the project, there were some backend structural questions that needed to be answered quickly in order to 2023 commits. To unblock engineering, we rapidly user tested lo-fi concepts that isolated key variables in question.
Open question: What navigation pattern should the campaign builder utilize?
Open question: When and where should multi-channel be introduced?
Open question: When and where should A/B testing be introduced?
“Alex was able to craft sets of designs using lo-fi mocks that we could quickly get in front of customers. Without Alex, I don’t believe that the project would be where it is today, and we might still be working on making user flow decisions.”
— Senior Product Manager (Campaigns) | 360 performance review
Define success
Our customer research findings enabled engineering to conduct more detailed scoping which allowed us to define an achievable goal toward our company initiative.
Deliver beta by Q4’23 where customers can create a multi-channel campaign within in a single workflow.
(Stretch) Increase the number of campaigns that contain more than one channel from 2% to 5%.
Design
With high level structural questions answered and goals defined, we moved onto the details of the campaigns experience. In approaching the design, we focused not only on facilitating a multi-channel campaign, but incentivizing one.
Test
We conducted usability testing with a high-fidelity prototype. Participants were asked to create a campaign using a single channel, then create a campaign using multiple channels.
Iterate & refine
Our researchers collected a long list of observations and compiled into a set of recommendations with prioritization. To the right is just one example of many iterations we made toward the final output.
Below are screenshots of our finalized designs
