Predictive Content Creation is where Salesforce Marketing Cloud customers can create Predictive recommendations for their email or website. A redesign was proposed to replace the unorganized flow with a redesign that supports users with varying expertise levels, upgrade to the new Salesforce Lightning Design System and improve the user experience of the product.
My role was the UX lead for this project. I designed low and high fidelity prototypes while mentoring another UX designer during this project. I worked closely with a UX researcher to design an end-to-end experience by using the information gathered from interviews, user walkthroughs, and other research methods. I maintained a strong relationship with Key Stakeholders to help define the project by understanding and communicating customer insights. I communicated often with development teams for transparency and to make sure the designs created were feasible from a technical and timeline standpoint.
The UX team took a human-centered design approach for this project.
I collaborated closely with the UX researcher throughout this whole project. We identified the personas we were designing for, their process and their pain points by conducting interview with the services team. The service team interacts with our customers on a daily basis so it was important for us to understand how our customers are using the product, the challenges they see, and what new functionality they would like to see added to the product.
Using the knowledge we gained from the services team members, we came up with a user flow. The current product lacks any kind of flow, we determined that providing a clear step by step flow would help increase the usability. To determine the user flow, we did some whiteboarding and sketching based on some heuristic evaluations. This exercise helped to identify all the steps that user must go through. I also designed these screens to be a visual process for the user rather than a technical process. We realized that the current product was very technical and the redesign had to accommodate non-technical marketers to create recommendations without the help of a developer.
Using the Salesforce Lightning Design System, I built a prototype that displayed the new flow and the visual studio. Sketch and InvisionApp were used.
We conducted multiple rounds of user walkthroughs with participants to evaluate if the new visual studio designs were effective. We wanted to determine how different the experience was for the returning user and understand how to support users with varying expertise levels. We also used the studies to gauge the use of normalized terminology and explore how we could step away from jargon. From the studies, we found that the visual studio was perceived well and we improved the ease of use ratings significantly by normalizing terminology, having a clear flow, and providing help in areas that required additional context.
We took the feedback from the walkthroughs and implemented the suggestions to our designs. We iterated, tested, and iterated for a couple rounds until we were satisfied with the ease of use ratings from our customers. Our final designs scored 6.25 out of 7.
I collaborated with the development team during this whole process. I joined scrum meetings and demo-ed the work for transparency. I wanted to make sure I was designing a product that was technically feasible and could meet the targeted release date. Once the designs were complete, I held a developer hand-off review to demo the prototype and answer any questions. Developers also held design reviews after each ticket was completed to make sure things were developed the way they were designed.
Overview - Before Predictive Email and Predictive Web were two separate products. From user feedback, we combined them into one product.
Email Image Template - "Display" was confusing to users. We created a poll to find a better alternative and "Email Image Template" was the winner. The new design is more visual compared to the old design. The old design required users to code their templates, the new design allows users to drag and drop attributes from their catalog from the left to the canvas. We still kept HTML for those who are comfortable with code. We also added additional capability, such as viewing the template with a specific item and providing a preview of the image since it may look different after image generation.
Algorithms - Before, the algorithms made no sense and users had to reference a document to understand what each algorithm was. We suggested using actual English to make it easier to understand. We also categorized algorithms so users could find the exact algorithm rather than using a dropdown like before. We also brought up some settings that were buried on the bottom and worked on help text.
Preview - Previewing recommendations is a crucial step to our users. That is why we included this in our flow as the step before deploying the code. We added new functionality such as viewing as a specific user, being able to see the specific user's top affinities as well as category affinities, and using a different reference item. Allowing our users to test with different users and products allows users to check if our system is working successfully.
From this project, I learned that...
1. Providing a clear flow is important. Without a clear flow, users feel frustration.
2. Provide content and help -- Especially when your product is very technical and jargon heavy, help and context is important, as well as providing best practices.