Role - Head of Product Design

I came on at Magnetic as a Sr. Product Designer. Within a year, I was promoted to Head of Product Design. Finally! The big chair! Kind of? Not really. The hierarchy of the product team was still flat, and the design team was small. I still did the same work: discovery, validation, wireframes, UI kits, mockups, prototypes, user testing, etc. People did seem to be friendlier. I will say that.

Product - Magnetic Force DSP

For many years, Magnetic ran programatic ad campaigns for clients, using tools (commonly known as DSP’s) like Tradedesk, AppNexus, or Google’s Doubleclick. After a while, they began to notice that they spent so much time and effort trying to hack these programs in order to get better performance, they decided to build their own, which they called Magnetic Force.

Force’s main value-add to the DSP space was its ability to automate the targeting process. While historically, ad-ops managers would adjust targeting parameters manually, with Force hundreds of attributes could be considered for every bid request.


Design Sprint - Ad Setup

After we launched, we did a usability survey, and we discovered that the one feature everyone was having trouble with was the ad setup, which had a number of drawbacks, the most painful of which was the fact that multiple creatives couldn’t be added at once. Rather they had to be added one at a time. This combined with some of our other ad set problems resulted in the perfect opportunity to run a design sprint that focused on revamping the ad setup/management UI.


Generating Buy-in

Design sprints were relatively new to Magnetic, so getting stakeholders onboard required some compromise on time/resource constraints along with some extra incentive. It cost me $80 in Starbucks gift cards and an extra $20 for the hipster who didn’t drink Starbucks. Another $100 on pizza, soda, and supplies, but in the end I was able to assemble a team of six consisting of myself as Facilitator, one Director of Engineering, one UI Developer, 2 Ad-ops Managers, and 1 Chief Product Officer.

“How Might We…?”

The design sprint involves several exercises that help teams focus on making sure they’re solving the right problems. One such exercise we do on the first day is called “How might we?” It’s where we figure out which problems we want to focus on and then turn them into statements that begin with the phrase “How might we?” We vote on the ones that are the most important, which we would sketch out solutions for later on in the sprint.

Screen Shot 2019-09-05 at 2.28.18 PM.png

How Might We Winners

  • How might we validate that an uploaded ad is working?

  • How might we simplify + combine bulk/single creative upload flows?

  • How might we bulk edit creatives?

  • How might we know when something is wrong with a creative?

  • How might we reuse new ad creation/management design during campaign setup?

Journey Map

Once we decided on our “how might we” statements, we drew a journey map illustrating every possible step in the ad setup/management user journey. The artifact below is a recreation of that drawing.

Magnetic Force Design Sprint Journey Map@2x (1).png

Wireframe Sketches

On the second day, we all went off on our own and did sketches on our own time. We would regroup on the third day when everyone would present their sketches and we would vote on the ones we want to see make it into a prototype.

Screen Shot 2019-09-05 at 4.07.38 PM.png

Hifi Mockups & Prototype

Using a combination of on the fly visual design with whatever I could muster from the UI Kit, I took the fourth and fifth days to turn our sketches into a hifi mockups that would eventually be turned into a prototype via InVision.


New Ad Sequence

Ad Management Sequence


Usability Testing Results

Fortunately, we had nice pool of in-house campaign ops specialists to test with. Not exactly statistically significant, but perfect proxies for our primary user persona. The feedback was very positive, and across the board, we heard the new design would save a lot of time, otherwise spent on tedious manual tasks.

BJ Fogg Behavioral Model Scores

  • 8 of 10 users rate 5 of 5 for usability.

  • 9 of 10 users rate 5 of 5 for motivation.

Iteration Candidates

  • “‘New Raw Image’ button”

  • “Pasting a single ad tag”

  • “Integration with ad server”

Outcomes and Takeaways

I felt great coming out of the design sprint. We checked off everything we set out to accomplish. But our company was acquired by Deloitte not too long after the sprint. The product itself was taken apart and eventually reassembled into what is now Hux Audience, so most of the features that cam out of the design sprint never actually made it into production. And so it goes.

On the other hand, it did set a precedent for how our team could collaborate on design going forward. It got more stakeholders invested in the design process, and it got everyone thinking critically about designing the right features, which is really the best outcome any product designer could hope for.