Optimizing Millions of Ads

Project Management, UI/UX Design, Wireframing, Lo-fi & High-fi Prototyping

Yieldmo

UI/UX Designer, SW Engineer

Jun 2016 - May 2017

How do you create high-converting mobile ads, without creating click bait?

No one likes ads. But why do ads have such a strong stigma when it comes to the digital world? It’s because most of them are poorly designed and serve to be clickbait that just drive more “clicks”.

But imagine a world where you’re reading SLAM Magazine, the basketball magazine, and all of sudden you see a coupon or deal for the Nike sneakers that you were planning on buying anyway? The ad ended up becoming a complement to your reading experience, without being another piece of clickbait. Plus, it saved you some money.

My task was to create, test, and deploy fully functional prototypes that drive this kind of positive advertising that would eventually exist in the form of a product.

In the World...

Here’s how Yieldmo’s Ad Format Lab is trying to reinvent the mobile ad via Venture Beat
‍A/B Testing At Yieldmo via The Yieldmo Blog
‍Pull up a Side of Meatballs via The Yieldmo Blog
Yieldmo Ad Formats via Yieldmo
Yieldmo Company Profile via Crunchbase

Identifying the relationship between performance, bad user expereince, and clickbait

Balancing Data Driven vs. User Driven Design

Data, data, data. Everyone’s all about the data. But is that a good thing? My team and I found out that we needed to be data-driven but had to balance that out strongly with the user driven outcomes we wanted.
‍
Sometimes we might see good, positive data points, but that doesn’t necessarily mean that we’re creating the expected user experience outcomes.
‍
For example, if the click thru rate goes through the roof, you might think that your ad is performing really well. But in reality, you have a bug that is forcing people to click on your ad regardless of what they do - even if they try to swipe away.
‍
You have to find that perfect balance between data and user experience.

These ads intrude the experience, and force you to click. It actually creates a negative user association with the brand.

This is a beautiful photo. But it does not drive performance at all. Do you know it's even an ad?

Design decisions must enhance the experience.

Ads should pique your interest, but never be intrusive.

As a member of the design team, we were very intentional about our design principles. I worked with senior interaction designers to create prototypes that were non-intrusive.

I created hand-drawn prototypes, user experience flows, wireframes, crudely coded prototypes, and finalized products that aligned with our design principles to develop great user-experiences.

Sample of the Windowplay Ad Format

Iterate. iterate. iTeRaTe. iTERATE.

In order to understand what type of gestures and interactions really work, you have to keep trying and trying new things. The benefit of being at scale meant that we could get statistical significance extremely quickly. We used data-driven ways to code and measure how users were qualitatively interacting with our ads.
‍

Read the full case study via The Yieldmo Blog

Prototype to Product

This is where things got real. A good rule of thumb that I followed was the 80/20 rule. Turning our validated prototypes into real products was the last 20% of work that took 80% of our time.

Crude Code is sometimes Good Code

The beauty of code is that you can always re-factor it. So, that’s what we did. In order to test and validate our prototypes, I coded the minimally viable prototype that we knew wouldn’t break.

It wasn’t the prettiest or most efficient code, but it test our assumptions with tangible results. From then on out, I started re-factoring the code to get it ready for production.

Getting ready for Production

Before launching the product, it was important to make sure that we had all of our flows and touchpoints set up. It wasn’t just about deploying code.

I worked with the marketing and sales team, our account managers, and product teams to make sure that I had properly communicated our rollout, the enhancement, but most importantly the benefits to every single stakeholder.

Launching a Product is a Multi-Team Effort.

Being both on the design side and engineering side meant that I got to sit in a really weird spot. I was in the middle of all the teams, and got to see how I could create the most seamless product rollout.
‍
I updated design and style guidelines, re-factored prototypes to production-ready code, and QA-ed each version of the product releases. I got an opportunity to work intimately with our engineering team and join their bi-weekly sprints and deployment cycles. I also updated sales deck and marketing materials so that our sales people could land more accounts.

I worked with designers to productize the final version. We designed extreme use cases, responsiveness, and user gestures.

I worked with engineers to build, deploy, and QA new versions of products into the real world for real clients.

I worked with sales and marketing teams to create case studies, client-specific demos, wrap decks, and more.

@iamtedko