Adam Baumgartner | Strategy


Creating a
New Ad Unit

Two examples of a responsive ad unit in vertical orientation

The Problem

Group Nine has four ad-driven websites, and no single display ad to run across each.

The Outcome

A unit that complements multiple content types and works across all screen sizes.

The Process

Research, Prototype, Evaluate Existing Resources, Templatize, Test

Group Nine is an ad-funded company made up of four digital media brands – Thrillist, NowThis News, The Dodo and Seeker. When creating a website for NowThis, it became evident the ads in the wireframes we used for our other three sites wouldn’t be sufficient for the brand’s video-only approach.

We rightly optimized the site for video views, but in doing so made it difficult to place standard banner ads into the site.

Rather than forcing a square peg into a round hole, I saw the opportunity to create a new ad unit. I recognized running the unit on all four of our sites would be important for ensuring similar KPIs across site and promising advertisers that audiences are seeing their creative.

The Breakdown

NowThis runs vertical, horizontal, and square video, and changes the aspect ratio of the player based on device. An ad should complement these sizes.

Sites like Thrillist primarily produce written content, whereas NowThis primarily creates video. The new ad unit should live next to both formats gracefully.

Any solutions have to lean on existing tools and resources, including human power, for decision makers to consider it viable.

My first explorations focused on impact and novelty. I wanted slight motion that wasn’t disruptive to capture audience attention, like a shiny object in the corner of your eye. In this example, the creative reacts to mouse or device movement. It’s meant to be impressive, spanning the width of the page and almost the entire height.

1. Research

I began by doing competitive research, looking at solutions from Vox Media, The Outline, Quartz and other organizations with large-format ads. I looked for patterns across creative, and for levels of interactivity.

At the same time, I read about the effectiveness of such ads. Across the board, these companies reported big payoffs from there investments in large-format creative.

2. Prototype

I created a series of designs while exploring solutions, and I quickly put together a prototype as a proof of concept to demonstrate what I was thinking. I also wanted to get my product manager excited.

The image above is an example of one such prototype.

3. Evaluate Existing Resources

Now I had to figure out how to use my company’s tools to execute my plan. Developers didn’t have the bandwidth to work on this project, nor would they for future implementations, most likely.

But we did have a license for a rarely used tool that allowed us to create rich ads at preset sizes and serve them across our sites. I opted to explore this tools for a few reasons.

  • The team that would ultimately create a new unit was already familiar with the tool, so training would be minimal.
  • It was already integrated with our sites, so it would be easy to stand up a test.
  • It would require no work from developers create or run a new ad.

None of the presets had the dimensions or functionality I desired, so I read through documentation and figured out a way to gerryrig a small-screen preset to display responsive, full-screen ads.

I knew I’d need to streamline the process of creating these ads in order for decision-makers to deem this solution viable.

4. Templatize

I’d figured out how to get an ad live, but the tools I used offered a new set of constraints. Beyond that, in the back of my mind was a constant hum reminding me that if we couldn’t produce these large, bespoke ads quickly and with little manpower, these ads would never get off the ground.

So I went back to my initial prototype and figured out a way to get it to meet the constraints of our ad builder. From that new design, I created another, and another, eventually figuring out patterns and compromises enough to create a template. And from that template, I designed and built a few more ads.

4. Test

By this point, I had a new unit, I had a way to build it, and I had a template to empower us to build quickly. I also had research to demonstrate these formats perform well. What I didn’t have was data to back up my hypothesis that such an ad would perform well for us.

So I crafted an experiment.

I designed a house ad with the large-format template, and my peer designed ads with the same creative in the dimensions of a leaderboard – for desktop – and a big box – for mobile.

We picked a single position on a page, and we performed an A/B test.

The Results

The ads ran until each received 72,270 views. The large-format ad had a CTR of 0.50%, whereas the standard banners combined had a CTR of 0.09%.
Standard Banners
The ad ran between two articles in an infinite scroll format. It appears to remain still as the user scrolls between two stories. It spans the width and height of the screen.


All in all, what I created isn't a new concept, neither in interaction nor size. The value of my research, explorations and tests was primarily in seeing how our user base would engage with such ads, and determining whether we could produce them at scale.

Given the opportunity, I would continue the experiment by testing other ad positions. I'd also have our creative team put together an ad for an actual client, and track the amount of time the process took in an effort to understand cost to produce over money earned.

All of this is part of an ongoing conversation at Group Nine about ad offerings in a constantly changing industry. Though no such units are live now, the work is a resource empowering people to make smarter decisions about what Group Nine presents to advertisers.


Gaining insights with little time

How I used feedback from our audience to identify painpoints and improve our design templates on a team with tight deadlines.

Learn More


Building a new tool from scratch

When my company needed a new media management tool, I helped create something valuable for 6 teams.

Learn More