Converged TV Optimization

Senior UX/UI Designer @ Mediaocean (Current 2021 Project)

Introduction

How I conquered a new vertical to design a product with a goal of revolutionizing the way TV planners purchase advertising.

The world of TV advertising is distinctly different from the world of Social advertising. Social is newer, so the workflows of brands and agencies are grounded in modern thinking that utilize innovative tech. In contrast, brands and agencies that plan to run TV advertising campaigns are somewhat outdated in their workflow and technology. Due to this, Mediaocean saw a major need in this lucrative market.

Thus we set forth to build the product, Converged TV Optimization. The project goal was to arm our primary persona, the TV Planner, with a tool that would use our advanced data science to identify how they should purchase their TV and Connected TV (streaming services such as Hulu) ads. If successful, this product would revolutionize the way TV Planners purchase advertising. It would infuse their old-school workflows with the innovative technology and data science of Mediaocean.

**A Section of the TV Planner workflow**

Research & Knowledge Gathering

How I immersed myself in the world of a TV planner to better understand their needs.

As the sole UX designer on this side of the company, I had to quickly ramp myself up and understand the complex workflows of TV advertising. Previously I had only worked on our Social advertising products, and I immediately realized the depth of knowledge I would need in order to be successful on this project.

I began by setting up knowledge sharing sessions with PMs, developers, CS, and data scientists that worked exclusively on our TV products. From there, I set up a handful of interview sessions with our primary persona to dive further into the world of TV advertising.

After these sessions, the primary PM and I attempted to map out the user flow. This served as a strong starting point and gave myself a base to move into lo-fi wireframes.

Lo-Fi Wireframes

The core design philosophy was to establish trust in the inputs, algorithms, and outputs.

Through the research, we now understood that our solution needed to do the following:

Allow users to “simulate plans” that would take their desired inputs and output an optimized analysis
Allow users to visually compare analyses against one another to help decide on the best simulated plan to purchase
Allow users to adjust specific TV or Connected TV spots that the analysis returned in order to maximize control
Allow users to visually compare the plan projected data against the actual data to build trust in our algorithms and product

Starting with lo-fi wireframes allowed us to translate the user flow into tangible UI. We could focus on the overall workflow and inputs without concerning ourselves with specific UI components and style.

UI Components

How I facilitated communication between the Front-end and Converged TV development teams to ensure alignment.

Concurrent to this project, I was simultaneously designing our next generation Social Ads product, the Unified Ads Manager. The Unified Ads Manager required working with the front-end development team to design and build out an entire new design library in React.

The plan for Converged TV Optimization was to leverage the components that were already built out in the design library as much as possible. Given the different nature of these products, there were undoubtedly going to be some new components that would need to be built.

It was my responsibility to identify where a new component would be needed. Part of that exercise was to ensure that the design of the new component would not be a one-off. The new components would need to be useful in both Converged TV Optimization and the Unified Ads Manager. Successfully building out a design library that has valuable components regardless of the product would ensure maximum efficiency for both myself and the developers.

For each new component identified, I wrote specs and collaborated with the front-end developers to ensure their timely delivery so that the Converged TV developers would have them available when needed.

Mockups & Validation

How we ensured that the design was intuitive and valuable.

We validated and fine-tuned the usability of the mockups through client feedback sessions and AB testing of multiple versions.

These sessions led to improvements particularly with the overall navigation and the validation paradigm.

The reception was extremely positive, and after improving the pain points, we felt comfortable heading into development.

Create New Plan

Allow users to define their desired plan inputs and submit it for processing. The plan will be processed through our data science algorithms and returned will be an optimized plan. The user can then compare the optimized plan against other plans, make edits to specific TV or CTV spots, and eventually decide on which plan to submit. The submission of an optimized plan will begin the execution of the identified TV/CTV inventory.

View Output: Actuals vs Projected

Once the plan has actually ran the optimized TV campaign, we connect the Actuals (the measurement of real data) to the plan. The user can now view metrics and data visualizations of Actuals vs. Projected, instilling confidence in our product and directly providing evidence that the product is providing tangible increases of performance to the user.

Today

How I am ensuring a smooth development and measuring our success.

As the product is currently being built today, I am coordinating and providing specs to development to make sure implementation is running smoothly.

I am also defining which metrics we should track to determine the success or failure of the product. The current thinking is to track the following:

The ratio of plans created vs plans purchased
The number of edits a user makes to an optimized plan
Evaluate the temperament of the end-users in regards to trust of our algorithms. Did our projections closely match reality? Did the performance of the optimized plan bring better performance than a manual plan?
Did the product save the TV Planner time compared to their more manual workflows?

In addition, the finalized mockups are currently being used in our pitch to try and land a 7 figure deal with Procter & Gamble. If we land that deal, it will be a major success for Mediaocean.

Conclusion

This was the largest and most challenging project of my career thus far. I ventured into the completely foreign and complex world of TV advertising and came out the other end designing a product that the entire company is leaning on to penetrate the lucrative TV advertising market. Time will tell if we have achieved that goal.

Below is a few of my most important learnings:

When designing for a new vertical, it is important to communicate timeline expectations. If I had not fought for additional time in the knowledge gathering phase, I may have been rushed to designing for a persona I did not yet fully understand.
Juggling a product of this scale while concurrently designing another product of similar scale (Unified Ads Manager), is not an ideal situation for a designer. It was challenging to retain the breadth of knowledge necessary for both products and pivot from one to another on a daily basis. Although we are currently hiring to help relieve some of the bandwidth concerns, I should have raised this issue earlier and with more urgency.
Being the liaison for the front-end development team and the Converged TV development team with regards to our design library requires constant communication. Look to PMs to assist with this type of management.
As a designer, have fun! Experiment with crazy concepts. Some of the most outlandish concepts I came up with had ideas within that had legs. Those ideas were refined and iterated on, and eventually made their way into the final mockups.