Redesigning Podcast Ad Inventory Search
For Flightpath, B2B SaaS Predictive Analytics Web App
Responsibilities: Product and Interface Redesign, Dev Support
Worked With: Chief Data Officer, CTO, Customer Experience Rep, Contract Developer
Primary Goal: Create a scalable sales inventory search experience that eliminates unnecessary complexity and reduces user error.
Secondary Goal: Update early days feature to match the new design system and experience principles.
Background
The Sales Availability Search (SAS) was Flightpath’s first major interface feature— the first time our AI-driven predictive technology was embedded directly into the product. At its core, SAS provides podcast networks with real-time download forecasts, using parameters like shows, ad placements, and dates to deliver industry-leading accuracy.
But as Flightpath evolved, SAS remained static. We built SAS before we had created a robust design system, and the interface had fallen out of step with the rest of the product. Unfortunately, tech debt and development constraints made updates difficult.
However, an update to the features of one of our core platform providers meant that we would have to take SAD back into development and address those constraints. Rather than just adding the new features, we took this opportunity to reimagine SAD for the next phase of Flightpath. A complete redevelopment meant we could finally integrate the improvements we’d envisioned for years.
The result: a more flexible, intuitive, and scalable tool that’s tripled user engagement, driven huge revenue gains for our clients, and maintained Flightpaths position as the industry standard in predictive analytics software.
Within two months, Report Builder was being used by 100% of our clients and became Flightpath’s second most-used feature.
I can't imagine doing my job without Flightpath reports. It doesn't just make my job easier. It lets me expand my scope and do better work than I've ever done before.
Improving Search Parameter Input with Modular Forms
Original Design
Research
- Surveys show that style and font inconsistencies make it unclear which fields must be completed before running a search
- Mixpanel data shows that most of our users only use the required parameters
- CX reports that Users adjusting optional parameters have received inaccurate data, creating an unacceptable failure point for an ad sales tool
Pain Point
The existing input form was rigid and cluttered, requiring users to navigate through unnecessary fields. Inconsistent styling and unclear helper text made it difficult to determine which fields were required, and incorrect adjustments to optional parameters sometimes led to inaccurate results, impacting ad sales. An ad sales tool can’t afford a failure point that produces bad data.
Design Solution
We introduced a modular input form that starts with only the required fields and allows users to add optional parameters as needed. This transformed the form from a rigid checklist into an adaptable tool, making searches more intuitive and reducing the risk of input errors.
On the backend, we designed the input form such that it could be customized on a client-by-client basis, allowing us to turn off specific parameters based on usage patterns.
We also aligned the form with our new design system to unify the language, font, and input parameters.
User Testing Results
✅ 100% of users successfully added the geotarget module to filter availability to the United States—demonstrating the form’s intuitive design for common use cases.
✅ 100% of testers completed a power user task by adding the position module, proving that even advanced workflows are accessible.
⚠️ 2 out of 10 testers relied on intuition rather than concrete knowledge to use the position module correctly—highlighting an opportunity for better guidance on ad placement logic.
💡 100% of users said they would feel confident using the feature in the future, confirming that the UI empowers users without creating barriers to learning more advanced ad tech concepts.
This test validated our core design philosophy: The interface doesn’t limit beginners, but it scales seamlessly as users gain expertise.
Gotta Select 'Em All
The announcement of Megaphone’s target exclusions meant we needed to take a new approach to our multi-select interface. In the original design, the user would scroll through a drop-down or type in the names of shows to add them to the list. But we would need a more robust system to handle shows, episodes, tags, networks, and exclusions.
After several sketches that pulled inspiration from different designs in other parts of Flightpath, I ended up moving forward with a design that borrowed from the Megaphone and Art19 multi-select modals, modifying it to match the design patterns of Flightpath and incorporating the capabilities unique to our platform.
One key feature of this design is the inclusion of additional data fields and context for the targets, which allows the user to find the targets they are looking for with more robust search and filter options.
For example, one user might select their video game network from the networks page to see all the shows on that network. Another might navigate to the shows page and filter down to the video game network so that they can select only the shows that they’re interested in selling. A third might filter the shows page to just gaming shows, and then order by show rank to grab their top ten shows.
Like the input form design, I don’t want users to change how they think in order to use the tool. I want the tool to work seamlessly with the thought patterns they’ve already developed. That’s the difference between immediate value to the user and promised value.
Changing Parameters from the Results Page
Original Design
Pain Point
One of the most common pieces of feedback we received on the old SAS design was that users were frequently returning to the form page just to confirm key parameters, causing unnecessary friction in their workflow. Additionally, results pages weren’t shareable, making collaboration difficult for remote teams who needed to quickly reference or adjust availability searches.
Design Solution
We redesigned the results page to display parameters clearly in individual boxes, making it easy to confirm search criteria at a glance. Users can now edit their parameters directly from the results page using the same form interface, eliminating the need to navigate back and forth between pages.
Tech Solution
We made results pages routable by encoding selected parameters directly into the URL. This allows users to copy, share, and revisit searches without needing to start from the form. What was once a static JSON file is now an interactive, persistent link—enabling seamless collaboration and asynchronous workflow continuity.
A Last Minute Opportunity
Let me tell you a story.
It was the final stretch of development for the updated SAD search, and while the developer was off toiling over the code, the team was convening at the office for our quarterly onsite.
One of the themes that kept surfacing was how we could encourage our clients to get the most out of their available inventory. The numbers showed us that revenue was being left on the table in unsold inventory that wouldn’t be sold in a traditional ad buy. If we could expose that inventory, we could work with the sellers to develop ways of tapping into that potential.
That sparked an idea!
Our chief data scientist and I had been talking for some time about ways that Flightpath could help identify shows that are similar in content and demographics. He had been working on a way of using LLMs to find these relationships, and it was getting close. We also had considered using a more advanced algorithm to find potential upsell opportunities by adjusting parameters.
I realized that in the final stretch of development, it might be possible to hand off one last element to the developer that we could turn off for now, and then turn back on once the AI was ready to be plugged in.
So in the hour before dinner, I booted Figma and whipped up these designs.
Putting together the design and mockup was easy—all I did was pull the interface elements from the table above and add a way to indicate the changes and the potential benefits.
But the exciting part is that the design works precisely because we made the pages routable. (See the dev corner above!)
I was able to change the button that reruns the search with the adjusted parameters into a button with an almost identical function—only this time, it reruns the search by appending the details of the new parameters to the link.
I walked my laptop over to the team, all standing at the door putting their coats on to go to dinner, and got immediate approval and excitement from all. I didn’t want to lose any of my thoughts, so I whipped up a few notes to prepare for more careful feedback and user testing in the coming days.
Unfortunately, all the nachos were gone when I arrived at dinner. I suppose I’ll never know what I missed out on, but I had fun pulling this together, and maybe that’s all that matters.
Results
This is the most positively received update we've put out since launch. Our clients are raving over how much they love using the new SAS search.
This redesign was a tremendous success. At the time of writing this, we’ve only had it live for a few weeks, but we’ve already seen a huge uptick in user engagement with the feature.
Some users who had completely avoided the SAS search before have started the process of migrating their workflows to be entirely in Flightpath. This has resulted in massive efficiency improvements for those teams. They’re also selling with more precision since they’re always using the most up-to-date numbers, dramatically improving their sell-through rates.
Reflection
This was probably one of the most delightful projects I’ve worked on at Flightpath. I don’t enjoy a more involved research process, but it was really cool to base the designs on two years of user feedback and usage data.
I learned so much from looking back at my work from two years ago and seeing how much my instincts, practices, and sensibilities as a designer have evolved. I now have a much better handle on design patterns and anticipating edge cases. I look forward to revisiting this page in two years and seeing how far I’ve come.