This Problem Cost Our Client Millions In Lost Revenue, We Solved It In 4 Weeks.

Theres a reason why technology and politics are not for the weak. The endless problems one encounters working in these fields can drive a person to madness. Good design however can solve any problem, no matter how big or messy. I joined the product team at the political technology startup Circa Victor in May 2016.We had 4 weeks to concept, design and launch Redline, an enterprise solution for multi-media corporations that are tired of losing money.

The Problem

The client (A well known Multi-Media Corporation) approached us with a problem.

  1. Political committees spend money faster than they can raise it, making it difficult to pay their bills.

2. Broadcasting companies lose millions each year when political committees cant pay for booked airtime.

3. Selling media services without knowing which committees can afford is risky.

Product strategy

Identify The Problem: A Multi media corporation hates loosing money when a political committee defaults and cant pay, this makes targeting new customers increasingly risky.

Service To Solve It: Circa Victor is here to design a product that prevents revenue loss through awareness. The product will inform the media corporation on which committee’s are safe to target based on the likelihood of default.

How to Achieve It: The product will determine the chance of default by tracking the committee’s expenditures, donations and cash on hand data. We will assign each committee with a healthy, risky or safe label based on defined data parameters.

The Process

According to the data, most of the defaults occurred when there was a spike in spending. We invited the clients analytics team to meet with our data scientists. The data team worked to refine the burn rate model by identifying the rate of money coming in to the rate of money going out.

Default Event- Any occurrence of a campaign canceling a buy after the cancel-by date, defaulting or announcing intention to default.

Burn rate — the ratio of a campaign’s spending to its assets.

Our data scientist worked closely with the client and created a predictive algorithm of when a campaign will miss its payments. With expenditure and donation data from early as January 1, 2010, we created simulated default-events around expenditure surging periods.

The results were 87% accurate, this confirmed our assumptions about the default events. The next step was to figure out the design for the data.

Designing The Interface

How does a human easily digest this information without feeling overwhelmed or out of the loop? We played around with geocoding and choropleth maps; using location to refine results and heat maps to relay risk. Unfortunately, the maps required more work for the engineers and we only had 4 weeks so we kept going.

The maps were labor intensive and not ideal for such a short deadline. Names and figures blurred per confidentiality agreement

We continued our design research: evaluating heuristics, exploring data visualization and graphing solutions towards a clean deliberate interface with interactive graphs and tables.

Design, Test, Revise

Designing with data allowed us to pinpoint problems faster and explore potential edge cases and their exceptions. Instead of coming up with simulated scenarios or false names in a prototype, the interface was designed with live data.

We designed the original risk concept with progressive and regressive bars but the flow was not as seamless as we hoped. We faced an inevitable user quandary with the risk visual: the bars say bad but the numbers look good.

Risk chart Before

We reworked the visual concept and went with a static front and center label. I imagine all elements of UX design as a sentence. Can the user look at your design and form a complete sentence based on the order of visual elements?

This sentence says: The first committee is risky with a 97% chance of defaulting.

Names and figures were changed per confidentiality agreement.

Redline rates the health of a committee by analyzing the current activity and recent trends. We designed a visual story about the committee’s annual spending activity and divided it based on the span of a campaign cycle.

4W = A month — a minimum span of time to view a trend
3M = A single fiscal quarter 
6M = Two fiscal quarters
1Y = Almost the entire campaign cycle

Names and figures were changed per confidentiality agreement.

The user can hover over any day within a 90 day date range and discover the committees exact amount of cash on hand.

Spending data was categorized for broken down into percentages. Clicking on a committee triggers a detailed spending report in a pop up modal.

Names and figures were changed due to confidentiality agreement. A detailed modal contains breakdown of committee spending.

The Final Product

Broadcasters and Media Corporations use Redline to target new customers, assess financial risk, and ultimately save hundreds of millions of dollars in lost revenue.

Future Applications

Thanks to the success of the product and the hard work from the engineers, Circa Victor developed an API with Redline data.

Names and figures were changed to per confidentiality agreement.

My Experience Working For A Startup.

Quote

Everything you do in a startup makes a difference. No longer are you surrounded by a safety blanket world where you’re a small cog in a large machine. In a startup, everything you do will contribute to the ultimate success or failure of the business. -Adam Arbolino, Tech Crunch

I was really nervous when I got this job because I didn't know what to expect. I don't consider myself an expert in politics and big data was intimidating to me. I felt comfortable doing user experience so I pushed my UX agenda from day one but I didn’t get far. The chief product officer made it clear that him and I would be involved in every aspect, not just user experience. We worked in an agile scrum environment with a weekly rundown on Tuesday and product recaps on Fridays. I was tasked with designing sales deck presentations for the growth team and made revisions based on feedback after sales meetings. I learned to love the capabilities of data and embraced the limitations and turning points in the design process. I led bug review meetings with engineer and used Github to track progress.

Most importantly, I worked in the same room with the three founders of the company. Not only did I have a front row seat when things went wrong but I was an audience participant that helped get the show back on track. I can't thank Circa Victor enough for giving me this opportunity and I'm so excited to see what they do next!