How We Fine Tuned Lyric Opera’s Marketing Efforts

January 3, 2017 Scarlett Swerdlow

Like many nonprofits in the arts, Lyric Opera of Chicago faces the familiar challenges of aging audiences, technological disruptions, and changes in cultural consumption. In response, Lyric Opera, one of the leading opera companies in the United States, has innovated in its programming and marketing. Lyric intensified their online presence, creating a new state of the art website and investing in content creation. They also took steps to build loyalty among their core audience. For their next act, they wanted to take it a step further, and expand their audience with the power and precision of data science.

Enter the Applied Data Science team at Civis Analytics. As a Chicago-based company committed to social good, we welcomed the opportunity to help Lyric Opera bring opera to more Chicagoans.

To start, we matched Lyric Opera’s current ticket-buyers to our national database, providing insights into the demographic and behavioral characteristics of their core audience. Consistent with national trends, Lyric’s typical audience member is more likely to be a woman, more likely to be 50 years or older, and more likely to be from a high-income household than the average Chicagoan who has not attended the opera.

Some of the results, though, were surprising. The single most powerful predictor of being a ticket-buyer was the likelihood of voting in an election—more predictive than age, gender, or income. Lyric Opera ticket-buyers were also much more likely to have made political and charitable contributions than their Chicagoland counterparts.

While we all love a crosstab, Lyric Opera wanted to take our analysis further and find their top marketing prospects. We used our machine learning algorithms to take into account hundreds of dimensions at once, and figured out which features were most predictive of whether a person would purchase a ticket. Using those characteristics we calculated a prospect score for every Chicagoan, and identified the set of individuals who were most likely to buy a ticket.

How Lookalike Modeling Works

Our methods take into account the important differences between arts donors and other nonprofit donors, and even the differences between opera fanatics and museum-goers. For instance, while all arts attendees are significantly more likely to have a graduate degree than the average U.S. adult, opera-goers are the most likely of all arts patrons to have completed advanced studies.

Because Lyric Opera wanted to be completely data-driven in their marketing efforts, we supplemented the targets we identified for Lyric with prospects they could have sourced through a traditional market segmentation. In this case, we supplemented our targets with households from Chicago’s wealthiest neighborhoods. Lyric Opera mailed to both lists—the Civis targets and the high-income segment.

The targets we identified converted to ticket-buyers at 3.7 times the rate of the other prospects. Through an individualized approach to marketing, Lyric Opera experienced an almost four-fold lift in ticket-buyers over more traditional tactics.

Lyric Opera Conversion Rate

Lyric Opera is renowned in Chicago and around the world for its artistic programming. Here at Civis, we are also pleased to celebrate Lyric for its commitment to being a data-driven nonprofit.

The post How We Fine Tuned Lyric Opera’s Marketing Efforts appeared first on Civis Analytics.

Previous Article
How the Robin Hood Foundation Uses Data Science to Fight Poverty
How the Robin Hood Foundation Uses Data Science to Fight Poverty

Two of the most effective federal anti-poverty programs in the United States are the Earned Income Tax Cred...

Next Article
Open Sourcing the Civis API Client for Python
Open Sourcing the Civis API Client for Python

We’ve written before about the Civis API and how we’re using it to do some serious data science. With the C...