MARKETING ON PURPOSE CASE STUDY DATA-DRIVEN MARKETING Reduced direct mail and overall member acquisition cost by 20% in two years, all while increasing overall membership by 16%. CLIENT Cincinnati Sports Club THE GOAL Cincinnati Sports Club approached in 2012, with a goal of reducing advertising cost in direct mail. We needed to be smarter and able to track what was working and worth repeating, reports Mary Frank, Sales and Marketing Manager at Cincinnati Sports Club. had already done a market-penetration analysis, and provided guidance on possible site-expansion opportunities using membership trends and robust demographic data. Their system is by far the most comprehensive we have seen. It gives us the ability to accurately track prospective client activity and adjust our marketing efforts, thus converting them into sales, says Frank. instinctiveinsights.com Page 1 of 6
HOW WE DID IT Establishing a Marketing Database The first step to achieving sound data-driven marketing is establishing a marketing database a repository that merges club-provided (organic) data with rich collection of demographic and psychographic elements. We compiled a comprehensive list of all consumer households within 25 miles of Cincinnati Sports Club, complete with more than 600 demographic variables of possible interest. We typically refer to this as a Sandbox, or Universe, as it represents the population at large, even though only a subset will likely be of interest and engaged. The Universe generated for Cincinnati Sorts Club s needs was matched to a standardized, deduped and rolled-up list of current and past members as well as prospective future members of the club. Each household was verified through NCOA (National Change of Address) updated where necessary and geocoded (roof-top latitude and longitude) to provide guidance on geographical-proximity. Data hygiene is always a big challenge, explains Jeramy Fishel, Vice President of Database Marketing at. If you don t start with something stable and sound, the conclusions you draw and the decisions you make from those conclusions won t be rooted in truth. developed a host of custom-processing routines for Cincinnati Sports Club to handle nuances specific to their CRM and data-collection methods. Members Referrals Demographics Psychographics Prospects Historical Efforts Marketing Database New Mover Birthday Transactions Member Segments CLUB CONTROLLED DATA Standardization Matching & Deduping Geocoding & NCOA Persona Segmentation Predictive Modeling Channel Attribution Marketing History Model Scores SOURCED DATA instinctiveinsights.com Page 2 of 6
Identifying Actionable Demographics While having access to over 600 demographic variables can be exciting, an immediate goal of datadriven marketing is to boil that rich ocean down to something manageable, explains Fishel. Looking at each variable, we identified 27 that accurately predicted the probability that a prospect would become a member. OCEAN OF VARIABLES PREDICTIVE VARIABLES The more commonly expected variables of importance, included: dwelling type, age, income, marital status, net wealth, presence and age of children, home value, ethnicity, and proximity to the club location. We found stronger correlations when those more common demographics aligned with indicators that suggested a prospect lived an active, outdoor, healthy lifestyle, explains Fishel. For example, prospects that matched the common profile and that purchased running, tennis, biking and/or swimming items were much more likely to become members. Our club has a strong community feeling to it, Frank reports. It was interesting to learn that our members were much more likely to volunteer and donate money to local, community-based groups, both political and apolitical. Even such seemingly disconnected qualities as these in our members give us insight into who we serve, and how to serve them better. Building a Predictive Model A predictive model was built to express these variables, and their relationships, by way of a simple score that guides Cincinnati Sports Club toward engaging the households most likely to generate new members and thus be more appropriate for direct mail spending. It s all very mathematical and scientific, explains Fishel, but ultimately these algorithms are translated into a comparative number that is easy to manage. If a prospect has a score of 55, they are more likely to convert than a prospect with a lower score of 40. These scores are grouped and clustered so we can measure the percentage of households converting within a score range, thereby helping Cincinnati Sports Club make sound investments with a predictable and positive ROI. MEMBERSHIP CONVERSION RATE BY MODEL SCORE Conversion of Prospects to Members 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% Predictive Model Score (highest to lowest) Actual model score and direct-mail conversion rate for one of Cincinnati Sports Club s Target Personas instinctiveinsights.com Page 3 of 6
Tracking and Attribution When designing a marketing database, it is imperative to be forward-thinking about what needs to be tracked and how, says Fishel. Cincinnati Sports Club s marketing database tracks every instance a household receives direct mail: when it was sent, what was sent, why a prospect was targeted, and their model score at time of deployment. Prospect households are monitored and measured based on the number of times direct mailings have been sent to them, what those mailings consisted of, and how long it has been since the last communication. Each month Cincinnati Sports Club provides with a current roster of active and lapsed members. then cleans, standardizes, and matches membership information with historic prospecting activity to determine conversion and return on investment (ROI). While of vital importance from a pure ROI and how are we doing perspective, this tracking actually becomes a variable within the predictive model, explains Fishel. It s just as important to track lack of response, since a prospect with a high initial model score may degrade in score after subsequent touches. Based on past direct mail experience, our model displays which prospects are significantly less likely to become members after x initial touches. Conversely the model also tells us when another touch is a smart investment. Developing Personas We already had a pretty good idea of how to segment and speak to prospects in thoughtful, personalized ways, says Frank. Young Professionals have a very different take on the club experience compared to a middle-aged family with young children or empty-nesters. Our creative messaging now speaks to the likely wants and needs of each persona segment. Cincinnati Sports Club expanded the personalization of creative messaging and offers by utilizing the demographic variables compiled by. It didn t take long before we broadened our offerings to variant household compositions, says Frank. A family with children aged 0 6 receives different imagery and messaging, from that which a family with children of ages 7 13 or 14 17 receives. and Cincinnati Sports Club have built a robust offer toolkit for an expanded variety of household situations. The predictive models have evolved to contain persona-specific algorithms that predict the likelihood a prospect household will convert based on the uniqueness of their family makeup. For example, we have a situation where Income-Producing Assets (IPA) are more highly correlated to the conversion of a middle-aged or older-family, whereas for a younger couple it really has little or no bearing, explains Fishel. As an offer toolkit is developed, we evolve the predictive models to incorporate our client s creative as well. instinctiveinsights.com Page 4 of 6
Demographic Refreshes Twice a year, we identify new prospects not yet part of the marketing database and refresh demographics for households. Refreshing demographic data isn t just about having current information, says Fishel. When we see major life changes, such as a new child in the home or a significant increase or decrease in household income that development is integral to the model as a changing value. Let s say it is important to target households with income of $75K or more. How does the importance of that income range change when a household has moved from $50K to $75K in a year s time? Is the change in income a stronger predictor than the income level itself? Our models monitor those changes and make scoring adjustments based on actual direct mail performance of the prospects we are targeting. instinctiveinsights.com Page 5 of 6
FINAL THOUGHTS Every engagement, whether through direct mail or via another tactic, should be viewed as an opportunity to test and learn something, says Fishel. Cincinnati Sports Club is a perfect partner for us as they always want to challenge the status quo and squeeze a little more value out of our efforts. We have launched several new investments because of our success reorganizing our direct mail spending, explains Frank. Partnering with, Cincinnati Sports Club launched both a B2C and B2B prospect-email marketing program, a sweepstakes contest on Facebook, and social media re-marketing. Cincinnati Sports Club has plans for website expansion, on-premise GPS and beacon technology, a membership app, and a Re-Marketing Program on Google, Bing, and Facebook. To learn more about data-driven marketing for fitness clubs, visit us online at instinctiveinsights.com. ABOUT INSTINCTIVE INSIGHTS is a full-service, data-driven marketing company headquartered in Columbus, Ohio, with employees in Cincinnati and Atlanta. Our purpose is to truly help Fitness Club Clients both big and small be more successful in acquiring and retaining members improving acquisition rates and lowering acquisition costs. In doing so, we empower our clients to allocate more time and funds toward providing an exceptional club experience. Great marketing is about truly making clients better off by understanding and satisfying their most important business needs. It is about providing real, measurable value. Visit instinctiveinsights.com to learn more. ABOUT THE AUTHOR Jeramy Fishel was born, raised, and educated in Seattle, WA. He was involved in number of tech startups before moving to Cincinnati to join a young, pioneering, and data-driven marketing company. After a decade of year-to-year growth, he went on to manage global market research studies for dozens of Fortune 500 companies and led an international technical sales team for one of the world s largest software companies. In 2013 he joined as Vice President of Database Marketing. His technical skills, passion, diverse marketing research, sales experience, and ability to think as both data scientist and marketer, produce unique perspectives and measurable results. instinctiveinsights.com Page 6 of 6