Zipcar traditionally messaged its members through email and social media. But those efforts didn’t reach all members.
“Not everyone follows us on social channels or through email, but everyone is on the web,” said Millie Park, Zipcar’s senior director of member marketing. “We wanted to find ways to communicate with them through appropriate targeted messaging.”
In March, Zipcar delved into display and began working with demand-side platform (DSP) ChoiceStream because it offered in-ad polling and creative services.
The managed-service DSP combined polling data with Zipcar’s first-party assets to build personal messaging via its creative services division. This support meant Zipcar’s in-house creative staff wasn’t overburdened because the company doesn’t work with any advertising agencies.
Information from ChoiceStream’s in-ad polls informed how Zipcar interacted with members, based on information like whether they knew of Zipcar’s one-way option in Boston or its airport services.
The initial test campaign used hashed member email addresses matched via a third party to see how members would react to the advertisements and polls it produced.
The poll answers sometimes differed from assumptions Zipcar had about members. “We learned a lot about members we didn’t know, and we can use those audiences to hypertarget messaging for next campaigns,” Park said.
But it still wasn’t reaching enough members. Email hashing only produced about a 20% match. Zipcar tagged its desktop and mobile reservation pages to expand the members it could reach and used cross-device mapping to target across devices.
Zipcar A/B tests everything and used a control group (which saw an ad for a nonprofit) to measure the incremental revenue lift from the campaign. Even if a user clicks on an ad and books a reservation, Zipcar only measures the extra reservations that came from the campaign – not an easy task.
Ultimately, Zipcar saw a 12% incremental brand lift. It also posted a 6X return on ad spend, using a “tight and defensible” three-hour attribution window, much smaller than typical seven- or 30-day windows, Park said.
While ChoiceStream’s polls help Zipcar understand its members better, it also wants to target messages to members based on its CRM database.
Park eventually wants to eventually connect different data points like ad engagement and acquisition to driving behaviors such as what, when and where a member drives.
Zipcar also wants to pull data from its reservation system, so it can ramp up display ads to drive bookings when a region or time looks sluggish.
“If we have a market that’s looking soft for Fourth of July, we can shift budgets and dollars to book ahead of the long weekend,” Park said, adding that she wants to automate that process. “We’re starting to look at refocusing budgets regionally, or toward different parts of week or holidays based on how we foresee utilization in certain markets.”
Her next goal is figuring out how much she can grow her campaign budget and still capture incremental revenue from members. The problem is that they’re a limited pool, Park said.
“There are still further edges to push,” she said, “and members we haven’t been able to find.”
