In August 2013 we announced the ability to quantify the impact of Promoted and organic Tweets on offline sales for consumer packaged goods brands in the U.S. Today, through our continued partnership with Datalogix, we’re pleased to introduce offline sales impact for U.S. auto advertisers to measure actual vehicle purchases from their Twitter audience.
New research: Households with Twitter users were 2x as likely to purchase a new car as the average U.S. household
By matching unreadable, scrambled (hashed) email addresses from Datalogix, Twitter provides information about activity on our platform that Datalogix can then pair to vehicle purchases captured by Polk (which was recently acquired by IHS). Polk’s comprehensive automotive data captures new and used vehicle purchases by make and model through DMV reporting. Datalogix’s reporting helps auto advertisers identify the Twitter audience segments that demonstrate the highest conversion, providing insight into campaign effectiveness and helping future media optimization.
Auto brands have previously been able to measure the engagement and brand lift achieved by Promoted Tweets, as well as how Twitter-specific engagements and impressions can lead to online conversions. Now they can measure actual vehicle purchases through offline sales impact. This helps Twitter campaign measurement from the top of the funnel to the very bottom.
Initial results on vehicle purchases
As part of the development of this new offering, we asked Datalogix to measure the car purchases of the total U.S. Twitter audience, those who follow auto accounts on Twitter, and those who were reached with Promoted Tweets from 10 different auto brands. The results were compelling:
1. Households with Twitter users were 2x as likely to purchase a new car as the average U.S. household
We learned that U.S. households with Twitter users were twice as likely to purchase a new vehicle as the benchmark, which included a group of randomly selected U.S. households for the geographical location of the measured Twitter audience. Not only did they exhibit a higher propensity to purchase; households with Twitter users also demonstrated vast reach, accounting for more than half of all U.S. vehicle purchases during the six-month measurement period.
2. Households with Twitter users who engaged with Promoted Tweets from auto advertisers were 32% more likely to purchase a new vehicle than the average Twitter user
We’ve already shown that Promoted Tweet engagement drives greater in-store sales for CPG brands. Now we’ve learned that Promoted Tweet engagement is also predictive of car purchases. Users exposed to an auto advertiser’s Promoted Tweets were more likely to buy a car than the average Twitter user, and users who engaged with those Tweets had an even higher buy-through rate: 32% more likely than the average Twitter user.
3. Auto brand Twitter followers were nearly 3x as likely to purchase a new car
Not only are an auto brand’s Twitter followers more likely to currently own a vehicle of that brand; auto account followers also were more likely to purchase a new vehicle, especially those followers acquired during the measurement period. Many of these auto brands were running Promoted Accounts to grow followers among users in-market for a new vehicle as well as brand loyalists post-purchase.
These results are just the beginning. This new product offering will help auto advertisers learn how to leverage the power of Twitter to enhance their media mix. If you are an auto advertiser in the U.S. and want to know more about qualifying for this measurement product, please reach out to your Twitter account representative.
How we protect user privacy
Consistent with our CPG offline sales impact capability, user privacy is our top priority with this new measurement product. The reports that Datalogix creates for U.S. auto advertisers include only aggregated anonymous information. Users can opt out of Datalogix measurement studies at any time. If a user opts out, Datalogix does not send the user’s hashed email address to Twitter.
For more information on our process and user privacy, please visit this Help Center page.
Did someone say … cookies?