DR Smart Pixel: Identity Match Meets Addressable TV
with Cross Platform Contacts at a Fraction of the Cost of Spot TV
IDENTITY MATCH in the form of our DR Smart Pixel connects one customer data set with another via a persistent identifier, such as an email address. (Usually, the identity is masked thru a secret 32-character code called an encrypted hash)
For instance, a Direct Response Advertiser might supply its customer list, and DR Smart Pixel will connect it through the email or another solid identifier to other data about that customer, such as their mobile device ID or an identifying cookie.
Now, the Direct Response Advertiser can directly and individually target its customers online. Hello addressable, 1:1 TV!
DR Smart Pixel uses the term “people-based marketing” to describe this identity matching. The idea is that it joins customer profiles that don’t just target the few behaviors of a cookie, but the actual type of person who may also have a high Intent to Purchase.
Today, the Las Vegas-based DR Smart Pixel is expanding its universe to include television and 19 media platforms such as HULU, Spotify, LinkedIn, and YouTube.
For a Case Study, consider a national DRTV campaign targeted at hybrid car buyers in the state of California. (Targeting people like my brother who drives in LA traffic 145 miles a day and gets real stabby not having a plug-in hybrid that qualifies for the car pool lane)
HOW TO TARGET JUST 10,000 VIEWERS
DR Smart Pixel selects 3 million users in its database with online browsing of websites about pricing of hybrid electric vehicles. Heck, they don't even have to live in California.
We then match these data sets to the hashed email of recent hybrid electric vehicle buyers of the client products.
Together we narrow the target list down to a list of 10,000 prime targets in-geo and in-market. We build a custom audience toward these users and hit only those targets--get this-- on the devices they prefer including email, Facebook, Hulu, Spotify and now Comcast.
Using a DSP or demand side platform, Comcast provides its subscriber data matches to our team, which matches the Comcast subscribers to the DR Smart Pixel targets through an email address or another persistent identifier. This yields, the 10,000 Comcast subscribers who are on our list.
The client can then direct TV ads ONLY to those subscribers with addressable set-top boxes. Or some of the matched Comcast subscribers might have our Custom Audience data indicating that they follow shows like “Modern Family” and they live in San Diego, so the ads can be directed in the old-fashioned way of buying a show and geographical market, but now propelled by data and individually addressed.
In this and similar use cases, DR Smart Pixel serves as a kind of match-making Switzerland. Comcast would likely prefer to give its golden subscriber list to a neutral party like DR Smart Pixel.
Our match-making is useful in determining cross-device identity, where an advertiser wants to direct ads to the same user across that user’s phone, laptop and tablet.
As television is currently such a fragmented platform, running TV ads across all its variations with our custom audience also solves the cross-device identity problem.
At DR Smart Pixel, we remain “very focused” on addressable TV and specifically look to boost the effectiveness and ROI of direct response advertising.
Winner! Winner! Chicken Dinner!
Making Machine Learning a Reality for Your Business
Home Price Predictions
Our team can help predict house prices based on several parameters including least squares linear regression and Bayesian linear regression.
Human Activity Predictions
Our machine learning can predict the human activity (walking, sitting, standing) that corresponds to the accelerometer and gyroscope measurements by applying the nearest neighbours technique.
Credit Card Fraud Prediction
We can create a fraud detection alert that predicts potential frauds using credit card transaction data. Machine learning can utilize the random forest method to identify fraudulent transactions.
Our team can create market segments using US Census datasets (or even hashed anonymous sales data) by applying the k-means clustering method.