By Daniel Williams Published May 26, 2016
Raleigh, NC, May 26, 2016 - A recent meeting with a client provided a great example on the immense revenue generating potential for newsmedia companies that have adopted a data-driven strategy for audience development.
Our client was seeking to understand how to leverage the LEAP audience database to call on the most sophisticated of local SMBs – banks, credit unions, retail, insurance, and entertainment venues - with the ability to take their customer files and identify not only “look-alikes” but, if possible, perform predictive behavior modeling.
Using tools like LEAP Dimensions, many of our clients are discovering that the audience database is an asset that not only supports audience engagement efforts, but can be used as a strategic asset to serve the local advertising market through a managed-services approach.
In order to provide data marketing services as a solution to local SMBs there are really 3 key elements required:
the data – the integration of 1st party behavioral (e.g. a customer file with transactions or web visit behavior) and 3rd party acquired data that is licensed or purchased (demographic / behavioral).
the technology to process this data – a core database engine (such as LEAP's audience management platform), as well as a reporting & analytical tool to summarize and display the modeled data (such as LEAP Dimensions, powered by Tableau).
the human resourcing to take the data and build the compelling story & marketing tactics to meet the organization's goal.
When LEAP deploys an audience database for its clients, we address all three of the above elements under our shared resource management approach, further described here.
Descriptive vs. predictive modeling
Through our partnership with Acxiom, LEAP appends a household-level demographics segmentation system - Personicx - to every address in a publisher's market. The segments on their own are of use, but are really most powerful when used as a component of our more comprehensive Targeted Growth Model methodology because we also incorporate additional lifestyle, psychographic, and channel purchase preference data into the scoring and segmentation process.
This can best be descriptive as a "descriptive" model. For example, “Mary Smith of 123 Main Street” may represent a high-value potential customer because she has characteristics in common with the advertiser’s typical customer based on our analysis and scoring of their actual customer file.
However, to build a truly "predictive" model, you need as an input a dataset that provides both successful & unsuccessful outcomes (for example, customers who bought as well as people that did not buy), and then apply a typical predictive modeling technique such as regression analysis which outputs a propensity and degree of confidence for a given outcome.
Because LEAP’s audience database is also a campaign automation system, executing your targeted marketing activities through it allows us to conduct not only the modeling, segmentation and target selections, but to track the campaign outcomes through the response reporting metrics. The latter item is critical to making the “LEAP” from a "good-enough" descriptive model to a "truly-great" predictive model, and puts you - the publisher - in the driver’s seat for providing an end-to-end marketing solution (data modeling, segmentation, targeting, execution and response analysis) to local SMBs. Better yet, through the continuous loop of data (input > output > input > output...so on) the predictive model is self-correcting and moves toward an optimal state for driving campaign response as you refine your target audience based on actual outcomes.
Working with your advertisers
If an advertiser can provide you with just their customer file with no transactional data (e.g. Mary Smith at 123 Main Street made a purchase in April), the best approach is to build a descriptive model (i.e. ‘lookalikes’) which can still be very powerful in predicting response.
Even better is if the advertiser can supply you with transactional data that accompanies the customer record (e.g. Mary Smith at 123 Main Street was targeted with a direct mail offer in February, and purchased a furniture suite for $2,500 in April on a 24-month finance term), in which case the more advanced “predictive model” can be performed.
In either case, your investment in an audience database, enriched with 1st and 3rd party data assets, and equipped with the ability to perform both descriptive and predictive analytics as a service for local advertisers, and most importantly identify and target the highest value potential customers, presents newsmedia companies with an exciting opportunity to serve and fulfill the needs of local SMBs.
For more information on LEAP’s end-to-end approach to audience management, please contact us at firstname.lastname@example.org.