In 2007, we adapted our bank reporting platform to support Hacker Group, an ad agency that manages direct mail for a large telco. In 2009, we were selected to build a list segmentation tool for the Direct Marketing Group. We've been helping send a lot of mail ever since.
These two engagements demonstrated that we can handle marketing campaigns of any sale and complexity. In fact, we thrive in a complex, ever-changing, high-touch advertising eco-system.
At present, we are retooling to use Python/Spark as a platform.
Enterprise Data is messy. One of the first things we do with a new client is use our custom ETL (extract-transform-load) tools to organize multiple data sources into a usable form.
This typically consists of isolating bad data, transforming cryptic codes into a human-friendly format, and applying meta-data to regularize data. We usually wind up integrating 5-10 different data feeds from various silos.
Because we have specialized tools, we can do this in an efficient, repeatable fashion.
Reporting covers a lot of ground. We have found that different users need different tools so we provide a variety of options.
We have a light-weight web-based exploration tool for getting answers quickly with minimal learning curve. We also have a standardized reporting capability for complicated reports that need to generated month-in and month-out.
For great visualization and interactive dashboards, we support Tableau. We also provide strong integration with Excel.
We enable advertisers and their agencies to handle complex marketing campaigns -- at scale with minimal labor.
We support multiple outcomes, multi-stage operations, suppressions, dynamic distance-to-store calculations, record limits, integrated historical response, two-variable selections, audit trails, and automated output. We have all of the features that you would expect in an enterprise-class list segmentation tool.
Using our tools, we were able to take campaign excecution from days to hours. Our next release takes them from hours to minutes. However, the real advantages is that they can tune their thresholds based on data rather than randomly picking records to meet a quantity target.
We are building a robust collaborative optimization solution.
We are not ready to talk about the details but we can talk about our result targets.
Our preliminary analysis indicates that we can get a minimum 10% lift over segmentation alone. It uses a white-box approach that doesn't require the advertisers to place blind faith in a magical algorithm. We think we can accomplish this without requiring an army of PhDs to maintain.