Why Data Science Needs To Be Simpler for Marketers
Overall IT budgets might be flat, but marketing budgets are growing as companies try to maximize outreach to customers across multiple channels. For the data-savvy marketer, the world of big data and data science offers tantalizing opportunities to create a one-to-one relationship with customers and leap ahead of competitors. However, the steep learning curve required to become proficient with big data poses a major hurdle.
While some marketing organizations are finding great success with predictive analytics, your typical marketing professional is feeling a bit overwhelmed by the complexity of the technology, according to Patrick Tripp, vice president of product strategy for RedPoint Global, a developer of marketing campaign management software.
“There’s definitely a perception of predictive modeling and how complex it is,” Tripp says. “There’s a perception of PhDs with red staplers in the basement trying to figure it all out, and nobody understanding it because it’s all advanced statistics and math. It goes against the grain of traditional marketers trying to be creative and innovative.”
Whether it’s de-duplicating data in Hadoop, extracting data from a Teradata (NYSE: TDC) warehouse, or creating advanced machine learning models in SAS, the state of the art in big data-powered decision making has clearly moved beyond the comfort zone of your average marketer.
With a new release of its Convergent Marketing Platform unveiled today, RedPoint is hoping to scale back the complexity without dumbing down the results.
The new release makes it easier for marketers to complete data analytics-related tasks, such as setting trigger events, configuring a machine learning algorithm, or executing a customer segmentation model, without resorting to extensive Python coding or requiring a graduate degree in computer science.
Many of these tasks can be automated to run from an intuitive GUI in the CMP. For example, a marketer can drag and drop icons to tell the system to automatically initiate outreach to a customer based on the presence of key events, such as a customer asking for a password reset or a customer abandoning a cart or a session.
Similarly, the software provides the marketer a list of commonly used predictive models to chose from, such as a regression models that seeks to identify which 5,000 customers out of a basket of 1 million will be most receptive to a certain message. It still requires sophisticated data science techniques and large amounts of distributed computing to pull off, but RedPoint is doing its best to hide those requirements from run-of-the-mill marketers.
“We’re really trying to bring the essence of data science forward to the everyday marketer,” Tripp tells Datanami. “That’s something I’ve seen a lot of vendors try to achieve unsuccessfully. It requires very sophisticated modeling tools and PhDs to use things like SAS. But what we have here is the ability for a marketer to go and grab from a library or a series of very common modeling techniques and be able to incorporate them into a multi-channel campaign.”
RedPoint may not be a well-known outside the marketing community, but the 11-year-old firm has developed a solid reputation for its collection of data management and data quality tools that rival what some of its bigger competitors offer. The software already runs in Hadoop and other distributed computing platforms, and now it’s trying to simplify how everyday users interact with its software, which is something that many big data software vendors are struggling with.
“If we can bring machine learning and predictive data science forward a little bit more for marketer, it would have a huge impact,” Tripp says. “We don’t want marketers to run away. You need to make it successful and simple.”