I’m a big fan of marketing automation. My experience goes back a few years. I was just getting settled at a new company when our CFO received a renewal invoice from a marketing automation vendor. She didn’t know how we were using the software, and neither did I.
Over the next 30 days I decided to immerse myself in all aspects of marketing automation. After all, you can’t make a recommendation to renew a platform without reviewing the alternatives. It was a steep learning curve that consisted of mock workflows, sample email campaigns, and CRM integration testing. I learned a ton. In the end we changed platforms and began the process of migrating our contact histories and email lists. That’s one benefit of working for a small company: You get to work in Marketing and IT.
Eventually, we were able to implement most of the core platform functionality. We launched a new blog, created landing pages, ramped up our email capabilities, and developed a few workflows. We had the inbound marketing channels under control, but something about our marketing automation capabilities still bothered me. Many of our automated actions were based on gut feelings and intuition.
I was left wondering:
· What exactly were we automating?
· Did we truly understand the customer buying process well enough to automate many of our interactions?
I know that many — most — marketers struggle with these same questions, and the challenges aren’t limited to marketing automation. Think about the automated decisions being carried out across ad servers and customer service desks. Are the analytics running behind the scenes as robust as the automation software that’s serving up the content?
Analytic prowess will separate the winners from the rest of the pack in digital marketing.
Of course, best-in-class analytics today are not the same as they were ten, or even five, years ago. Back then you could develop predictive models with static data from your data warehouse and use the scoring output to develop complex rules-based systems or segmentation lists. The process would easily take six months to deliver value. Those days are long gone. Now, a wider breadth of data flows around the clock, the traditional role of the middleman is being eliminated, and new forms of intermediaries are making informed decisions at lightning speed.
Today, the machine-learning algorithms we deploy must run against flowing data streams and adapt over time to changes in market dynamics and consumer behavior. Subtle changes in the customer buying process are buried in the data, and we can fish them out — in near-real time. In the marketing world, this is exciting stuff!
Consider how one global travel company is applying data science to power its customer management program. By looking at nano-segments of individual travel habits and browsing patterns, the company was able to double the take-rate of online campaigns. This granular analysis can distill other findings such as which upcoming trips are business versus leisure, the degree of price sensitivity for specific trips, and the propensity to upgrade seating, to name just a few examples. All these factors contribute to a truly personalized experience for the customer.
The use of data science to produce signal-based treatments is revolutionary on a very broad scale. First, we now have unprecedented amounts of accurate, predictive information. Second, analytics are now focused on individual behaviors, not demographics or macro-segments. Third, the extraction and application of real-time predictive signals gives marketers entirely new ways to manage customer relationships. It’s at this point, when advanced data science can completely reinvent the customer experience, that marketing automation makes sense to me.
The relationship between marketing automation and data science is symbiotic. And as we look over the horizon at content automation and location-based media, these dependencies will only become stronger. So get to know your resident data scientists, and share your marketing automation vision. You may be surprised at what’s possible.