Nuts & Bolts - Case Study: Dell Uses Data to Improve Social ROI
No one has to convince Dell's C-suite that social media marketing matters. Michael Dell himself—chairman and CEO of Dell and the man behind @MichaelDell—started using SMM in 2006. That meant that a couple years ago, when the social media analytics group made an interesting discovery, rollout of a new social media marketing program wasn't far behind.
What the group found was B-to-B buyers of servers, networks and other expensive items were performing research and asking questions on social media before reaching out to the call center to convert. Buyers of lower-priced items tended to ask questions on social media, then convert online. That prompted the group to ask a question of itself.
"Can we actually start to build a predictive model which tells us, based on their consumption of content [whether they'll buy]," asks Munish Gupta, director of social media analytics at Round Rock, Texas-based computer company Dell.
The answer was "yes." In 2012, Gupta and the social media analytics group worked with Northbrook, Ill.-based management consulting firm Mu Sigma on the development and execution of those predictive models.
"When a customer … comes on social media, they are probably trying to be more open in terms of what they want to see and what they want to learn," Gupta says. "I think that's a behavior which is converting over to a better performance or a better close rate."
To create the models, Gupta says Dell used a mix that included its social media and online sources, as well as internal CRM information.
The models keep track of which content matters at various points in the customer lifecycle. Models take into account the company size, location and industry, as well as where the content consumers are in the customer lifecycle. The models pay attention to a whole lot more than the last click, he says.
In the pilot program that started and ended in 2012, Dell saw social media marketing using predictive models result in three times the normal close rate. Gupta says "normal" is the close rate Dell sees in its other marketing programs, which include other channels.
"The way we actually defined the close rate was the sales opportunities which we were actually able to book, divided by the number of opportunities which we had created from those customer targets," Gupta says.
Helping leads along also shortened the sales cycle by 25 percent, he says, and Dell saw 75 percent accuracy in customer targeting with the predictive models.
According to Gupta, those numbers have held fairly steady since the program's rollout in late 2012.
"We are continuing to expand this program," Gupta says. "We started this program in North America. Now we are actually expanding this program in other regions. … We are not only looking at this program … for generating customer targets, we're also looking at it as our broader strategy of 'What is the type of content we should create and what is the type of content which is useful to our customers?'"