- Dell implemented predictive analytics for a broad range of marketing tactics
- A consistent cycle of planning, piloting, testing and measuring results is required to effectively leverage a new technology
- Marketing practitioners should take an expansive view of how a new technology can impact their organizations
A challenge for every organization is the need to balance stability and change. If there are too many innovative changes being implemented at the same time, near-term performance will be compromised. But if there’s insufficient readiness to change, the organization will fail to respond adequately to new opportunities or requirements.
For Dell’s Enterprise Global Demand Generation organization, this balance is maintained by a dedicated innovation team that works with each marketing group to evaluate and help implement creative solutions. “We act as internal consultants to the marketing organization,” explained senior demand generation strategist Jeff Siegel during a recent conversation about how the Round Rock, TX–based computer technology company fast-tracked its ambitious implementation of a predictive analytics solution over the last 18 months. “We sit outside day-to-day execution, and our job is to drive continuous innovation within marketing.”
Dell’s recent predictive implementation began in the summer of 2015, when its evaluation of potential solutions led to the selection of 6Sense as a predictive technology partner. “That type of intent-based predictive analytics was appealing, because it extended other efforts that we had previously done at Dell,” Jeff explained. “We liked the concept of a vendor having a large data network that could be used to monitor engagement, and using that intelligence to predict where buyers are in their journeys.”
After Dell began engaging with 6Sense, the next phase of the project involved several months of planning and infrastructure-related tasks to support a broad-scale implementation of predictive-enabled marketing strategies. This work required partnering with Dell’s IT team and working closely with the 6Sense customer success team to connect and integrate Dell’s expansive data systems and establish the necessary security and privacy safeguards and clearances.
“By January 2016, we had operating models, and we started doing pilots in the first quarter,” said Jeff. “Then in Q2 and Q3 we branched out to other delivery vehicles and additional predictive models. Each new vehicle comes with a full round of testing, and making sure that the data systems support it. Each new vehicle also has its own timelines and different execution schedules.”
While the project focused initially on improving the precision of targeting for teleprospecting programs, its scope quickly expanded once the effectiveness of the predictive insight was demonstrated. The Dell team began scaling across multiple predictive models to support various lines of business at Dell, and began to expand the program’s reach globally to Dell’s other operating regions. The Dell team also expanded the scope of the project, enabling marketers to use predictive hypertargeting not only for outbound tactics, but also for inbound lead generation, email nurture, media targeting for banner ads and social media, and orchestrating how sales reps interact with customers.
“We’re now building predictive intelligence into our backend systems,” said Jeff, “so that when a lead comes in or when we target accounts to go to market with a campaign, we have more insight into determining what’s the next best activity they should be routed to – based on understanding how heavily engaged that account is and at what stage of the purchase cycle.”
When asked how the results of the project have been measured, Jeff cited gains in conversion rates, pipeline generation, deal win rates and average deal sizes, as well as improved deal fidelity – the percentage of prospects who buy the type of product that they have been targeted for, instead of proving to be interested in a different offering. For example, in one teleprospecting pilot, Dell saw a higher than 70 percent increase in target-to-SQL conversion rates, a 25 percent uplift in opportunity open rates and 1.4x higher deal fidelity when compared to a standard control campaign. Jeff stated that while the expectations for most implementations are not quite that high, the Dell team has seen similar upticks in performance consistently across its predictive campaigns.
Based on his experience at Dell, Jeff recommended the following success factors for marketing practitioners and technologists to focus on when implementing and optimizing a predictive solution:
- Make sure you have a strong foundation. The organization must have the right systems, data integration, lead management and campaign execution processes – whether automated or manual – to gain the intended value from a predictive solution. For example, campaign managers need to understand when and how to deploy hypertargeting across multiple delivery vehicles. Clear service-level agreements and processes across marketing and sales for qualifying and following up leads are also essential. If there are gaps in systems or processes, identify what needs to be done to eliminate them.
- Rely heavily on pilots and testing. Rather than sitting outside the execution environment and dictating changes, partner closely with the groups that can benefit from predictive modeling (e.g. campaign teams, teleprospecting, digital marketing, sales) and determine how the innovation can work for them, then pilot it and prove that it works – or if it doesn’t work, figure out how it needs to be adjusted.
- Check that you have the content to support your targeting. If the predictive solution is capable of precisely segmenting prospects with specific interests, then the content that is provided to those prospects must have the same level of specificity. For example, if predictive intelligence identifies that a prospect is interested in a particular type of server, sales enablement scripts must focus on that specific product rather than exploring the entire server category.
- Don’t focus exclusively on the highest-propensity accounts. While a hypertargeted account-based marketing approach is valuable, it’s still necessary to continue broader, above-the-funnel engagement. “As soon as you target only those customers that are already looking to buy, you stop talking to those who will want to buy at some point in the future,” Jeff explained. “You still need to apply budget to that broader outreach to keep bringing new people in.” While buyers that are identified as having initiated an active buying process can be routed to a sales rep, earlier-stage prospects can be routed to a lead development function or nurture flow.
- Work with innovative partners. Jeff emphasized the need to partner with other organizations and groups, including solution providers, media partners and internal IT teams, to fully leverage the insight that is becoming available from predictive analytics. He cited Dell’s success in working with Oracle to set up a data mart integration so that 6Sense data for Dell is now available for programs executed via its programmatic system. Dell also developed innovative relationships with media partners such as Forbes to enable specific hypertargeted media campaigns on their properties. “A lot of the infrastructure of our industry is still rooted in older ways,” he explained. “Now that we have this level of customer intelligence, we want to make sure that’s how all of our third-party partners work with us to go to market.”
- Look beyond vendor use cases. Think big – develop a broad vision for predictive analytics, based on a holistic understanding of how it can provide both buyers and sellers with the information they need. Think about how to apply predictive insight to all of your marketing and sales efforts. Instead of limiting the application of a predictive solution to a single use case or scenario, think about how it can support your entire marketing stack – your vehicles, content and integrated campaign execution plans
“As a whole, our industry can benefit from thinking about what our next steps are,” said Jeff. “How do we take these new technologies and integrate them across all of marketing – how do we weave it all together? As practitioners, it’s up to us to take these tools and weave them together into a cohesive system. So I’m very interested in hearing from others about what breakthroughs they are making.”
Editor’s Note: To read more about predictive use cases, see Kerry Cunningham’s post “It Slices, It Dices: Problems Predictive Can Solve”.