You can’t read an article, go to a conference or go into a marketing planning session without discussing “predictive analytics.”
Because today’s reality is that most marketers are still trying to get their arms around digital marketing and demand generation, they simply don’t have the ability — or infrastructure — to fully utilize this advanced capability. However, this hasn’t slowed down the marketing trend-setters who are diving into predictive to accelerate marketing and sales efforts.
The Seductive Power of Predictive Analytics
What's the appeal of predictive analytics?
There’s increasing interest due to emerging predictive analytics vendors leveraging data scientists and technology to develop real-world solutions. What makes it particularly exciting today is that we have never before had so much data and processing power. This means we have tons of signals to refine predictions, and we can do it at higher speeds and lower costs than ever before.
Data is captured from numerous sources — internal (e.g., marketing automation like Marketo and Oracle Eloqua, and CRM systems from providers like Salesforce and Microsoft) as well as external (e.g., third-party data providers) — and then processed by predictive software to generate models to achieve specific business goals.
Am I Ready for Predictive?
Predictive isn’t for every B2B organization.
Predictive analytics has the biggest potential when marketing organizations have core components in place. Because the technology is emerging in front of our eyes, I asked a few of the leading predictive pros to weigh in on this question. Adding perspective to the readiness factor is: Sean Zinsmeister, director of product management at Infer; John Bara, CMO and president at Mintigo; and Amar Doshi, VP of products at 6Sense.
So, how do you know if your organization is ready?
How Much Is Enough Historical Customer Transaction Data
The most important factor in building an effective predictive model is having enough historical sales win/loss data, according to Zinsmeister. Depending on the provider, the exact number varies pretty widely on exactly how much transaction data you'll need.
Some providers look for a minimum of a few hundred examples of the customer organizations they want to target, while some predictive providers recommend a minimum of a thousand customer opportunities. Most importantly, Doshi stresses the importance of having enough “deal” history and transaction volumes to make buying behavior data valuable in models.
Do You Know What You’re Trying to Predict
Like any smart investment or initiative, predictive starts with defining and agreeing on business goals and desired outcomes. Everything starts with understanding what you’re trying to predict, according to Zinsmeister. Companies should start by defining the business challenges they’re trying to solve before launching any predictive initiative.
Doshi also underlines predictive is optimal for organizations that “sell products or services that are 'considered purchases,' where prospects conduct online research before engaging with the organization.” This captured data is a vital ingredient to predict future success.
Marketing Fully Aligns with Sales
A powerful outcome of predictive is its ability to identify buyers faster and better target the accounts that fit your optimal profile. This increases the odds of winning new opportunities.
Therefore, success relies on sales-marketing alignment. Mintigo’s Bara advises that your marketing organization must be fully committed to helping sales win and must also accept equal ownership and accountability for the entire sales process.
Why? If marketers are going to use predictive, simply providing leads to sales with basic firmographic or professional contact data isn’t enough. Predictive must be baked into the entire marketing-sales workflow and customer journey.
Your Database and Systems Are Ready
All agree you need core systems and technology in place — such as marketing automation and CRM — to maximize the value of predictive analytics. These systems capture and house the key marketing, event and web data required for predictive work, according to Doshi. Surprisingly, hygiene or the actual size of an organization’s database (number of contacts) is usually not a critical success factor for predictive initiatives: “This is a common misperception we see all the time,” states Zinsmeister. “The reality is predictive works with datasets that are small or large, clean or dirty.”
The bottom line: predictive analytics isn’t for every B2B marketing organization. At this stage, its success depends on having the core requirements and clear purpose outlined above, as well as a commitment to testing, the smart use of customer data, and working closely with your preferred predictive providers to get meaningful results you can repeat and scale.
It may be early, but many are predicting a bright future for predictive analytics.
Title image Thomas Brault