Predictive Intelligence Might Pick Out Your Next Customer

It's 2016, and data science is going mainstream. Predictive analytics—the output of models built on complex algorithms for a specific use case—are becoming more accessible to companies in every industry, with platforms focused on business intelligence at the forefront of the movement. In the business-to-business (B2B) world, platforms that provide predictive analytics are able to show sales and marketing teams exactly who to target for their next sale.

Until recently, using predictive intelligence was prohibitively expensive for any organization that wasn't already in the business of big data. Without it, sales and marketing teams pursue leads according to their own set of qualifications, like company size or industry. If the lead turns out to be a bad match then they eat their losses. But now that predictive tools are affordable, any company can use them identify prospective customers.

What's So Exciting About Predictive for Business?

There is a class of emerging tools in the B2B market that use customer data to predict who's most likely to buy next. Some platforms are focused on broader business intelligence, like Adobe Analytics, while others are tailored to specific markets and have their own pools of data that contribute to the accuracy of their predictions. But they've all emerged because of timing: the need for better analytics is there, and the technology is ready to support it.

6sense, a platform that identifies buying signals for their customers, recognized an opportunity to use their founders' experience in both data science and business growth to claim a place in the B2B predictive space. CEO Amanda Kahlow explains why the popularity of these platforms are rising fast:

"There are a few reasons predictive is making such an impression on the B2B market right now. One is the technology—it's here. Big data tech stacks are usable, fast, affordable. Second, the data is here. People are leaving more of a trace than ever as they go about their research online, and they don't have time to pick up the phone and talk to a salesperson. The third factor is that the pain for B2B companies is so great. It's hard to know who's in the market to buy, and for sales and marketing teams at a particular company to agree on an approach. These businesses are ready for a better way, to know what the indicators are that will get them their next customer. Predictive does this."

Forrester analyst Laura Ramos, another leading voice in the B2B industry, spends her time researching predictive marketing vendors. She and her colleagues are already seeing big payoffs for companies that implement this technology: "When you're using traditional tools for identifying leads, they're scored through an automation system that sees activity within your site or campaigns. It's not necessarily about fit. With predictive, we've seen an impressive lift in conversions because of better targeting."


In an Idio webinar, Forrester's Ramos shared a colleague's example of how predictive can increase profits.

Predictive vs Traditional Analytics

Today's marketing stack is full of analytics tools that gather past data to help marketers report on what worked best. They're extremely useful, but come with some setbacks:
  • Keeping data clean is a full-time job, and it has to be done in-house. Cleaning data is mostly manual, so there's no real scalability.
  • Even with clean data, someone has to work to interpret the analytics coming from various tools. This again takes manual effort and must be repeated regularly or data won't be actionable.
  • Even with clean data and good reporting, the results are always based on the past. It's easy to see a trend and attribute it to a past action based on correlation, but often there are variable unaccounted for that make predictions virtually impossible.
Predictive tools, however, are different in nature. They use machine learning algorithms to compute data that's gathered from sources like a marketing automation tool, Google Analytics, a customer's Hadoop data lake, or a third-party site that provides data on user activity. Some initial data is fed into a machine learning model to train it, then the quality of the predictions is tested and adjusted until the customer is happy with the results.

Predictions are typically shared in the form of dashboards or alerts. Customers can see things like which accounts are in-market to buy, which people the customer service team should reach out to, or which prospects should be targeted with the next email campaign.

"Examining past analytics to find companies that look like who you've sold to before just helps you built a target account list. It doesn't tell you if they're going to buy."

Predictive, when used as an added layer in the marketing stack, has the potential for a huge ROI. While initial setup is still somewhat manual, the output of predictive models becomes an automated source of intelligence that entire marketing teams can use.

Kahlow explains a key difference between the old and the new: "Examining past analytics to find companies that look like who you've sold to before just helps you built a target account list. It doesn't tell you if they're going to buy. The data signals that we find to be predictive, that might have an 80% accuracy rate, is the activity data. It's looking at the research people are doing and finding signals that they'll buy in the future."

Navigating the Predictive Market: Hiring, Outsourcing, and Jumping In

As a business investing in a predictive platform, you'll have to do some of the initial legwork. There's a general consensus that a good chunk of time should be spent on deciding how predictive will help reach your business goals before shopping for a vendor.

"Be specific about the scenario that you want the vendors to solve for you."

When that's figured out, approach vendors with the right questions. Ramos suggests what to ask: "Be specific about the scenario that you want the vendors to solve for you. Construct the conversation around two things: one, what is it about the vendor's process that's distinctive? And two, how is their data differentiated from other people's data? They should be able to explicitly express what that advantage is, citing things like quality of data or a deep understanding of your market."

Knowing when to hire and when to outsource is a bit trickier. On one hand, an in-house data science team can be a great asset, even building and maintaining the models themselves. On the other, a shortage of data scientists and the difficulty of scaling big data infrastructure internally makes it attractive to outsource everything. According to some of the leading minds in predictive and machine learning, it's likely that most companies will lean heavily on vendors, making it even more important to vet your choices well.

What Does the Future of Predictive Marketing Look Like?

Predictive intelligence for business is a young and promising industry. Ramos, who also identifies when markets are ready for a full, in-depth evaluation, says predictive for sales and marketing is still in its early stages. But she does expect it to grow and thrive over the next few years.

"I expect some consolidation of predictive vendors," Ramos explains, "along with acquisitions by large marketing cloud companies, and of course for some to struggle as competition rises. The technology will either become part of the martech stack, or it will edge toward specific use cases that will be used for executive decision-making. Either way, the winners are going to be the ones selling the best tailored solutions—those who really understand the business their customers are in."

Kahlow envisions a future where predictive can make far more impactful decisions than which customers to pursue. "Imagine a day when data can tell you what's going to happen in the world. Predict the stock market, tell you when layoffs might happen, or when hardware might fail. This technology will be making predictions on what vertical to go into, or whether to enter a particular market at all. It will affect the direction of your business for years, not months, ahead."

Today it just matters that you get in the game. Companies thinking of buying predictive tools should feel comfortable that the market is ready. With the right preparation, adopting predictive will pay off as the rest of the competition struggles to catch up.