As a former sales engineer and product marketer, I’ve seen both the sales and marketing sides of the sales cycle. Even further back, as one of the people who built early SFA and CRM systems, I’ve always seen the importance of a closed loop system. We’re still looking for such systems. The amount of data collected in both sales and marketing has had developers focused on each subject with only a little thought given to a true integration. In recent years, that has slowly begun to change; and artificial intelligence (AI) might be providing the final piece necessary to supply a more robust solution.
While there are many places where sales and marketing overlap, the most critical is the lead cycle – how to understand, qualify, and track leads. It has been an almost intractable problem thanks to the lack of integration between systems and also in the complexity of lead qualification. AI can provide insight to help speed and improve the accuracy of analytics that provide organizations the ability to improve sales.
Marketing can always generate leads. The challenge is not providing names, it is in qualifying leads. If someone interested in your product doesn’t have budget, a purchase is not going to happen. Well, not always. What if you’re in a “land and expand” account, on department doesn’t have budget, but the sales team knows people in the CFO organization and can prove ROI? An enterprise sale might still happen. What can be seen from that example is that qualifying leads is a bit more complex than many believe.
Even when only one person is involved in the decision, lead tracking can be complex. I have worked in more than one company where the sales and marketing systems didn’t talk. Marketing might think it’s gathered a new lead, but the sales team talked to that person months previous. On the other side, a salesperson thinks she’s closed a sale based on a sales lead, but the initial entry in the SFA system came from marketing while not tracking the original marketing lead. Who gets credit in those situations?
Companies are working with AI to improve analysis of all customer contact points to both identify leads and weigh lead quality. That includes ingesting information from web pages, email campaigns, phone calls, and much more.
From talking to a number of companies in the space, I’ve seen they’re all young and their AI focus has been on natural language processing (NLP). The background classification and correlation are being done in statistical methods.
One interesting area is in that analysis of buying stage – where in a purchase cycle a person or company is at a point in time. “While a purchase is a binary event, the move down a sales funnel has starts and stops,” said Viral Bajaria, CTO and Co-founder, 6Sense. “Modeling time series activities through Ensemble Learning is showing promising coverage and accuracy gains for predicting buying stage.”
One point from my soapbox: Notice the mentions of statistical analysis, NLP and random forests in this one example. It should always be remembered that AI is a tool, it is not the solution. Building an efficient and effective solution means using the right tool at the right point of the process.
The 6Sense team not only uses NLP, but they’ve also built some classifiers to analysis types of pages and emails, link them to campaigns, and learn from prospect passages through different touch points. When I asked about which types of AI systems they used, the honest, obvious, and correct answer was the old “it depends.” As Mr. Bajaria pointed out, when correlations have low cardinality (Low cardinality: Only “Yes” v “No” responses; High cardinality: 100,000 IoT devices sending different information), random forests can be used to test for multiple correlations and combinations of features. When the cardinality becomes large, it’s better to apply deep learning.
The company is young, so they are still testing different AI models and engines against a growing body of training data. That’s appropriate for the stage of maturity of AI in business. What’s also important to know is the ecosystem. While it would be nice to see SFA and CRM merge, it’s important to understand the existing landscape. Their partner page shows they are building relationships to underpin and share information with multiple players, both large and small.
It’s one thing for marketing to talk about the number of leads generated, it’s another to turn leads into customers, and quite another to understand how to understand leads in more detail in order to improve movement down the sales funnel. There is a lot of data, potential information, floating around that can help business to better understand prospects, to better track and analyze leads. With the ever expanding amount of touch points and data, artificial intelligence is becoming an important tool for combining corporate information in order to better analyze leads.