categories

HOT TOPICS

NEWSLETTER

If you are considering becoming a 1M/1M premium member and would like to join our mailing list to receive ongoing information, please sign up here.

Subscribe to our Feed

Building an AI-Driven Sales Optimization Company: Vincent Yang, CEO of EverString (Part 4)

Posted on Saturday, Aug 1st 2015

Sramana Mitra: I’m going to start asking you very deep questions because I did a startup in the sales lead generation area back in 1997. It used NLP and the whole AI stream. It was a bit early. It was well before the Internet had completely established itself. I know a lot about this area in general. Tell me more specifically about exactly what you do by applying your technology to the sales lead generation.

Vincent Yang: We have two use cases. First, we come into any company. We say, “Let us link to your internal work flow data whether it’s CRM or marketing automation data because we want to learn who, historically, are your good leads and who are the bad leads.” We do this for labeling. Then we start using our crawling engines to crawl every single information about those leads.

The underlying assumption is there must be something in common about those good leads. Why are they all converting? Why do those leads that you interact with never convert? We wanted to use a mathematical formula to describe it. This is what we call audience selection.

Sramana Mitra: Tell me more about what parameters you map on in determining the properties that you’re basing your leads distinction on.

Vincent Yang: On an average, we analyze tens of thousands of features. For example, some of the basic ones are revenue, employee size, and industry. Those are very basic. Let’s go to the deeper ones. For example, product offerings. For those leads, what exactly are their product offerings? This is very hard to do because product offering is not like one word for every industry. We use NLP to analyze each web page. We extract all the keywords out. That’s just one of the category of features that we are using.

The most difficult part is when we apply machine learning on top of that, everybody can do machine learning based on numerical data. How do you actually do machine learning on unstructured data, which is text. That’s something we spend a lot of time and even filed a patent. We filed four patents around this area. For example, hiring. Who do those leads hire? Are they hiring marketing people right now? Every single hiring indicates some business indication. If you’re hiring customer success folks, what that means is it’s a very good time for ZenDesk.

I started to realize those small patterns and signals when I was working in investment banking. I remember there were two companies which we ended up investing in. They started to announce that they were hiring Head of Global Sales. Why would you hire Head of Global Sales? It clearly means that you’re expanding globally. Hiring is a very strong category. Then things like news. How do we actually analyze the news? How do we extract the news out? That’s number one.

This segment is part 4 in the series : Building an AI-Driven Sales Optimization Company: Vincent Yang, CEO of EverString
1 2 3 4 5 6

Hacker News
() Comments

Featured Videos