Deep Learning and AI for lead generation from job postings
How do you know which company or target market will need your product or services the most?
This is a tough question. Often people start with guesses. For instance Bay Area Companies with greater than 300 employees, with more than 10 million dollars in revenue, having an internet website, etc. Then when such a list comes along, an approach might be to distribute this list to your sales / lead gen team. What they do is then find the contacts they have within their network that work in those companies or can make a relevant introduction. The next step is to write an email that is targeted to that connection. Often that email should be personalized. You get the intro and set up a meeting and define your value prop. More often then not the person is not relevant and he decides to make an intro to someone that is relevant. Your lead gen team keeps sending follow up emails so that your connection can make the intro. Few do. Now you supposedly do get an intro, and lets assume that intro does reply to see a demo of your product or value offering. Now you have to guess what the relevant person wants to know, his questions and concerns might be very different from previous meetings. Lets say you pass this hurdle, you finally realize that he does not have the budget or the authority to buy your software or hire your services. All this is assuming that you have done an excellent job in identifying who your target customer is, what their characteristics are, etc.
How about we turn this around? What signals are companies giving that they need your services?
Here is a list of sources where you might be able to get that information from:
1. Job Postings: when companies post a job, quite likely they have the budget and the need. If you can find the hiring manager, you also now know who the decision maker is Glassdoor will also tell you which positions the company is hiring for. If they're looking for new employees in a division related to your product, that's definitely a good sign. The organization is clearly investing in that area of their business.
2. Press and Media Releases: companies make press releases often about new product launches or new strategies. These press releases might contain relevant information about what the company’s top priorities are. Also any new launches that might be coming up.
3. Recent News: news related to a company. For instance tech crunch might publish an article about a company about a new initiative or a new product launch.
4. Recent Talks: talks given by company personals. In these talks the company personals might be discussing topics relevant to your products or service offerings
5. Recent Shares: look at what is being shared by the company on social media such as twitter, linkedin, or Facebook. You should also look at what individual employees at the company are tweeting or sharing
6. Competitor's Press and Media Releases Pages: same thing as above except this time you are scanning competitors
7. Company Blog: just like press and media releases the company blog contains information about what is current and new
8. The Company's Financial Statements: it might be a good idea to check out its most recent financial reports on the SEC’s website. Checkout especially check out “risk factors” section
9. Quora: using Quora to understand what your prospect is hoping to learn
10. Datanyze or similar tools can tell by analyzing a company website what tools they are using.
Job Postings : a case study
Now for the purpose of our discussion lets take Job Postings. One way would be to set up alerts on some keywords related to your business. However, there are several problems with this approach
1. The keyword might not be exact. For instance, you are providing services for Deep Learning. You might add the keyword deep learning but it might turn out that the job posting might say Artificial Intelligence, Machine Learning, or Predictive Analytics. All these terms are semantically close to Deep learning however since it is not an exact match the alert won’t get triggered
2. Alerting might not even be available in many systems. Imagine managing these alerts across multiple sources!
3. The keywords are not finite and keep evolving. For instance, you just published a post how deep learning can help in anti-money laundering patterns. Clearly a job posting that talks about anti-money laundering is now relevant to you. However, you would never get an alert because that keyword is not in your list
How can deep learning help here?
Deep learning is quite good at semantically understanding words and a system can be trained to understand semantic similarity between two sentences. So you can build a system that say scans a job posting, but also scans your website, because that is where all the content is getting posted, and then compares sentences in the job postings with the sentences in your website’s content to come up with a similarity score. In the end you will get job posting ranked with similarity to your product, services, and content. From there you can easily get the name of the company, the division and then start targeting people in linkedin with targeted messages.
Is job posting enough?
In a way they are a great source. But a full system should be architected in a way that each source should be configured as a stream. And that source should be scored for similarity there by notifying you if a company pops up in the radar that is relevant.
In our experience, we have seen that the most effective way to get a qualified lead is getting the timing right. Most lead generation systems fail at that. At SublimeAI we have built a novel system that turns the lead generation process upside down and uses AI to get leads which are qualified from a timing perspective. We use this system to do lead generation ourselves. To learn more please email at email@example.com.