The short answer is both. Most of Infer’s customers are selling B2B products, so the first order of business is to make sure that the companies they target are good fits for their offering, and then look at who the individual buyers are. Even if a lead comes in with a CEO or VP title, the chances of converting that lead to a customer may be very low if their company is not a good fit for your product.
On the flip side, a lead from an admin at the right company could actually be worth looking at. When someone is expressing a need and there is a fit at the company level, who knows – that admin might even work directly for the CEO. That said, the ideal scenario is to have the right person at the right company, so we start with a lot of company-level signals and then add data about the individual into Infer’s scoring models. And when looking at something like the person’s title, we do deeper analysis to extract things like whether that person works in the IT dept, or is a C-level executive, etc.
Another related signal we factor into a lead score is whether something looks like a spammy lead. Any time you get leads from the web or another signup source, there will be a certain amount of ‘junk’ that you need to deprioritize. Interestingly, when spotting spam, individual-level factors come first. We filter leads out immediately when we detect fake names, suspicious email addresses, or short query distances on keystrokes. Next, we go on to check if the person entered a legit company name into the form.
The last thing you want to look at is the individual’s and/or their colleague’s engagement with your company. For example, if you offer a free trial of your product, we can incorporate behavior and usage information into your scoring model. In addition, marketing automation systems like Eloqua, Marketo, and Pardot capture awesome engagement scores that monitor what pages a lead has visited on your website and what content they’ve viewed. Some of our customers keep their activity and Infer “fit” scores separate at first, while others have us incorporate engagement scores into Infer’s predictive model so they can bring everything together into a single score.
Any other questions or comments about Infer’s predictive lead scoring models, let us know!