I was asked a question recently regarding lead scoring and I think I botched the answer a bit. I kicked myself for a while but then realized that I didn't necessarily botch the answer, it was just a very complicated question with lots of factors and variables.
So let's talk lead scoring a bit.
First off, lead scoring is determining how viable a lead is against a collection of other leads. If a lead is a guaranteed sale, it gets a high score and vice-versa. We'll use a scale of 1-100 for this example.
Now let's parse the customer list down a bit by reducing a person's buying interest and capability by examining some of the known factors.
- How did we collect the lead?
- How long were they engaged?
- How well did they fill out the forms?
- Are there strong competitors in their area that might impede our sale?
- Have they bought from us before?
- Have they engaged again after the initial contact?
- What kind of capital do they have (ie. do they have the money to pay and at what scale)?
Now that we have a list of questions, we can start creating a matrix (not a figure) for each of the answers. We can't assume that just because someone filled out a web form that they get negative points if they also stayed engaged on the site after the form was submitted. So we'll need to delve a bit into our stats and analytics. This can be done with some programming (see our Programming For Marketing page) but let's assume you've got a programming team that can work with the data exports and stats.
We now create a bar graph of engagements based on that data. Don't think of a lead score as a single score at all, someone can be a good lead (or bad) for a number of reasons. Our bar graph would contain information based on the mitigating factors we've identified. In this case let's say Competitors, Interest Level, Capital. We'll create a score in each of those areas.
A couple of typical cases might be;
Someone who submitted a form online, then stayed to read a paragraph, then went to the main site for more information and hit several pages would get a 100 in the engagement graph. That person has never bought from us before but has been in business for a long time, they'll get a 40 in the competitors score since we know they've bought from someone else in the past. A quick D&B lookup using their API tells us that the customer has financing and has spent money with those competitors in the past, so we'll give them a high capital score of 100. Now average the three graphs to get a score of 80.
A user submitted a form but did not fill it in completely. The email address was from gmail and not a corporate domain. They will get a very low score, maybe a zero on the Interest graph. There are no competitors in their area according to our geo lookup so we'll give them a high number in competitors, perhaps 100. Now we have some fake data and no way to look up the credit so we've got to give them a low score for credit. In this case a complete failure so we'll have to give them a 0. Total score is a 33.
So how did we get these values? Therein lies the fun part. We've got to identify every possible scenario and account for that with a number. This is where the programmers come in. Programmers can do lots of If-Then statements to do a lot of this work automatically so it's worth dedicating some time and resources to the task.
For example; IF web form submitted AND after submission engagement > 1 (they hit other pages) AND len(company name field) is not blank THEN score = 100
This is by no means a complete tutorial on the subject but as I found out when answering the question earlier this week, you could drive yourself crazy trying to account for every possible scenario. My suggestion is to focus on the big ones and generate an If - Then chart on your own.
Hope this helps!