Blog with useful information and tips about Marketing Automation

Blog

Dive into the world of Marketing Automation

Practical & helpful knowledge that makes your daily life with Marketing Automation easier!

Lead scoring part 1 - How to calculate lead scores

Do you manage to generate many leads, but you don't know which ones are interesting for your company? Then it's time for lead scoring. Marketing automation tools offer a powerful instrument with automated scoring to better assess the potential of your generated leads. In this first part of the series on lead scoring, you will learn about different analysis methods to predict the potential of leads. Essentially, it's about understanding the different methods you can use to score leads and the advantages and disadvantages of each approach. But one thing at a time:

What is Lead Scoring?

The goal of lead scoring is to predict the purchase probability of leads. Leads are compared based on various characteristics and activities in order to specifically identify leads with a higher purchase probability. This allows you to invest your time and money specifically in leads with higher buying potential. Lead scoring is based on assigning numerical values to characteristics and activities, e.g. 20 points for downloading a white paper or if a lead has the desired job title (this is also called grading). The points are added up for each lead.

To use scoring in a meaningful way, it is essential to implement a lead qualification strategy. For example, in most companies, there is a process where a lead is passed from the marketing team to the sales team. Typically, this is when a lead moves from a Marketing Qualified Lead (MQL) to a Sales Qualified Lead (SQL). The lead qualification processes can then be mapped and automated in the marketing automation tool or CRM.

Analysis methods

The first step is to determine factors that predict a certain outcome (e.g. lead qualification or purchase). There are various methods for this.


Qualitative methods such as interviews
Qualitative methods are based on subjective assessments of persons such as employees or customers. For this purpose, the persons of interest are interviewed or asked to fill out a questionnaire. In terms of scoring, two parties come into question: sales team members and customers.

Sales teams often have a pretty good idea of which measures and contents work well with customers and which do not. They also know from experience customer-related characteristics that are promising. This experience can then be transferred into a scoring model, and assets that employees say work better can be given a higher score.

Another option is to ask customers directly. The customer perspective can provide important insights into what helped them at what point in the sales process.

Pros:
  • flexible application
  • Possibility to ask for background information and clarify ambiguities

Cons:
  • In order to make valid statements, a representative and sufficiently large sample is needed
  • Evaluation is time-consuming
  • Transfer of qualitative data (statements of a salesperson) into a quantitative score difficult
  • Analyses have to be carried out again and again and scoring has to be adjusted manually

Attribution Report
An attribution report is a useful tool for determining the effectiveness of campaigns and individual marketing assets. Marketing attribution allows marketers and sales teams to see the impact of marketing efforts on a specific goal, e.g., purchase or conversion. This helps to evaluate campaigns and assets. For example, through which campaign did a lead come into contact with your brand or the last interaction with your brand before buying a product. These interactions can be weighted differently depending on the impact.

Pros:
  • Relatively easy to implement
  • A good overview of the entire customer journey

Cons:
  • Only online marketing
  • Weighting of individual touchpoints often not easy
  • Scoring has to be adjusted manually over time

Quantitative data analysis such as logistic regression analysis
The basic idea of logistic regression analysis is to predict the expression of a categorical variable with one or more variables. What does this mean? In simple terms, we are trying to statistically predict whether a lead will do something specific or not, e.g. will a certain lead convert or will a lead buy one of our products? So we want to predict a yes/no question for each lead. For the prediction, we take the expression of one or more other variables. Here's an example: Let's say we offer a product and we have a specific page with product information on our website. Now we can ask ourselves whether the number of visits to this product information page increases the probability of a purchase. Or to put it another way, how likely is it that a lead who visits this page 1 time will eventually buy this product? What about 2 visits, or 3, and so on. Perhaps we find that the probability of a purchase is increased if leads visit a product page more than 3 times. This can be transferred to our scoring model and the appropriate steps can be taken, e.g., a sales rep should then contact the lead by phone.

Pros:
  • Quantitative analysis facilitates the conversion of results into scores
  • More precise compared to previous methods

Cons:
  • Complex and requires statistical knowledge


Predictive scoring with artificial intelligence
Many marketing automation tools now also offer lead scoring through machine learning. This involves an algorithm making predictions about the decisions and behaviour of leads based on previous data. This means that an algorithm "recognises" commonalities of successfully converted customers by analysing thousands of data points, and can thus reflect the potential of each lead in a score. This process is constantly being adapted and becomes more accurate as the data points increase.

Pros:
  • Many marketing automation tools include predictive scoring
  • Scores are automatically adjusted
  • High accuracy if enough data is available

Cons:
  • Sufficient data must be available
  • Predictive scoring often only included in higher priced marketing automation tools
  • Data scientists are needed for custom developments

Scoring models are useful to assess the potential of newly generated leads on the one hand and to determine their readiness for purchase during their customer journey on the other. There are several methods that can be used, depending on resources and competences.

In the next part of the series, we will show you how to manually implement a scoring model and how to improve it step by step.

References:

https://trailhead.salesforce.com/content/learn/modules/pardot-lead-scoring-and-grading/get-started-with-lead-qualification
https://www.pardot.com/blog/multi-touch-magic-pardots-campaign-influence-attribution-models/
https://www.datacamp.com/community/tutorials/logistic-regression-R

https://help.salesforce.com/articleView?id=sf.pardot_einstein_behavior_scoring.htm&type=5

By accepting you will be accessing a service provided by a third-party external to https://www.waym.ch/

Get your Marketing Automation news

Waym - Die Marketing Automation Agentur in Bern

Waym Marketing Automation
Schlossstrasse 3
3098 Köniz
Tel: +41 31 371 63 03
info@waym.ch