hubspot lead scoring

HubSpot scoring is a way to assign a value to your lead. The value is numeric and can be positive or negative. The score itself has no relevance until you create a criteria in which you identify if the score is good or bad. The score itself is based on adding or subtracting points based on data points and behaviours. The idea is when a certain score is reached, that lead becomes marketing qualified and goes to sales or for other purposes of scoring.

You can score more then just leads, you can use the same methodology for rating how well someone responds to your email campaigns, for customer service and for sales to identify the likelihood to close. Regardless of the way you use it, or what value score you assign to the ‘desirable leads’, the most important part is to get the scoring right.

Since the score will change with every data point and behaviour, it is advisable to review your lead scoring periodically. Depending on the volume or leads and sales cycle you should do a Hubspot lead scoring audit on a quarterly or yearly basis. Setting and forgetting will make the score obsolete over time.

How to use data science to identify your HubSpot Scoring Criteria?

Since HubSpot scoring consists of both demographic/firmographic and behaviour attributes, its ideal to look at them separately. If your business is established it is recommended by HubSpot Analytical Companies to use historical data to come up with not only what your ideal lead looks like but more important what fields are predictive of future success. For those businesses that are just starting out with HubSpot, you have two choices: either guess your idea target lead and their behaviour or wait until you accumulate enough data. Either way, it is best to review your score every quarter or two to make sure that nothing changes.

Scoring Demographic/Firmographic:

To determine what fields to use in your demographic/firmographic score, start by downloading all available data. Ideally, your data will contain success metric such as a sale or earlier indicators such as a sales qualified lead. Without the success metric it is impossible to run a predictive model because there is no predictive variable or data point.

Now that you have all data with your predictive variable, your next step is to remove data points that are incomplete. Any field that has data below 50% completeness should be excluded. You should also remove data points that cannot be grouped such as first name, last name, email because they are all unique or those that are identical.

Your dataset should now be smaller, for those companies that have over 1 million records, you would not be able to run the predictive models in excel because of scale, you would need to use R, Python, SPSS or SAS. Alternatively, you can look at last 12 months of data to make it more manageable.

Next you can run either regression model or decision tree model in order to identify which data fields have the biggest impact on the success metric such as sales or SQL. As HubSpot Analytics Agency we prefer decision tree model as it gives you the interaction between different fields. For example, we can say industry A, B, C with Titles X,Y,Z perform better than Industry Z, D, E and Titles D,B,C.

In order for the decision model to work, your data fields should be standardized. Your HubSpot Data Consultant, can help with standardize fields such as Title or Industry.

Scoring Behavioural Data:

Scoring Behaviour Data is utilizing the same concept as for Demographic/Firmographic however, the data will be inside your HubSpot and certain metrics you want to check include: avg. page-views, what form was submitted, or number of forms submitted, avg. email opens, clicks, delivered and any other behavioural metrics that make sense for your business.

Using the same data analysis technique you want to establish which metrics make the most sense for your business and at what value.

Implementing your HubSpot Score

Now that you have established your criteria for what fields matter and what values can predict a sale or whatever success metric you are scoring against, it is time to create your scoring system. To start identify what metric you want to use, most people use score out of 100, while simpler models can be using a score of 10. Next you want to assign a value point system to your fields, that can be done as part of your regression/decision tree model based or if you want to be less scientific you can assign a value yourself. For example, you can say anything that is super strong will get 30 points, or a specific form such as contact us is automatically at 50 points.

Do not forget to also assign negative values such as unsubscribed or bounced email should automatically have a score of negative in order to prevent those leads from moving forward.

To implement HubSpot scoring, click on settings, then properties, HubSpot Score or create a unique custom score property. For those companies that are using different scoring models for different purposes. Next add your criteria in the Positive and Negative sections. You can use test criteria button if you want to validate a few records. Once you save it will take a few minutes to a few hours to run through your database.

Validating Your HubSpot Scoring

Now that the score has run through, you can validate it by doing a few spot checks using testing model within the properties or alternatively you can download your database and check different scenarios to see if the score matches for the fields that you wish to score against. If it was implemented correctly, then your score would match what you would calculate manually.

To validate if you scored your model correctly, the best course of action is too wait for new leads to make it over to sales or other departments (depending on the purpose of your score) to get feedback if in fact it is performing as it should and there is hopefully a lift in sales or other measurement you are measuring.

Establishing at what score value do the leads make it to Sales

There are a few ways that you can establish when a lead should be sent to Salesforce/ sales or a different department depending on the purpose of your scoring. One of the is to establish how many leads sales needs on a daily basis so based on quantity rather than quality. Second is to look at historical data in order to identify what the threshold should be. And Third combination of the two. As a HubSpot Scoring Consultant, we recommend using historical data to establish the threshold, however, for business needs we realize that a third combination maybe the best bet.

  1. Score based on number of leads that Sales needs to receive on a daily/weekly basis. Look at the last 2-3 weeks, to establish the average daily lead count. Identify what percentage should you be sending over. Establish what score gets you to that number.
  2. Base your score on probability of success. Run a probability model, the model should identify your sweet spot.
  3. To combine the first method and the second. In this scenario you are assuming you have more leads that score high then your sales team can handle. Great problem to have. In this case, increase the score until you get the maximum number of leads you want to send to sales.
Monitoring and Long Term Lead Scoring

Depending on your business it is important to monitor your lead scoring to ensure that it still produces enough leads and that they are high quality. One way to do so is to do a quick audit every couple of months or a year to confirm that your lead score at the time of conversion generate expected results. Another way for faster paced businesses, is to set up a dashboard inside HubSpot to monitor number of leads that come in based on score and the outcome. Keep in mind the score is a moving target. However, if you are sending leads to Salesforce or another CRM at the time that a score has to be above a certain threshold you can use that as an indicator if success of sales is going up or down. Another way to use dashboard with lead scoring is to identify sources that produce high quality leads vs. not so high, if your lead scoring is built right then you should predict sales value prior to sales even receiving the lead. Bonus you can also calculate estimate MROI on your campaigns.

Hiring the right HubSpot Analytical Partner

Most companies will guess their criteria, not test it, and have sales complain about quality, quantity or both. StrategicDB can fix your HubSpot Scoring. What we do is not only set up your scoring based on statistical technics but also, we will also work with you to ensure best practices are used in monitoring your HubSpot lead quality. We also understand that many companies struggle with lead scoring simply because not enough data is available or it is available but not in a grouped way. We will help normalize/standardize your data and put in places ways to fill in data gaps regardless if you are using third party data sources, updating your marketing forms and lead gen practices. Contact Us today.