Platforms

How to Develop a Product-Scoring Model

BY: Maude Belanger

PUBLISHED: 8/3/2023

Nowadays, it is well known that understanding the interests and preferences of your prospects is key to delivering personalized and targeted messaging. Lead scoring is a highly effective tool to identify the most promising leads and to help sales prioritize their efforts; however, lead scoring tends to fail when sales teams are organized around lines of business or products.

Product scoring is an emerging alternative to lead scoring. Marketing operations professionals can leverage product scoring to help tailor communications to prospects based on their interests and also to assess when they may be sales-ready for a particular product.

While lead-scoring strategies are well-documented, information on implementing a product-scoring model can be scarce. In this article, we will provide you with a step-by-step guide on how to develop a product-scoring model in Marketo, empowering you to effectively prioritize and engage your prospects based on their product interests.

Step 1 – Determine the desired outcome

 

Before diving into the technical aspects, clarify your goals and evaluate whether using product scoring is the best way to go. Consider how you plan to use the product scores. Are you aiming to tailor communications based on prospects’ product interests? If so, have you considered other solutions than product scoring that might be faster to implement for similar results?

Or are you looking to revisit how MQLs are identified to align with your sales department, where reps are responsible for selling only one product? In this case,  have you considered the impact of changing your MQL process will have on your lifecycle model?

Depending on what you are trying to accomplish, the architecture of your scoring model and other operational programs will vary, processes might have to change, and you will want to make sure you are well-aligned right from the start with the other departments which such changes will impact.

Step 2 – Identify the similarities and differences between products

 

To structure your program and model effectively, start by identifying the groupings or lines of business within your organization. Determine if these groupings share common elements, such as Ideal Customer Profiles (ICPs). If similarities exist, consider centralizing the demographic score while compiling the behavior score on a per-product basis. Simplify your model as much as possible, centralizing what can be centralized and keeping scalability in mind.

Be mindful of leads who demonstrate brand interest without showing specific product preferences. There are multiple ways to account for these signals, and the one that typically works best for our clients is to assign lower scores to these non-product-specific actions across all product scores and create a Person Score to identify leads with high overall interest but unclear product preferences.

 

The marketing team can later add these leads to campaigns with a call to action that helps identify those leads’ main product of interest.

Step 3 – Tie the interest signals to their associated product

 

Associating actions with their corresponding products is the foundation of product scoring and requires a combination of methods:

Naming Convention: Establish a disciplined naming convention that includes the product(s) tied to a program or asset.  Since each campaign should be tied to a Marketo program where program statuses are updated to reflect a person’s engagement with the initiative, it becomes easy to score only the actions that reflect an interest in the product.

Using this methodology, you can, for example, add 20 points to the “ABC” product score every time a member of a program containing [ABC] in its program name gets its program status updated to “Attended”.

You can also deduct points when someone has not positively engaged with any asset or campaign tied to the ABC product. This can be done by using a combination of filters such as “Not Opened Email, Email contains ABC”, “Not Program Status was Changed, Program Name contains “ABC”, success is “true,” etc.

Maintaining a consistent naming convention for new programs is a straightforward task. However, to achieve success in product scoring, it’s essential to revisit and update the name of existing, active programs to ensure they align with the established naming convention, enabling accurate and effective product scoring.

Web Activities: Identify all the URL strings corresponding to each product. While some URLs may directly contain the product name, they can often include keywords associated with the product’s solution or with the pain points it addresses.

But there are more than web page visits and clicks on pages; consider including in your scoring model other behaviors that reflect an interest in a product but might not be associated with a standard Marketo trigger. Using custom javascript, you can, for example, increase a product score when someone watches at least 75% of a video tied to a product, when they favorite a product in your web interface, etc.

Forms: Forms can be used in different ways for product scoring, such as by identifying which form is tied to a specific product or by identifying picklist field values that indicate an interest in a particular offering. We’ve seen many instances where forms were not leveraged to their full potential for product scoring.

 

If you have a demo form for example, make sure you capture the product the person is interested in getting a demo for rather than keeping this form generic. It will help associate the product score that should be increased, but it will also accelerate the routing of the request to the right sales representative and allow you to tailor future communications with this lead appropriately.

Interesting Moments: Interesting Moments is another way to recognize an interest in a product if they follow a naming convention that incorporates a product identifier.

Step 4 – Identify the thresholds

 

Determine the score thresholds that will trigger the actions your organization wants to see occur (ex.: addition to a product-specific nurture program, considering the person as MQLed for a product, etc).

We recommend using product grades as thresholds rather than a total product score. The product grade is obtained by combining the product-specific demographic (which reflects the level of fit with the ICP) and behavior score (which reflects the level of engagement) into a matrix where the demographic score corresponds to a letter, and the behavioral score corresponds to a number.

This method allows more flexibility and granularity than a simple person score threshold (e.g., “100 points”) and can reflect your organization’s priorities with respect to demographic fit and engagement.

For example, a person with a low demographic score (26-50) is in the “C” tier for fit. They will only be considered MQLed if they have very high engagement (70+), in the “1” tier for engagement. However, a person with a very high demographic fit of 76 (in the “A” tier for fit) will qualify with a much lower behavioral score, since they are presumably of much higher interest to the sales team.

 

Different combinations or grades can be used to trigger other actions such as sending a personalized offer, adding the person to a nurture stream, etc.

This matrix idea can also be used for combinations other than demographic and behavioral scores. In some cases, organizations might need to sum scores together. In such cases, using external tools such as FlowBoost can be a great way to achieve the desired results.

Whatever method you choose to identify the leads ready for the next step in their journey, make sure their product of interest is clear. Stamping the name of the product of interest into a dedicated string field once they reach the MQL stage for that product is a great way to communicate which product-aligned resources the prospect is engaging with.

Step 5- Stay organized!

 

It’s good practice to keep all the information about the signals you want to score on, the ways each of those actions will be tied to each product, and the score increase/decrease associated with each of those elements.

We recommend creating a scoring matrix that will present for each product the behavioral and demographic signals you will monitor, how they will be tracked, whether they will create an Interesting Moment, the product score value increase/decrease they will trigger, as well as the token name they will be associated with.

Create this matrix before you build the program in Marketo; it will make it easier for your team to identify any discrepancies than it will be in the platform, and it can act as a blueprint for building out the program in your instance.

Conclusion

 

With various product-scoring models available, it’s essential to weigh their pros and cons to determine the most suitable one for your needs. Centralizing and organizing the model while keeping scalability in mind will ensure its effectiveness in the long run.

Invest time in creating a robust methodology that enables your team to accurately associate interest signals with the relevant products. Remember, testing and continuous adjustments are typical and expected. Just like lead scoring, implementing a product-scoring program is an iterative process that requires patience and adaptation. The program should evolve alongside your audience and the ever-changing market dynamics.

By adopting these principles, you will unlock the power of product scoring, enabling you to deliver targeted marketing efforts, enhance sales prioritization, and drive revenue growth.