Guide to Understanding and Using Affinity Scores for New Users

Welcome to the exciting world of affinity scores! This is a metric for assessing the similarity between your brand and current and potential customers. This type of insight can help with identifying customers and prospects that align with your brand and those that don’t. In this guide, we’ll help you understand what affinity scores are and how to use them to optimize your outreach and engagement.

A Real World Example

How is an affinity score useful? Let’s look at an example. In a scenario where you’re looking to launch a new product, you want to be strategic about where you spend your time and energy in order to begin recognizing an ROI as soon as possible. One way to do this is to target prospects who already have a high affinity with your branding and messaging, versus those that would require nurturing, or who simply aren’t a good fit for your new product. By using a description of your new product on the Wrench platform you can rank your customers and prospects by how closely they align with your new product. This makes it possible to automatically create a segment of potential customers who are more likely to engage and buy when your product launches, rather than take a “spray and pray” approach.

Pinpoint Your True Top Competitors – and Potential Partners

You can also use affinity scores to assess your brand’s affinity with competitors in your industry. If a competitor scores high, they are likely offering a product and messaging similar to yours, and you’ll want to keep a close watch on their customer communications and product launches.  If they have a lower score, they likely have a different value proposition and could be considered for a strategic partnership if they offer a product or service that fills a gap you have.

What Does an Affinity Score Measure?

It’s a numerical indicator of the degree to which current and potential customers share similarities with your brand. Similarities might be expressed interests or self-descriptions that align with content mentioned in brand documents As mentioned above, you can also generate affinity scores for competitors to evaluate their relationship to you.

Affinity scoring is not word searching. The AI part of affinity scores is understanding the probability of any word or group of words being in a similar space. It is looking at the way in which words share a common space; or how frequently a word or phrase is found in some proximity to other words or phrases when a large corpora of text is analyzed. The idea is that close proximity means there is some type of shared meaning. It is that measure of a shared meaning that lends itself to the psychological mirroring that is at the root of affinity’s success.

How Does an Affinity Score Differ from a Lead Score?

A lead score is a numerical value assigned to a prospective customer (a lead) based on a range of criteria. It's used to determine a lead's level of interest in a company's product or service, and their potential to convert into a paying customer. A lead score is relying on all available statistical predictors to generate a model that successfully calculates the likelihood of a conversion. The proof of its success is the number of times the model successfully predicts a conversion without knowing the conversion beforehand (i.e., blind prediction using a holdout sample that was not used in the modeling). A lead score model must have a known desired outcome with which to build a predictive model. Affinity scores may be predictors in this model, but are not a prerequisite.

Lead scoring models vary by organization, but they generally consider factors such as:

• Demographics: Information like job title, industry, company size, location, etc.

• Behavior: Online activities like website visits, clicked emails, downloaded content, webinar attendance, etc.

• Engagement: Frequency of interactions with the brand, responsiveness to outreach, etc.

• BANT (Budget, Authority, Need, Timing): This traditional sales qualification method can be part of the scoring system, too.

The lead score helps sales and marketing teams prioritize their leads, respond appropriately, and increase efficiency. For instance, high-scoring leads could be fast-tracked to sales as they're deemed most likely to convert while lower-scoring leads could be nurtured further by marketing teams until they're ready for a sales approach.

An affinity score is not predicting an outcome or behavior. Rather, it is measuring similarity between two pieces of information. Using the psychology of mirroring, higher degrees of similarity indicate the extent to which there is shared meaning, and shared meaning indicates a better chance that there will be a better understanding or sympathetic disposition between two objects with a high affinity (e.g., a prospect and brand).

One of the clear advantages of the affinity score over the lead score is that affinities do not require a pre-existing outcome variable in training data. Relying on unstructured data, the affinity score can readily use product reviews, emails, social media posts and profiles, or service logs to understand the relationship of the authors of these documents with brand and product descriptions that often require significant structuring to work in a lead score model. Affinity scores can be a proxy for a lead score (i.e., where high-affinity scores suggest a greater openness to a brand). These scores can also be used as a foundation for segmenting market prospects into useful segments along an adoption curve, thereby enabling a differentiated approach to segments like early adopters or late adopters. Combined with firmographics and organizational culture metrics, affinity scores can be used to select prospects that mirror ideal customers. Affinity scores can also be used to evaluate brand differentiation and identify the element of a brand that should be emphasized in marketing campaigns.

How to Interpret Affinity Scores

Here’s where we’re going to get a little technical, and if you are so inclined, you can skip to the TLDR at the end of this answer.

The interpretation of affinity scores depends on the mean (average) score and the standard deviation (variability). Here are three sample scenarios.

• Example 1: Average score = 60, standard deviation = 5

Justine Smith has created four affinity scores looking at different aspects of her brand. She has applied these scores to a new prospect list. The average of the four scores indicates that there is a high overall affinity for his brand for this prospect list. The standard deviation tells us that the average is not being inflated by a couple of outlier high scores, but indicates that the average is accurately depicting the overall tendency of the prospect list. Bottom line: This is a great list.

• Example 2: Average score = 25, standard deviation = 20

Justine has applied her four affinity scores to another prospect list with very different results. The low average score suggests that this list shares little affinity with the brand. However, the high standard deviation indicates that the average may not accurately represent the prospects on this list. Further investigation may reveal that there is a small subset of high-scoring prospects that could be targeted. Alternatively, the low average and high standard deviation may indicate large differences in the affinity between different aspects of the brand. While Justine’s initial assessment may conclude that this prospect list will underperform, the large standard deviation indicates the need for deeper evaluation before rejecting the list.

• Example 3: Average score = 20, standard deviation = 3

Justine has applied her four affinity scores and is right to conclude that this is a consistently poorly performing prospect list. The low average and small standard deviation indicate that the vast majority of prospects have a low affinity with her brand. This is not a list she should pursue.

Another Real World Example

Let’s put this in the context of a prospective customer Miles Morales. His affinity score for a new product is 65, placing him in the 80th percentile of contacts that show a high degree of similarity with the new product. In other words, this indicates that Miles shares more similarities with your product than 80% of other prospects. In other words, Miles is in the top 20% of prospects having a high affinity with the product. The higher the percentile, the more highly ranked the affinity customers and prospects share with your new product.

As seen in the standard deviation examples above, keep in mind that the total range of affinity scores can influence this interpretation and percentile reflects ranking within the group and may not reflect a corresponding rank if compared to a global benchmark.

TLDR: When it comes to your brand or a new product (or whatever it is you are seeking an affinity score for) the higher the affinity score of a prospect or customer, the higher the likelihood they are to engage with your outreach and purchase, compared to those with affinity scores that are lower.

Actions to Take Based on Affinity Scores

This metric can provide data-driven guidance for making strategic decisions.

Affinity scores are a powerful tool, but they work best when combined with an overall strategic objective; they can provide data-driven guidance on who to target so you can focus on an audience most likely to engage and convert. Happy strategizing!