Before launching into how to interpret a match score, allow me to address how we define one. At a technical level, match scores measure the affinity or likely similarity in meaning between two objects. In the world of AI, objects are usually some form of text that represents how an entity communicates who or what it is.
Here’s an example of where you would use one: Company A wants to launch a new product, so it will create a compelling message for a promotional campaign that describes its unique offering – and to maximize resources, budget, and efficiency – will only target customers with a match (or lead) score over a certain number. The higher the number, the higher the indication that a customer is likely to engage at some point during a promotional campaign. Those who engage are also likelier to convert, or purchase.
How can you match likely customers to the new product? The first step is to identify the two objects – we also refer to them as entities – to measure the affinity between the two. In the example above, the first object would be a description of the product, while the second would be a description of the customer, which could be a social media profile, like a LinkedIn profile. Note that if you have a small customer data set, you could do the matching manually. It might take a lot of time, but it’s possible. Let’s say you have a customer data set of one million customers; there’s no other way to do this than through automation (that’s where we come in, as we specialize in using AI for very large data sets).
I will spare you the technical details, but once we have two entities we can see how closely the language surrounding them shares similarities or affinities.
The power of the match score lies in its inferential power, or its ability to predict the likelihood of a strong match or a weak match.
For a sales or marketing team, high match scores between customers and a brand suggest that the brand’s message or description will have a positive resonance with high-scoring customers and a less positive resonance with low-scoring contacts. Notice that I did not use the term “negative resonance”; customers may have lower scores because they are not as familiar with a brand, but with a nurturing campaign they could eventually exhibit a higher match score because they are signaling more familiarity with the brand.
Conversely, high-scoring contacts could indicate that their public personas are more informed about the brand category and would therefore not require the same degree of education as their low-scoring counterparts.
The question most clients have is: “What constitutes a high or low score?” Generally, scores can range from 0 to 100, with a high score being anything greater than 60. Individuals scoring over 60 usually indicate someone who is an innovator or someone who is publicly expressing a higher degree of familiarity with a brand or a product.
Individuals with scores less than 35 can be considered uninterested or unfamiliar with the content of the comparison entity.
The most important thing to note is that scores need to be viewed in the context, which includes the population sample (are you matching warm leads from your CRM or a cold list?) and industry (is your product super technical, or easy to understand?), and possibly other variables. Match scores can provide statistically significant guidance on who to target based on the goal you are seeking to accomplish. Marketing and sales efforts that incorporate match scores are much more likely to be effective because they take into account more informed targeting in promotional and outreach efforts, rather than the typical casting of a wide net, where everyone is considered to be part of the same playing field.
