Expert matching algorithms are essential in many fields, and L&D in particular. With an expert matching method, technology platforms can accurately pair a “client” with a “provider” so that the needs of both parties are met to an optimal degree. Such algorithms are constantly evolving, thanks to improvements in artificial intelligence and other technologies.

What Is an Expert Matching Algorithm?

An expert matching algorithm allows someone searching for some kind of knowledge authority to find exactly the right person. 

Such algorithms are a particular application of general matching algorithms. We use these technologies constantly in areas such as:

• Question-answer apps
• User-item shopping platforms
• Entity-relation matching (the concept behind graphs)

Specifically for expert matching, current applications include:

• Author identification
• Collective intelligence
• Learning and development 

For many of these applications, the matching process has traditionally been semi-automated. For example, when you search for a product on Amazon, you tend to receive many pages of suggestions. This is because there are so many dimensions to a typical product search that it takes a person, who might only have a vague idea of what they want, to make a final decision. This is known as “Human in the Loop” data integration. 

Enabling Advances

Recent technology innovations are moving us more towards a solution that reduces human input to a greater extent. These advances include:

• Artificial intelligence
• The Semantic Web
• Process management
• Big Data mining

People Are Still Important

Obviously, the ultimate goal of algorithm development in the field of expert matching does not lead to removing people entirely. At one end of the process, there will always be the “client” who needs to input their search parameters. In the case of an expert search, this means a list of the qualities that they are looking for in the expert. Alternately, when it comes to learning and development, this can include the areas in which the employee requires improvement, for example:

• Professional focus area, for example, leadership
• Skill set, for example, leadership style
• Related skill areas, for example, integrity, risk-taking, and situational leadership

At the other end of the process, there is the “provider” – a coach, mentor, or technical trainer who needs to input what their professional abilities are so that the algorithm can identify them. 

The difference that an advanced algorithm makes is that it boils down the search results to only a few highly relevant matches. Using the Amazon example as an analogy, you would only have to view one page of search results before making a decision when an advanced algorithm is in use.

The Process of Expert Matching in L&D

At present, using technology to find learning and development experts still demands human input at certain stages. However, technologies such as workforce analytics, when integrated with expert matching algorithms, have the potential to reduce even this level of effort. 

Identify Skill Gaps

The first step in expert matching is to analyze the kinds of skills that your employees should develop. You can use various processes here, including a skills gap analysis, career development initiative, and manager recommendations.

The result of your analysis will be a list of skills and skill elements. For example, let’s say that your organization needs more leaders. Working with an employee who is interested in a leadership position, you’ll identify the leadership skill elements where they require development, such as communication, change management, or strategic thinking. 

Apply Taxonomy

Taxonomies are essentially methods of categorization. In terms of L&D platforms, you’ll be using two kinds of taxonomy (through the interface) to select the areas that require development and the experts to teach them. 

The first is granular skill type. As mentioned, there are varying levels of skills and skill elements. It’s important for a platform to allow multiple taxonomy levels so that you can choose specialized training areas. Above, we’ve used skill taxonomy of three levels (professional focus area, skill set, and related skill areas) to illustrate this fact. Without such an ability to refine your search, you’ll have to choose from a generalized skill area that might not be relevant to the employee in the training program. 

The second is the parameters for the experts you want. Typical taxonomies used to categorize experts include their industry, hierarchical specialty (e.g. working with executives or new employees), function (e.g. sales, strategy, operations), and mode of learning (coach, mentor, or technical trainer). 

The result of your search should be a selection of a few experts who specialize in the kind of training that you specified at the outset. 

A Note on Self-Reporting

One potential weakness in the expert matching system is that the expert has the ability to represent themselves as having experience in certain areas but without providing proof. This risk is enhanced by the fact that there is no legal requirement for people who describe themselves as a mentor, coach, or trainer to have certification.

In light of this, HR teams should use L&D platforms that:

• Only list coaches with ICF certification (for example)
• Have an evaluation mechanism that provides feedback and grading regarding an expert’s abilities
• Readily display both certifications and evaluation results

Growthspace’s Expert Matching Algorithm

The Growthspace platform has many features, with its expert matching algorithm lying at the heart of its success. With a 95% accuracy rate, the algorithm connects L&D departments with workplace training experts – coaches, mentors, and technical trainers. These experts form the Growthspace database, cover a wide range of professional subjects, and come from around the world.  

Together, these innovations promise Growthspace clients a maximum chance of finding precisely the highly graded expert they need for delivering very granular L&D courses.