Suggested Contacts is Beamery’s AI-powered matching engine that allows you to easily find and review relevant candidates for your Pools, minimizing the manual element of sourcing and giving you more time to make meaningful connections with the best candidates.
Table of Contents
- How does Suggested Contacts work?
- How does the Match Score work?
- What profile information will Suggested Contacts use to determine the ideal persona and match score?
- How can I improve the suggestions for my Pool?
How does Suggested Contacts work?
We hear you! AI can be mysterious and sometimes hard to trust, but it doesn’t have to be this way. Here’s what to know about Suggested Contacts so you can have confidence in the AI behind the feature.
First, Suggested Contacts will analyze the contacts already in a Pool to find commonalities between them. Suggested Contacts looks for commonalities in the contacts’ locations, job titles, languages, skills, and keywords found in their profile attachments.
Then, Suggested Contacts will analyze the list of contacts you dismissed from a Pool. These are contacts that were previously suggested for a specific pool, but which you or another team member dismissed as unsuitable.
From this analysis, Suggested Contacts will determine a hypothetical ideal persona for your pool. This ideal persona is not visible within the platform but will be used behind the scenes to power suggestions.
Finally, Suggested Contacts will assess the contacts in your database against this hypothetical ideal persona, and each contact will be given a ‘Match Score’ based on how similar they are to the ideal persona. The top 20 matching contacts are then presented to you in the list of suggestions, where you can then review the profiles and add them to the Pool, or dismiss those you find to be unsuitable.
The more you add and dismiss candidates, the more Suggested Contacts will be able to refine the ideal persona for your Pool, and the more refined your suggestions will become.
How does the Match Score work?
When Suggested Contacts runs, contacts in your system will be compared to the ideal persona in all of the above categories and will be assigned a match score for each of these areas. The Match Score is calculated and used only in the back end of Suggested Contacts, and you will not see a numeric ‘score’ associated with a candidate profile. The combined match score for all of the areas will determine how closely a contact matches the ideal persona of the Pool, and the contact with the overall highest match score will be considered the strongest suggestion.
The most relevant contacts in your database will be shown in the list of suggested contacts, ranked by Match Score. The first contact will be the contact with the highest Match Score, and the subsequent contacts Match Scores will be determined in relation to this first contact’s Match Score.
In the example below, Hein Behrens is the contact with the strongest Match Score in the CRM for this Pool, so this profile is shown at the top of the list, with a full green bar in the Match Score column. In comparison with Hein Behrens, Dwayne Smith does not match as strongly with the ideal persona, so his Match Score is shown as a partially filled yellow bar.
You can further dig into a contact’s Match Score by opening the mini-profile and reviewing the Match Score area at the top right-hand side of the profile.
In this Match Score area of the profile, you can expand an area where there is a match, to see what information from the contact profile matched with the ideal persona.
What profile information will Suggested Contacts use to determine the ideal persona and match score?
Suggested Contacts analyzes candidate location data in two ways, using map coordinates associated with geolocation data, and by matching text strings of relevant address, city, and country data. These methods are used in combination to suggest contacts for your Pool.
Using geolocation data, Suggested Contacts finds the average of the latitude and longitude positions of contacts in a Pool. It will then draw a circle from that point and allocate a higher Match Score to those within the area. Contacts outside the area are not excluded but are given a lower score.
When matching location keywords, Suggested Contacts finds common words and phrases in the location fields of the contacts already in your Pool and uses those as part of the ideal persona.
Significant combinations of terms are recognized and given great emphasis. For example, if many contacts in your Pool have “United Kingdom” in their location, then that term will be added to the ideal persona.
This same type of matching is carried out for addresses, cities, and countries to be sure the candidate suggestions you receive for a Pool are as geographically relevant as possible.
When Suggested Contacts analyzes the job titles of contacts in a Pool, it assesses the frequency of significant words in those job titles. For example, if a Pool of 50 candidates includes 45 profiles with the word ‘designer’ in their job title, ‘designer’ would be ranked highly, and would be added to the job title information of the ideal persona.
Then, when generating suggestions for the Pool, the frequency of a significant word in job titles affects how high the Match Score is for a contact’s job titles. So, in the previous example of ‘designer’ being added to the ideal persona, a profile with five different roles that include the word ‘designer’ would have a higher match score for job title than a profile with only one role containing the word ‘designer’.
Suggested Contacts will also analyze how common a word from a job title is in all contacts in the CRM. If a word is particularly rare, this will be given more significance on the ideal persona, and contacts who have job titles that include this word will have higher match scores for job title.
When Suggested Contacts analyzes the languages on the existing profiles in a Pool, it looks for the languages in common between these contacts, and for the frequency with which these languages appear. For example, if in a Pool of 50 contacts, 45 of these contacts have the language ‘French’ on the profile, ‘French’ will be added to the language information on the ideal persona.
Then, when Suggested Contacts looks for contacts to suggest for this Pool, those who have the language ‘French’ would have a high match score for Languages.
When Suggested Contacts analyses the skills on the existing profiles in a Pool, it looks for significant combinations of words, and for the frequency with which these combinations arise. For example, in a Pool with 50 candidates, if 30 of these candidates have the skill “‘Search Engine Optimization’, this will contribute to the Skills of the ideal persona.
Then, when Suggested Contacts looks for contacts to suggest for this Pool, those contacts who have the skill ‘Search Engine Optimization’ would have a high match score for Skills.
When analyzing attachment keywords to suggest contacts for your Pool, Suggested Contacts uses the analysis already done on job titles and adds this job title information to the ideal persona’s attachment keywords information.
Then, when Suggested Contacts looks for contacts to suggest for this Pool, those who have attachments that contain the attachment keyword information from the ideal persona would have a high match score for attachment keywords.
How can I improve the suggestions for my Pool?
Now that you know which areas of the profile contribute to the ideal persona, test Suggested Contacts by adding 5 to 10 contacts who have information in their profiles that match the type of candidate for which you’re sourcing so Suggested Contacts has good information from which to build the ideal persona.
For example, if you’re trying to fill your pool with Senior Engineers who live in London and speak Ukrainian, add 5-10 profiles that meet these criteria to your Pool to point Suggested Contacts in the right direction.
Then, spend a short while adding and dismissing the contacts that are suggested to you. Remember, every time you refresh this list of contacts, Suggested Contacts will review the additions and dismissals further refine the ideal persona.