Group composer
Build groups seat by seat from your audiences, so that every group follows the exact composition you set, and choose how often people can be rematched.
“Group composer” builds groups seat by seat: you decide who sits in each seat of a group, and RandomCoffee only creates groups that follow this composition exactly. For example, in groups of 2, seat 1 for a mentor and seat 2 for a mentee.
The algorithm stops creating groups as soon as it can no longer fill every seat. Not everyone is necessarily matched: it depends on the size of each seat and on the rematching option you choose (see Groups distribution and rematching).
When to use it
Use “Group composer” when every group must have a strict composition. For example, to:
- match new hires with members of the management team;
- connect members who are looking for a new role with members who are hiring;
- pair each new hire with a buddy (see Onboarding buddy program).
For a dedicated mentoring program, with mentor and mentee roles, milestones and a mentoring page, use a mentoring program instead: see Create a mentoring program.
Before you start
- Create one audience per seat, for example a “Mentors” audience and a “Mentees” audience. Seats are filled with audiences, static or dynamic. See Create an audience.
- Check how many people your program includes. On the Free plan, a program only includes the first 20 people of its audience. The Pro plan removes this limit.
Set up Group composer
-
Open the matching settings of your program.
Open your program, go to the “Settings” tab, then to the “Matchmaking” section.
-
Choose the group size.
In “Group size”, choose the number of people per group (from 2 to 10). Each group has as many seats as people.
-
Choose the algorithm.
In “Matching algorithm”, choose “Group composer”. The “Groups distribution” settings appear below it.
-
Choose a rematching option.
Choose one of the three options described in Groups distribution and rematching.
-
Fill each seat.
If the seats ask for category values, click “Switch to audiences” just above the seats first. Then, in each “Seat” card (“Seat 1”, “Seat 2”…), select one or more audiences. The number below each card is the number of active members in these audiences.
-
Add matching rules (optional).
Under “Matching rules”, click “Add a rule” to refine who is grouped with whom. See Matching rules and priority.
-
Click “Save”.
How groups are built
For each session, RandomCoffee takes the members who take part in the session (with “Ask for participation”, only those who accepted) and places them in the seats whose audiences they belong to. Then it builds groups one by one, taking one person from each seat.
- A member who is in none of the seat audiences is not matched, even if they take part in the session.
- A member who is in several seat audiences can fill any of these seats, but never two seats of the same group.
- The number below each seat counts all the active members of its audiences. In a session, the seat only contains the members who take part in it, so it can be smaller.
Groups distribution and rematching
Under “Groups distribution”, choose how often the same person can be in several groups during one session:
| Option | Number of groups created | Who is matched |
|---|---|---|
| “Allow unlimited rematching - some people will be rematched several times so that everyone gets matched.” | As many as people in the largest seat | Everyone. The people of the smaller seats are rematched in turn, so that each person of the largest seat gets a group. |
| “Allow rematching - up to a maximum number of [N] rematchs per session” | Until the smaller seats run out of people | Each person can be in [N] groups at most per session (2 or more). The people of the larger seats who cannot get a group are not matched. |
| “Do not allow rematching” | As many as people in the smallest seat | Each person is in one group at most. The people of the larger seats who exceed this number are not matched. |
It is always the people of the smaller seats who are rematched: each person of the largest seat is in one group at most.
Each group is a separate match. With rematching, match messages are sent to each group separately: a mentor matched with 3 mentees receives 3 different messages, each one introducing 1 mentee.
Example
You create groups of 2, with 1 mentor and 1 mentee in each group. “Seat 1” contains a mentor audience of 41 members, and “Seat 2” a mentee audience of 113 members. All of them take part in the session.
| Option | Groups created | Mentors | Mentees |
|---|---|---|---|
| “Allow unlimited rematching…” | 113 | 31 mentors have 3 mentees, 10 mentors have 2 mentees | All 113 mentees have a mentor |
| “Allow rematching…” with a maximum of 2 | 82 | Each of the 41 mentors has 2 mentees | 82 mentees have a mentor, 31 are not matched |
| “Do not allow rematching” | 41 | Each mentor has 1 mentee | 41 mentees have a mentor, 72 are not matched |
With unlimited rematching, mentors are rematched in turn, so the mentees are spread evenly: 113 mentees for 41 mentors makes 2 or 3 mentees per mentor.
These numbers assume the “Optimize people participation” priority. With “Optimize matching quality”, some of these groups can be removed (see below).
Matching rules and priority
The seats decide how many groups are created and their composition. Matching rules decide who is grouped with whom within this composition: for example, a “Category” rule on your department category with “Is different” pairs each mentee with a mentor from another department.
The “Matching priority” then decides what happens to the groups that do not fully respect your rules:
- “Optimize people participation” (default): RandomCoffee keeps all the groups, even if some rules are not perfectly met.
- “Optimize matching quality”: RandomCoffee removes every group that breaks a strict rule. Fewer people are matched, but every group respects your rules.
If the “Matching priority” setting is not shown below the algorithm, your program uses “Optimize people participation”.
With “Optimize matching quality”, the strict rules are “Category” rules with “Is the same” or “Is different”, “Category values”, “Audience” and “Specific members” rules, and the two default rules. For example, two people who have already been matched together are not grouped again, unless one of them has already met everyone else in the session. Rules that compare values (“Is the closest”, “Is the furthest”, “Has the most in common”, “Has the least in common”, “Category overlap”) never remove a group. See Matching rules.
Seats with category values
Seats can also be filled with category values instead of audiences. In that case, the seat cards list your categories and their values, and a “Switch to audiences” button appears just above the seats. Click it to fill the seats with audiences, as described on this page: an audience can combine several filters, and a dynamic audience updates by itself.