Back to guides
Live event runbookUpdated July 22, 20263 min read

How to run repeat-aware speed networking rounds

The schedule is only half the event. Good speed networking also needs the right conversation size, visible transitions, a late-arrival plan, and assignments that a producer can execute while the host keeps speaking.

A cheerful event organizer with a stopwatch keeps speed networking conversations rotating between small tables.

Guide visual

New conversations each round

Short rounds optimized for fresh contacts.

30 attendeesgroups of 35 rounds

Choose pairs or triads deliberately

FormatBest whenOperational trade-off
PairsThe goal is focused one-to-one introductions.Odd attendance creates a single person, bye, or deliberate triad.
TriadsYou want easier recovery from absences and less pressure on shy participants.Each round consumes two new contacts per person, so novelty runs out faster.
Groups of fourConversation needs more perspectives than individual introductions.Turn-taking becomes harder and every round creates three contacts per person.

Budget the whole round, not just the conversation

  • Allow time to announce the next assignment and let people move or join rooms.
  • Give a visible halfway or one-minute warning rather than ending mid-sentence.
  • Use the same prompt structure each round so participants do not need another long explanation.
  • Leave a buffer for the producer to place late arrivals before the next round starts.

Prepare and run the rotation

  1. Build one clean attendee list

    Use stable names or IDs. Keep walk-ins on a separate list until the producer adds them deliberately.

  2. Choose the conversation shape

    Set pairs, triads, or another size based on the event goal and expected attendance volatility.

  3. Generate every planned round

    Keep repeat minimization enabled and solve the complete event at once. Save one assignment sheet per round.

  4. Brief a producer

    One person should operate room moves or call tables while the host explains prompts and keeps time.

  5. Handle changes between rounds

    Place a single late arrival into the smallest viable group. If attendance changes materially, regenerate the remaining rounds rather than pretending the old repeat score still applies.

  6. Close with a deliberate final round

    Tell participants when the last rotation starts and provide a way to exchange contact details with consent.

Separate networking reach from conversation quality

  • Your groups answers the producer’s immediate question: who goes where in each round, and did the schedule preserve the intended pair-or-triad format?
  • The Pair Meeting Histogram describes reach. The one-session bar counts pairs that met exactly once, “Repeated” counts pairs that met more than once, and the zero-session bar counts pairs that never met.
  • The Pair Meeting Matrix identifies the conversations behind a repeat. Selecting a cell reveals the pair and the sessions they shared.
  • Unique Contacts and Avg Contacts / Person are useful for comparing candidate schedules with the same roster and round count. They measure reach, not whether the conversations were useful.
  • Once attendance changes, the old histogram and contact totals describe the old event. The remaining rounds need a fresh result before those diagnostics are meaningful again.

Worked example: thirty people in triads

Thirty attendees form ten triads for five rounds. Each person has ten contact slots out of twenty-nine possible contacts, so the simple capacity bound permits a repeat-free schedule.

Setup facts

  • 30 attendees
  • 10 triads per round
  • 5 rounds
  • no fixed hosts or relationship rules

Review the result

  • Check that every attendee receives five assignments.
  • Inspect whether any repeated contact appears despite the available capacity.
  • Recalculate after no-shows; the original arithmetic and schedule no longer describe the live room.

Try this setup in GroupMixer

This tool is preloaded with the example from this guide. You can edit the participants, constraints, sessions, and balance settings before generating groups.

Enter one person per line in the first column. Optionally, add attribute columns such as gender, role, or skill-level on the right. Those attributes can then be used to balance groups and set additional rules.
Name
Use 1 session for a single round of groups. Increase sessions when you want multiple rounds with new group assignments for the same people.
People
30
Groups
10
Approx size
3
Force specific people into the same group in every session. Write one "clique" per line, with names separated by commas. Example: "Alex, Sam" keeps Alex and Sam together.
Prevent selected people from being placed in the same group. Write one set per line; everyone on a line is kept mutually apart. Example: "Alex, Sam, Jordan" keeps all three separated.
Pin specific people to a specific group across all sessions. Enter one name and one group number per row. This is useful for leaders, presenters, or anyone who must stay in a known group.
Name
Group
Set target counts for each attribute value inside each group. Keep auto distribute enabled for fair automatic targets, or edit the counts manually when you need exact control.
Add an attribute in the participants list to use this section.
Need even more control?
  • Partial attendanceSet which participants attend which sessions instead of assuming everyone is present every round.
  • Custom group capacitiesGive each group its own capacity and override those capacities for specific sessions when room sizes or staffing change.
  • Session-specific constraintsApply scoped rules only where they matter, then use global objectives such as Avoid Repeats for whole-schedule quality.
  • Weighted preferencesAdd preferences that can be violated when needed, then tune their weights relative to other goals.
  • Encounter targetsTarget how often selected people should all meet across the schedule, including exact, minimum, or maximum encounter counts.
  • Group visit targetsControl how often selected people visit selected groups across sessions for station rotations, task exposure, or visit caps.
  • Advanced constraint tuningFine-tune repeat limits, attribute-balance modes, penalties, and other constraint details beyond the landing-page controls.
  • Solver settingsAdjust runtime limits, deterministic seeds, solver family, and other optimization settings.
  • Result analysisInspect score breakdowns, constraint compliance, penalties, and saved results in more detail.
Quick Setup and the scenario editor use the same saved scenario. Settings hidden here remain intact.

Related guides