The event networking playbook
How to design conference networking that works for people who hate networking, with formats, timings, room layouts and what to measure afterwards.
The short answer
Networking fails when it requires initiative. Design the first interaction so nobody has to start it, keep groups under six, give people a reason to talk that is not themselves, and make exchanging details a single action rather than a task for later.
- The barrier is starting, not talking. Remove the start.
- Groups of four to six work. Groups of ten do not.
- Standing rooms mix; seated rooms do not.
- Most connections are lost between meeting and following up, not at the meeting.
- Written by
- The event67 team
- Last updated
- Length
- 1,384 words, about 12 minutes
The event networking playbook
Ask attendees what they came for and most say the people. Ask them what worked and most describe a session. The gap is a design problem, and it is solvable, because the barrier is almost never conversation. It is starting one.
Everything below follows from that. If the format requires an attendee to walk up to a stranger and introduce themselves, it will work for the third of the room that finds that easy and fail for everyone else.
The four conditions
Conversations happen reliably when four things are true. Design for all four and the rest is detail.
A reason to talk that is not “networking”. A topic, a shared problem, a task. “Table 4: hiring” produces conversation; “networking lunch” produces phones.
A group small enough to join. Four to six. At eight, someone is silent. At ten, it is an audience with a chair.
A defined end. People commit more when they know it stops. “Twenty minutes, then we move” removes the fear of being stuck.
A next step at the moment of meeting. The connection has to be captured while both people are standing there, or it is a promise nobody keeps.
Formats that work, by size
Under 100 people
Assigned tables at the first meal. Name cards. Mix by company and role. This one intervention is worth more than everything else you could do at this size, because at under a hundred people everybody meeting three others means the whole room is connected.
A go-round. Thirty seconds each, one question everybody answers. Works up to about forty and does not scale past it.
Hosted topic tables. A sign and a person. The person’s job is asking the second question.
100 to 500
Topic tables at lunch on day one, with named hosts. The workhorse format. Print eight signs, brief eight hosts, done. Publish the topics in the app the morning of so people can choose.
First-timer orientation, twenty minutes before the opening. Hosted by someone approachable rather than someone senior. Half of your first-timers will attend and most of them will meet each other.
Speed networking, once, optional, facilitated. Three minutes per pair, a bell, twelve rounds. It needs a real facilitator and a room with movable chairs. It is not for everyone, which is fine because it is optional.
Track meetups at the end of day one. People who sat in the same sessions have a shared reference. Twenty minutes, in the session room, no programme.
Over 500
Structured small groups within a large room. The room size does not matter if the tables are six.
A hosted-buyer or matchmaking programme, if your event’s commercial model depends on exhibitors meeting buyers. This is a real product category and a real cost. An attendee directory is not a substitute for it, and neither is event67 — it has no meeting scheduler and no matchmaking. Say this to your exhibitors before they assume otherwise.
Interest-based meetups distributed across the venue. Publish locations in the app the morning of, not in the printed programme, so you can move them.
Room layout is the intervention
Physical arrangement decides more than the programme does.
- Standing mixes; seated does not. A seated reception is a set of fixed conversations. A standing one recombines every few minutes.
- Round tables beat long tables. At a long table you can speak to three people. At a round table of six you can speak to all five.
- Put food and drink where you want people. Circulation follows the coffee. If the sponsor area is empty, move the urn.
- High tables for two to four. They create the small-group unit without assigning anyone.
- Bottlenecks are features. A single door into the break area produces more collisions than three doors. Do not over-optimise flow.
- Leave a quiet room, unbranded and signposted. Some of your attendees need it in order to stay all day, and they will not ask.
Timing
| Slot | Minimum that works | Why |
|---|---|---|
| Morning break | 20 minutes | 15 is a coffee queue |
| Lunch | 60 minutes | 45 means eating, not meeting |
| Afternoon break | 25 minutes | Energy is lowest; the conversation carries it |
| Reception | 90 minutes | 60 ends as it warms up |
| Unstructured block | 60 minutes, once | The single most valuable hour on the programme |
Fewer, longer breaks beat more, shorter ones every time. Two 25-minute breaks are better than three 15-minute ones, and cost you the same programme time.
Making the exchange a single action
This is where most networking programmes leak. Two people have a good conversation, agree to follow up, and neither does — because following up requires remembering a name, finding it in a directory, and writing a message, and each step loses people.
The fix is to capture the connection at the moment of meeting, in one action, while both people are standing there. Scanning a code on the other person’s phone is the version of this that works, because it takes three seconds and it happens before the memory fades.
In event67 that is what QR connections are: an attendee scans another attendee in person and the two are connected, with browse-and-request available when they are not in the same room. It is deliberately attendee-to-attendee. It is not exhibitor lead retrieval, it does not push anything into a CRM, and it should never be sold to a sponsor as if it were.
The attendee directory sits behind that as a lookup tool: searchable by name or company, with each attendee controlling what appears on their own profile and able to stay out of it entirely. That last part is not a compliance footnote — directory participation being visibly optional is what makes people comfortable being in it.
Designing for the person who came alone
Solo attendees are the group most likely to leave early and least likely to tell you why. Four things help, in order of effect:
- Assign the first interaction. A table, a group, a host. They do not have to choose to join.
- Give them a role. Table host, session introducer, question wrangler. A job removes the need for small talk entirely.
- Name the format honestly. “Hosted table for people who came alone” gets attendance. “Networking” does not.
- Introduce them by name. If you know two people should meet, walk over and say so. Nothing else you do all day has a higher hit rate.
What to measure
Three numbers, from the app and the survey, held across years:
| Number | Where it comes from | What it tells you |
|---|---|---|
| Connections made per attendee | The app’s connection records | Whether the formats produced meetings |
| Share who made at least one connection | Same, as a percentage | Whether it worked for the room or for the extroverts |
| Share who expect to follow up | Survey question | Whether the connections were worth anything |
The second number is the important one. A high average with a low share means a few people met everybody and most people met nobody, which is the failure mode that looks like success on a dashboard.
The follow-up window
Most value is lost after the event, not during it. Two cheap interventions:
One email the day after. Not a highlights reel. A short prompt: here are the people you connected with, here is how to reach them, this is the week to do it.
A named reason to reconnect. The recording of the session they both attended, the report you promised, the answer to the question that got the most upvotes. A follow-up email with something in it gets opened; one that says “great to meet you” does not.
What not to do
Icebreakers with props. They select for the people who were already comfortable and embarrass everyone else.
Networking bingo. It turns people into a checklist and produces thirty seconds of contact thirty times.
An unstructured drinks reception as your only networking. It works for people who already know somebody, which is the group that needed the least help.
A directory presented as the networking programme. It is a lookup tool. It supports the programme; it is not one.
Filling the unstructured hour. The first thing to get cut when the programme gets crowded is the thing most attendees actually came for. Protect it.
What it costs
- Free up to 50
- activated attendees, on one live event
- $5
- per activated attendee beyond 50
- $4,750
- the most one event can ever cost, whatever happens past 1,000
Activated means the person claimed their invite and opened the app. Not when you import them, not when the invite is delivered, and not if they register and stay home.
Questions people actually ask
How do I make conference networking less awkward?
Remove the need to initiate. Assigned tables, hosted topic groups and facilitated openers all work because nobody has to walk up to a stranger. Once a first conversation has happened, most people are fine for the rest of the event.
What is the best networking format for a conference?
Hosted topic tables at lunch on day one. They are nearly free, they scale from 50 people to 2,000, and they give everyone a reason to talk that is not their job title. Speed networking works but needs a facilitator and a room; unstructured drinks works only for people who already know somebody.
How long should a networking session be?
Long enough for a second conversation. Twenty-five minutes minimum for a break, ninety minutes for a reception. Fifteen-minute breaks are a coffee queue, and a sixty-minute reception ends just as it warms up.
How do I help attendees follow up afterwards?
Make the exchange happen at the moment of meeting, not later. A scanned code beats a promise to find someone in the directory. Then send one email the day after with a prompt to reach out, while the memory is still specific.
Does an attendee directory actually help networking?
It helps people who already know who they are looking for, which is a minority. It rarely creates a first conversation on its own. Treat it as a lookup tool that supports networking rather than as the networking programme.
Keep reading
Make the exchange one action
Attendees scan each other in person and are connected, with a directory behind it that each attendee controls and can opt out of entirely.