MatchMaking Match Criteria
Build a ranked preference list per matchmaking attendee by CSV import, string matching or AI, then bulk book appointments from it.
Match Criteria builds a ranked preference list for every matchmaking attendee and books appointments from it in bulk. Open Match Criteria in the show's MatchMaking sidebar section. The screen is a three step flow: pick how attendees get matched, review and curate each attendee's list, then schedule the pairs for a meeting day.
Step 1: Select Matching Criteria
Three matching engines, each with its own on and off switch. Switch on the ones you want, configure them and select Save. The switches control step 2: a generator button only works while its engine is switched on.
- Preference Matching imports ranked preferences from a CSV. Download template gives the format: a header of Email, Email Preferences, then one row per attendee with their own email first and their preferred attendees' emails after it, in rank order. Import reads the file: rows that cannot be used (an unknown email, a bad address) land in a failure file you can download, and never stop the rest of the file. Every import appears in the history with its counts; Roll back removes the preferences an import created, except rows that were already booked into meetings.
- String Matching compares attendees on one profile list field: Commodities, Sub Commodities, NAICS Codes, Diversity Type, Services or Sub Services, shown with your show's own field names and only when your attendee form collects them. Attendees whose selections overlap score higher, and an attendee's first choices count more than their later ones.
- AI Matching scores attendees over weighted criteria blocks. Each block picks one or more profile fields and a weight from 0 to 100 (50 to start); Add Criteria adds blocks and the bin removes one. The weights are part of the scoring: a block with weight 80 moves the result four times as much as a block with weight 20.
Step 2: View Matches
The left card lists every active, booked matchmaking ticket, ten per page, with search. Pick an attendee to see their ranked preference list on the right:
- Drag a row by its handle to re-rank it; the order saves as you drop.
- Type in Search attendees to add and select + to add someone manually. The screen refuses an add that breaks the rules and says why: attendees cannot prefer themselves, the same person cannot be added twice, the target's ticket type must be in the attendee's Show Matches list from the MatchMaking setup, and the ticket type's Max Interests cap applies. A ticket type without a cap has no limit.
- The x on a row removes that one preference. Clear Preferences empties the attendee's open list after telling you how many rows it will remove; rows already booked into meetings stay.
- Add Preference from AI and Add preference from String Matching run the engines over every attendee as a background run with live progress. Both write through the same rules as a manual add, so running an engine again tops lists up instead of duplicating them. Engine rows carry their similarity score.
Step 3: Schedule Appointments
The preview lists each attendee with their preferred matches as pairs. The Event Date chips pick the meeting day, the filter shows All, Unscheduled or Scheduled pairs, and search matches names and companies. A pair that is already booked shows a Scheduled badge; the bin on an open pair removes that one preference. Export Schedule downloads the pairs as a CSV with each side's values for the configured matching fields.
Tick the pairs to book and select Schedule. The run books each pair through the same rules as the Appointments screen:
- The earliest slot of the selected day where both people are free, neither has blocked the slot and the slot still has room.
- Nobody is double booked: a pair that already has a meeting anywhere in the event is skipped and reported.
- Each person's appointment limit applies in its real mode, Per Day or Per Event.
- A ticket type with the Approve action books its meetings as Pending, decided pair by pair.
- The usual confirmation emails and notices go out, exactly as for a hand booked meeting.
The run shows live progress and finishes with a real result: how many pairs booked and how many failed. Failed pairs (no common open slot, limit reached, already meeting) are listed in a downloadable error file with the reason per pair. Booked pairs flip to Scheduled on the preview and in each attendee's list.
Still need help?
Contact your Map D account team. Include your show name and, if something looks wrong, a screenshot of the page.
Last updated November 12, 2026