Our method for festival and event studies
The direct shock in a festival study is net new spending: money spent in the host region that would not have been spent there without the event. We measure it with on-site intercept surveys. Field staff sample across days, gates and times of day so the sample reflects the whole crowd, and record home ZIP code, whether the festival was the main reason for the trip, party size, nights in the area and spending by category.
Attendees who live inside the study region are excluded, since their spending would most likely have gone to other local businesses. Visitors who were already in town for another reason are excluded or counted only for extra nights they stayed because of the event. We divide spending by party size and multiply by unique out-of-region attendance, estimated from ticket scans and a days-attended question so multi-day attendees are counted once. Ticket purchases are left out of attendee spending because they fund the organizer's budget. Organizer spending is modeled separately and includes only payments to local vendors, crews and venues; fees paid to touring artists mostly leave the region.
Typical inputs we ask for
- Ticket sales and gate scans by day and ticket type
- Ticket buyer ZIP codes, if your ticketing platform provides them
- Site access for survey staff and a map of entrances and stages
- The organizer's production budget, split by local and non-local vendors
- Hotel occupancy or lodging tax data for the event dates, if the city or visitor bureau shares it
- Any results from prior-year attendee surveys
Questions clients ask
Why is your number smaller than total spending at the festival?
Total spending includes residents who would have spent the same money at local restaurants and venues that weekend, so it shifts activity within the region rather than adding to it. We report net new activity from out-of-region visitors as the headline. For Zootown Music Festival that approach produced $9.1 million in net new output for the Missoula region, a figure sponsors and city officials could defend.
How many survey responses do you need?
It depends on attendance, the share of visitors from outside the region, and how finely you want results broken out, for example by origin state or ticket type. We set a target in the proposal that gives usable margins of error for spending by origin group. Spreading interviews across days and times matters as much as the total count, because early arrivals and headliner crowds differ.
Can the survey cover satisfaction and sponsor questions too?
Yes. Our Zootown survey also documented attendee satisfaction and likelihood of returning the following year, and multi-year events can track Net Promoter Score over time. We keep the spending questions first and the full instrument short, since a long survey lowers completion rates at a gate or between sets. Sponsor recall questions can be added if a partner needs them.
Related reading
Sources
- Bess and Ambargis, Input-Output Models for Impact Analysis: Suggestions for Practitioners Using RIMS II Multipliers, BEA working paper WP2012-3 (only net changes in final demand should be applied to multipliers; displaced local sales are netted out) Source
