Audience Analysis¶

Access

Audience Analysis is available to Admin and Theater users. Theater users see only productions belonging to their theaters.

Navigation: Admin Menu > Analysis > select "Audience" as the Analysis Type

Shortcut from a production

On any production's admin page (Productions > [Production Name]), click the Audience Analysis button next to Edit / Ticket Classes. This jumps straight to the analysis form with that production pre-selected as the target and the production's own theater seeded as the comparison group. One more click on Run Analysis produces the report.

What it answers¶

For a selected production, Audience Analysis answers questions about the people who attended it:

How many of them are new to your programming (no other attendance in a given lookback)?
How many came back from the prior show(s) in a comparison group you choose?
How many are repeat patrons of that comparison group (2+ or 3+ visits)?
How many have attended every comparison-group production whose run overlapped a window?
How do those same numbers look when you widen the lens from "this company" to "the entire facility"?

The output is a single table organized by lookback window (3 months / 6 months / 1 year / 3 years / 5 years / Ever), with rows broken out by metric, and grouped by comparison scope. The "Ever" column has no lookback bound — it considers every prior production that ran up to the anchor date, so it's the closest thing to a true lifetime-new-vs-lifetime-returning view.

Picking the comparison group¶

The Audience type replaces the "Comparison Shows" picker (used by Rate of Sales) with a Comparison Group picker that selects theaters, not individual productions:

Theater autocomplete -- Search by theater name or by any tag applied to a theater. The current production's theater is added automatically as a seed when you switch to Audience.
"All theaters tagged X" group entries -- When the autocomplete finds a matching tag, it offers a one-click bulk add of every theater carrying that tag.

You can pick any combination of theaters; at least one is required. The facility-wide metrics (every theater in the system) are always shown alongside the comparison-group metrics, so there's no separate toggle for "compare against the whole building" -- those numbers appear in the lower section of every audience report.

The Back to Analysis button on the results page preserves your full comparison-theater selection, so you can iterate on the comparison group without re-picking theaters every time.

Anchor date¶

All windows look backward from an anchor date:

Closed productions -- The anchor is the production's closing date. Windows look back from there. Orders for shows that ran AFTER the production closed are excluded.
Still-running productions -- The anchor is today. The same post-anchor exclusion applies: anything dated after today doesn't count.

This means a customer who attended The Ally and then returned four times after it closed shows up as a first-timer for The Ally's analysis -- but a returning customer when you run the analysis on the LATER shows. Each analysis is anchored relative to its own production.

The metric table¶

Header rows¶

Selected production attendees -- The cohort being analyzed. This is the count of distinct Address records linked to paid (non-comp) Processed or Fulfilled TicketOrders on the production. The system already merges duplicate Address records, so each address_id represents one real person. Addresses are included if they have either an email or a street address (line + zip); addresses with neither are dropped because they can't represent a real person, and addresses flagged "Not a ticket buyer" are excluded. The cohort itself is comp-filtered (comp recipients of the selected show are NOT counted as audience members); cross-attendance to other shows counts both paid and comp tickets.
Returning attendees ([Production]) -- One row each for the three most recent comparison-group productions whose run ended before the selected production opened, named with the production. The count is how many cohort members attended that prior production at any point in history.
Returning attendees (any production) -- How many cohort members attended any production in the comparison group, ever.

These rows have a single value (not broken out by window) because they're lifetime counts -- "did this person ever attend X".

Per-window rows¶

Under the "VS. [Comparison Group Names]" subheader and the "VS. FACILITY" subheader, each metric is shown across all five windows:

Other productions -- The number of distinct OTHER productions whose run overlapped the window in the scope (comparison group or facility). This is the ceiling for the visit-count metrics: if only 1 production ran, the maximum visit-count is 1.
First Time (by recent months) -- Cohort members with zero attendance to other productions in the scope whose runs overlapped the window. Wider windows generally yield fewer first-timers (more prior visits get discovered).
Dedicated customers -- Cohort members who attended every comparison-group production whose run overlapped the window. Zero by definition when no productions ran.
2+ visits in comparison / 3+ visits in building -- Cohort members who attended that many or more distinct productions in the scope whose runs overlapped the window. The narrower comparison-group threshold (2+) reflects how few productions usually run in a single company's slate, while the broader facility scope uses 3+.

Cells with value 0 are rendered as blank for readability.

What counts as a "visit"¶

A visit is a distinct production attendance, not a per-performance count:

Buying two tickets to the same performance of a show = 1 visit
Buying tickets to two different performances of the same show = 1 visit
Buying tickets to two different shows = 2 visits

A production is considered "in" a window if its run (first to last performance date) overlaps the window -- not if the customer's specific order date fell inside the window. This avoids edge cases where a production that ran across a window boundary was undercounted (e.g., a show that opened just before the 1-year mark and closed just after).

What's excluded¶

Addresses with neither an email nor a street address (line + zip) -- they can't represent a real person.
Addresses flagged "Not a ticket buyer" in the address record (placeholder records).
Comp-only orders to the selected production -- people in the cohort had to buy at least one paid ticket. (Comp recipients to OTHER shows still count as having attended those shows.)
Orders whose performance date is after the anchor date -- "ignore orders to other shows AFTER this date".
Productions on Presale status are not eligible as the selected production but are not filtered out of cross-attendance.

Reading the numbers¶

Two important sanity checks:

Wider window = fewer first-timers, more repeat-visitors. This is the expected direction. A 3-year window finds more prior attendance per person than 3 months, so "First Time" shrinks and "3+ visits" grows as you read left-to-right.
"Dedicated customers" is bounded by "Other productions". If only 1 production ran in the comparison group in 6 months, "Dedicated" = the count of cohort members who attended that 1 production. If 5 ran, only patrons who saw all 5 qualify.

If "Dedicated customers" is unexpectedly 0, check the "Other productions" row in the same column -- often the comparison group simply didn't run enough shows in that window for the bar to be meaningful.

Tips¶

Use the default comparison group (the production's own theater) when asking questions about how a specific company is retaining its audience.
Add additional theaters when asking questions about a coproducing relationship or a shared lane of programming (e.g., all musicals, all theaters tagged "Storefront").
The "VS. FACILITY" section is always shown, regardless of which theaters you pick -- use it to answer "how many Ally attendees are new to anything we host" as a counterpoint to the comparison group's "new specifically to Theater Wit-produced work".
Tag your theaters meaningfully (Theater tags) so you can build comparison groups quickly across recurring partnerships, neighborhood circuits, or programming categories.
The "Returning attendees ([prior production])" rows are the strongest predictor of loyalty -- a sizeable number there means your previous show carried over directly into ticket sales for this one.

Future capability¶

A planned next phase will let you export the cohort (and any sub-segment, like "3+ visits in building") as a CSV for outreach and marketing campaigns. The metric queries are already cohort-shaped, so the same numbers shown in the table will be exportable as the actual email list behind them.