5  What do large donors attend?

Donors are split by lifetime giving — smaller donor (under $1,000 lifetime), large donor ($1,000–4,999), and major donor ($5,000+) — and compared with households that have never given. The question is whether larger donors engage differently: do they attend more, take more classes, and prefer different performance genres? Attendance is headcount (succeeded ticket attendees); see Methodology. Size tiers are defined in How has giving by donation size changed over time?; the binary donor view is in What performances do donors attend?.

import pandas as pd
import plotly.express as px
from hh.io import read_parquet
from hh.viz import style, BLUE, CATEGORICAL, SEQUENTIAL_BLUE, pct
from hh.analytics.productions import is_excluded, count_succeeded_attendees

GROUP_ORDER = [
    "Smaller donor (<$1,000)",
    "Large donor ($1,000–4,999)",
    "Major donor ($5,000+)",
    "Non-donor",
]
GROUP_COLORS = {
    "Smaller donor (<$1,000)": SEQUENTIAL_BLUE[3],
    "Large donor ($1,000–4,999)": SEQUENTIAL_BLUE[8],
    "Major donor ($5,000+)": SEQUENTIAL_BLUE[12],
    "Non-donor": "#c9c7c0",
}
GENRES = ["theater", "music", "opera", "dance"]
GENRE_COLORS = dict(zip(GENRES, CATEGORICAL[:4]))

hh = read_parquet("processed", "households_summary.parquet")
regs = read_parquet("processed", "registrations_enriched.parquet")


def _lifetime_group(r):
    if not r["is_donor"]:
        return "Non-donor"
    t = r["total_donations"]
    if t >= 5000:
        return "Major donor ($5,000+)"
    if t >= 1000:
        return "Large donor ($1,000–4,999)"
    return "Smaller donor (<$1,000)"


hh["lt_group"] = hh.apply(_lifetime_group, axis=1)

5.1 How engaged is each donor tier?

eng = (
    hh.groupby("lt_group")
    .agg(
        households=("id", "count"),
        patron_share=("is_patron", "mean"),
        class_share=("is_class_taker", "mean"),
        avg_perf_regs=("n_performance", "mean"),
        avg_class_regs=("n_class", "mean"),
    )
    .reindex(GROUP_ORDER)
)
eng_show = eng.copy()
eng_show.columns = ["Households", "Attend performances", "Take classes", "Avg performance regs", "Avg class regs"]
eng_show.style.format(
    {
        "Households": "{:,.0f}",
        "Attend performances": pct,
        "Take classes": pct,
        "Avg performance regs": "{:.1f}",
        "Avg class regs": "{:.1f}",
    }
)

	Households	Attend performances	Take classes	Avg performance regs	Avg class regs
lt_group
Smaller donor (<$1,000)	1,154	69.8%	47.6%	3.9	5.7
Large donor ($1,000–4,999)	129	83.7%	54.3%	16.5	8.0
Major donor ($5,000+)	49	91.8%	63.3%	40.3	12.3
Non-donor	3,322	56.4%	36.2%	1.2	1.6

eng_long = (
    eng[["patron_share", "class_share"]]
    .rename(columns={"patron_share": "Attend performances", "class_share": "Take classes"})
    .reset_index()
    .melt("lt_group", var_name="Activity", value_name="share")
)
fig = px.bar(
    eng_long, x="lt_group", y="share", color="Activity", barmode="group",
    category_orders={"lt_group": GROUP_ORDER},
    color_discrete_map={"Attend performances": CATEGORICAL[0], "Take classes": CATEGORICAL[1]},
)
fig.update_layout(
    xaxis_title="", yaxis_title="Share of households", legend_title="",
    legend=dict(orientation="h", y=-0.25),
)
fig.update_yaxes(tickformat=".0%", range=[0, 1])
fig.update_traces(hovertemplate="%{y:.0%}<extra></extra>")
style(fig).show()

Figure 5.1: Share of each lifetime-donor tier that attends performances or takes classes

5.2 Which performances draw the large-donor tiers?

Performance headcount by genre, split by lifetime-donor tier:

m = regs.merge(hh[["id", "lt_group"]], on="id", how="left")
perf = m[(m["event_majorcat"] == "performance") & ~m["event_name"].map(is_excluded)].copy()
perf["attendees"] = perf["tickets"].map(count_succeeded_attendees)

gxg = perf.groupby(["lt_group", "event_minorcat"])["attendees"].sum().unstack(fill_value=0)
gxg = gxg.reindex(columns=GENRES).fillna(0)
genre_mix = gxg.div(gxg.sum(axis=1), axis=0).reindex(GROUP_ORDER)

mix_long = genre_mix.reset_index().melt("lt_group", var_name="genre", value_name="share")
fig = px.bar(
    mix_long, x="lt_group", y="share", color="genre",
    category_orders={"lt_group": GROUP_ORDER, "genre": GENRES},
    color_discrete_map=GENRE_COLORS,
)
fig.update_layout(
    xaxis_title="", yaxis_title="Share of performance attendance",
    legend_title="Genre", legend=dict(orientation="h", y=-0.25),
    barmode="relative",
)
fig.update_yaxes(tickformat=".0%", range=[0, 1])
fig.update_traces(hovertemplate="%{y:.0%} %{fullData.name}<extra></extra>")
style(fig).show()

Figure 5.2: Performance-genre mix within each lifetime-donor tier

Major donors are the most likely to attend performances (92% of them are patrons) and the most likely to take classes (63%). The genres they attend mirror the rest of the audience — theater first — but opera and music take a noticeably larger share of their attendance than of non-donors’.