3 How has giving by donation size changed over time?

Each household’s giving in a given year is put into one of four size tiers — <$100, $100–999, $1,000–4,999, and $5,000+ — so we can watch the mix of small and large donors shift over time. Giving is counted the same way as elsewhere: SUCCEEDED individual gifts (DONATION and PLEDGEPAYMENT), rolled up to the household (see Methodology). Donors are defined in Are large donors going away?; where they live is in Where do donors live?.

Code

import pandas as pd
import plotly.express as px
from hh.clean.accounts import clean_accounts
from hh.clean.donations import clean_donations
from hh.analytics.donors import annual_gifts, donors_by_size_year, TIER_LABELS
from hh.viz import style, SEQUENTIAL_BLUE, money

SIZE_COLORS = [SEQUENTIAL_BLUE[i] for i in (1, 5, 9, 12)]
COLOR_MAP = dict(zip(TIER_LABELS, SIZE_COLORS))

ag = annual_gifts(clean_donations(accounts=clean_accounts()))
sby = donors_by_size_year(ag)

# the latest year in the data is the current, partial year — keep it out of the trend
max_year = int(ag["year"].max())
complete_years = [y for y in sorted(ag["year"].unique()) if y < max_year]
plot_df = sby[sby["year"].isin(complete_years)].copy()
ytd_donors = int(sby.loc[sby["year"] == max_year, "donors"].sum())
ytd_dollars = float(sby.loc[sby["year"] == max_year, "dollars"].sum())

# lifetime tier totals (all years) for the summary table
ltot = sby.groupby("tier").agg(donors=("donors", "sum"), dollars=("dollars", "sum")).reindex(TIER_LABELS)
ltot["dollar share"] = ltot["dollars"] / ltot["dollars"].sum()

3.1 Donors and dollars by tier, by year

Code

fig = px.bar(
    plot_df,
    x="year",
    y="donors",
    color="tier",
    category_orders={"tier": TIER_LABELS, "year": complete_years},
    color_discrete_map=COLOR_MAP,
)
fig.update_layout(
    xaxis_title="Year (gift date)", yaxis_title="Donor households",
    legend_title="Annual giving", legend=dict(orientation="h", y=-0.2),
)
fig.update_traces(hovertemplate="%{y:,.0f} donor-years<br>%{fullData.name}<extra></extra>")
style(fig).show()

Figure 3.1: Active donor households by annual giving tier, by year

Code

fig = px.bar(
    plot_df,
    x="year",
    y="dollars",
    color="tier",
    category_orders={"tier": TIER_LABELS, "year": complete_years},
    color_discrete_map=COLOR_MAP,
)
fig.update_layout(
    xaxis_title="Year (gift date)", yaxis_title="Donated dollars",
    legend_title="Annual giving", legend=dict(orientation="h", y=-0.2),
)
fig.update_yaxes(tickprefix="$", tickformat=",.0f")
fig.update_traces(hovertemplate="%{y:$,.0f}<br>%{fullData.name}<extra></extra>")
style(fig).show()

Figure 3.2: Donated dollars by giving tier, by year

The small-donor base (under $100 a year) has grown sharply in recent years, while the number of households giving $1,000 or more in a year is small and uneven — its dollars, though, move the total most. The share of dollars coming from the top tiers:

Code

share = plot_df.assign(share=plot_df.groupby("year")["dollars"].transform(lambda s: s / s.sum()))
fig = px.bar(
    share, x="year", y="share", color="tier",
    category_orders={"tier": TIER_LABELS, "year": complete_years},
    color_discrete_map=COLOR_MAP,
)
fig.update_layout(
    xaxis_title="Year (gift date)", yaxis_title="Share of donated dollars",
    legend_title="Annual giving", legend=dict(orientation="h", y=-0.2),
    barmode="relative",
)
fig.update_yaxes(tickformat=".0%", range=[0, 1])
fig.update_traces(hovertemplate="%{y:.0%}<br>%{fullData.name}<extra></extra>")
style(fig).show()

3.2 Where the dollars concentrate

Across all years combined, giving is heavily concentrated in the top tiers:

Code

ltot.index.name = "Annual giving tier"
ltot.style.format(
    {"donors": "{:,.0f}", "dollars": money, "dollar share": "{:.0%}"}
)

	donors	dollars	dollar share
Annual giving tier
<$100	2,344	$64,224	4%
$100–999	1,762	$405,568	27%
$1,000–4,999	275	$455,736	31%
$5,000+	66	$562,241	38%

2026 is year-to-date (partial year), so it is left out of the trend charts. So far in 2026, 299 donor households have given about $37,158 — the full-year figure will be higher.