2020 Post-Election Analysis

May 2021

1
Project Overview
The fundamental question: What happened down ballot in 2020?

2
 Project Overview

A Message from Marlon Marshall and Lynda Tran be a standout in the last half-century of congressional gains for
Democrats and approached our analysis with tempered
expectations about what “should” have happened across the
Our team set out to conduct a postmortem analysis of the 2020
battleground maps. Lastly, we were conscious of the ongoing
cycle that would offer an honest and data-driven look at both what
debate around polling accuracy and effectiveness, and we
worked well for Democrats running for the U.S. House of
evaluated the public and proprietary polling we received
Representatives and the U.S. Senate – and where the challenges
accordingly.
our candidates experienced call for meaningful solutions before
voters head back to the ballot box in just two short years.
Over 6 months, we explored the national and regional message
themes that played out in the campaigns’ own media and other
While we did not embark on the project with preconceived notions
assets, reviewed the infrastructure and ecosystem that
about what strategies, tactics, or issues impacted electoral
supported these campaigns, and conducted our own deep dive
outcomes, we believe we performed our work with eyes wide open
into the data and analytics using both final state voter files and
to the public discourse and historic health and economic crises that
public and proprietary polling. While we intentionally devoted
touched every major campaign – and every individual in America –
the bulk of our time and effort to gathering first-person
in 2020. We knew it was important to explore the COVID-19
observations based on the campaigns’ lived experience,
pandemic and its impact on everything from how campaigns
throughout the project, we continually sought out additional
reached their supporters to whether and how voters were able to
make their voices heard. And we were professionally and data, integrated new analyses of 2020 demographics and
outcomes, and incorporated the wisdom and experience of
personally conscious of the heightened role that race played in
expert voices from across the Democratic political spectrum.
these elections – both in the GOP’s bold-faced efforts to divide
Americans and in helping to drive civic engagement and voter
Although we explicitly focused our efforts on answering the
enthusiasm in critical races nationally.
question of “What happened in 2020?” rather than prescribing
what should happen next, we hope this analysis informs
We also went into the analysis grounded in many cycles of
Democratic Party and broader movement leaders as they
experience and data on voter participation and longitudinal turnout
begin planning for 2022 and beyond.
trends. As such, we understood the “blue wave” 2018 elections to

3
What We Found

4
 What We Found | Our findings

1. Voters of color are persuasion voters who need to be

Our findings convinced

2. Republican attempts to brand Democrats as “radicals” worked

3. Polling was a huge problem – even after 2016 adjustments

4. COVID-19 affected everything

5. Year round organizing worked, as did cross-Party collaboration

6. Our hopes for 2020 were just too high

5
1. Voters of color are persuasion
voters who need to be convinced

6
 Our Findings | Voters of color are persuasion voters

Our assumptions about Dem support among voters of color –

Our approach and the lack of differentiation in our messaging and outreach
within demographic groups – cost us support in key races

to voters of
• As in previous cycles, Democrats generally treated Black, Latino, and AAPI
voters as GOTV targets, concentrating outreach efforts closer to Election
Day and in some cases after early vote was already underway

color • Despite historic turnout, even where Black voters were key to Democratic
successes this past cycle – including in GA, AZ, and MI – the data show drop-
off in support in 2020 compared to 2016 and 2018

significantly • Drop-off in support among Latino and Hispanic voters were the lynchpin in
Democratic losses in races in FL, TX, and NM – especially among working

hurt our
class and non-college voters in these communities
• Drop-off in AAPI support fueled losses in key races in CA, especially among
Vietnamese and Filipino voters

outcomes • Campaign messaging to these groups typically did not account for differing
perceptions among gender, age groups, educational attainment, geography,
or country of origin – and there was a dearth of message research on Black
voters in particular
• Black, AAPI, and Latino and Hispanic voters were targeted by misinformation
and disinformation efforts both online and on-the-ground – yet campaigns
were largely unable to either assess the reach or respond in real time

7
 Our Findings | Voters of color are persuasion voters

The Party We heard that there were substantial Party and campaign
misfires for engaging key demographic groups as a monolith –
most notably Latino and Hispanic voters

treated ● National strategy failed to take into account regional and local differences,

voters of
socioeconomic status, urbanicity, or country of origin – despite higher
support for Democrats among voters from Mexico, Puerto Rico, and the
Dominican Republic compared to Cuban-American voters

color as a
● Latino and Hispanic voters were broadly treated as GOTV targets rather
than audiences for persuasion earlier in the cycle, and the modeling,
polling, and subsequent campaign decision making reflected this

monolith –
assumption
● Campaign messaging didn’t always reflect the differing values and
priorities of urban Hispanic voters vs. rural Hispanic voters, much less

especially account for what would most persuade Hispanic men in the Rio Grande
Valley, oil and gas workers in New Mexico, or Latinas in South Florida

Latino voters

8
 Our Findings | Voters of color are persuasion voters

Latino voters Turnout among Latino voters grew dramatically in 2020 - rising
by more than 30% compared to 2016

made a
● Latino support was critical for the top of the ticket but especially in tight
states like AZ, PA, and NV where the margin of victory among Latino
voters exceeded Biden’s win margin with the overall electorate

difference in
● Latino voters were also a lynchpin in Sen. Kelly’s win in Arizona where his
support nearly matched 2016 levels – but Biden saw heavier losses among
Latino voters, indicating that some split ticket voting took place

top races Despite increased Latino turnout overall, Democrats saw a

significant dip in support in places with high concentrations of
Latino and Hispanic voters – including in south FL and west TX
● Part of the underperformance in 2020 is a reflection of Clinton's strength
with Latino voters in 2016, when she outperformed Obama's 2012 support

● In FL-26, Republicans took a big lead in voter registration in 2020, saw

higher turnout among Latino voters, and netted more of their support
● In TX-23, despite growing their share of the overall electorate, support
among rural Latino voters shifted right by 13 points in 2020, mostly due to
higher turnout among Latino Republicans

9
 Our Findings | Voters of color are persuasion voters

Latinos continue to be 10 points more Dem than the rest of the electorate, but
our advantage has shrunk*

Latino voters have grown their share of the electorate in recent cycles, but in 2020, Democratic support among Latino voters
dropped in South Florida and parts of Texas

FL-26: TX-23:
Latino Precincts (>80%) Maj. Latino Precincts

64% 66% 65%

58% 59%
52%

40% 38% 40%

34%

'16 CD CD '20 CD '12 '16 CD '20 CD

'16 '18 '20 '18 '20

*Source: Immigration Hub Analysis on Latino Voters in 2020 10

 Our Findings | Voters of color are persuasion voters

Even in Arizona where Mark Kelly retained support from Latino voters,
Republicans made inroads at the top of the ticket

Presidential + 2020 Senate Special

Support-wise, Sen. Kelly was able to retain near 2016 level Dem vote share by demo composition
support among Latino voters, while Biden saw heavier
losses among Latinos than Kelly - indicating some split White Precincts (>90%) Latino Precincts (>60%)
ticket voting took place

77% 79% 77%

75%

37% 39% 39%

34%

'12 '16 '20 SEN '20 '12 '16 '20 SEN '20

11
 Our Findings | Voters of color are persuasion voters

In FL-26, Democratic registration among Latino new registrants dropped by 8

points since 2016, while Republicans saw a 13-point increase
Meanwhile, Latino voters across Miami-Dade were less likely to identify as Democrats in 2020 relative to 2016
and more likely to identify with neither party

Party Registration in 2016 Party Registration in 2020

New Registrants New Registrants

DEM OTHER REP DEM OTHER REP

OU TER R ING SUB URB S 34% 48% 18% OU TER R ING SUB URB S 27% 43% 30%

INNER RING SUBUR BS 37% 46% 16% INNER RING SUBUR BS 30% 43% 27%

UR BAN M IAM I-DADE 33% 49% 18% UR BAN M IAM I-DADE 25% 44% 32%

FL K EY S 23% 43% 34% FL K EY S 25% 35% 40%

WHI TE 26% 46% 29% WHI TE 25% 38% 37%

BLAC K 64% 33% BLAC K 67% 29%

LATINO 32% 49% 19% LATINO 24% 44% 32%

LATINO (CU BAN) 30% 48% 22% LATINO (CU BAN) 21% 42% 37%

LATINO (NON-CUB AN) 33% 50% 17% LATINO (NON-CUB AN) 26% 45% 29%

12
 Our Findings | Voters of color are persuasion voters

Black voters The data show that strong vote share among Black voters –
despite some downward trends – likely made the difference in
top races

made a ● Dramatic shifts in the electorate helped many Tier 1 districts become bluer

difference in
in 2020 – but Biden was more successful in cementing this baseline
support while many Congressional candidates lagged behind, including in
CA-39, TX-23, and TX-24

top races ● Although Democrats saw a modest dip in support among Black voters
(particularly in VA-07 and the NC Senate race), Black voter turnout
increased substantially nationally — resulting in more net Democratic
votes from Black voters in 2020 than in 2016
● Black voters were critical to Democrats’ victories in the Georgia runoffs —
with higher turnout and less drop-off than other groups heading from the
general into the runoff elections

13
 Our Findings | Voters of color are persuasion voters

Black voters were critical to Democrats’ victories in the Georgia runoff

elections and helped cement Democratic control of the U.S. Senate
Although turnout among Black voters dropped relative to 2016, Sen. Warnock overperformed both Clinton’s 2016 and
Biden’s 2020 support, and performed similarly to Obama’s 2012 margin

GA: Presidential + 2021 Runoff

Dem vote share by demo composition

White Precincts (>90%) Black Precincts (>90%) Lat-Am Precincts (>20%) AAPI Precincts (>20%)
N=423 N=340 N=41 N=28

98% 96% 97%

95%

65%
61% 61%
57% 58% 59%
53%

41%

19% 21% 21%

18%

'12 '16 '20 '21 '12 '16 '20 '21 '12 '16 '20 '21 '12 '16 '20 '21

14
 Our Findings | Voters of color are persuasion voters

Black voters In 2020, Black voters comprised a smaller share of the overall
electorate. A substantial boost in turnout netted Democrats more raw votes from
Black voters than 2016, but the explosive growth among white voters in most races

turned out in
outpaced these gains

Despite increased turnout, Democrats saw a modest dip in support

big numbers – ● In the North Carolina Senate race, Black voters grew their share of the electorate
in 2020 and saw increased turnout over 2016 – but the GOP also saw a slight

but their ●
improvement among Black rural / exurban voters

In VA-07, the turnout gap between all voters of color and white voters increased

overall share
compared to 2016, benefiting the GOP. Democrats’ support also fell in majority
Black precincts, although turnout in these areas was up relative to 2016

dropped
Still, Black voters were critical to Democratic wins in the GA runoff
elections and helped cement Democratic control of the U.S. Senate

● Turnout among Black voters dropped relative to 2016, but Sen. Warnock
overperformed both Clinton’s 2016 and Biden’s 2020 support, and performed
similarly to Obama’s 2012 margin

● While turnout among all racial groups dropped heading into the runoffs, the drop-
off was the least concentrated among Black voters

● Drop-off was also less prominent among voters in urban/inner ring areas – key
constituencies for both Dem candidates – while drop-off was steeper for white
rural voters (which hurt Republican candidates)

15
 Our Findings | Voters of color are persuasion voters

Despite
In MI-08, Slotkin made gains over 2016
Democratic support has slightly
Dem performance in areas with more
eroded in majority-Black precincts
voters of color, but lost some support

robust
in NC over the past three cycles
since her first cycle in 2018

support from
MI-08: Dem Support in NC Senate: Dem Support in
Diverse Precincts Black Precincts (>75%)
(Voters of Color >30%)

Black voters,
PRES CONG

95% 92%

there are
91% 91%
80% 78%
76%

warning signs 73% 76%

for Dems

'16 '18 '20 '12 '16 '20 '20

Pres Sen

16
 Our Findings | Voters of color are persuasion voters

In NC, Republicans made greater gains among Black supporters, particularly in

non-urban areas
Modeled Dem Turnout Prev Reg Modeled GOP Turnout Prev Reg

2016 TURNOUT 2020 TURNOUT 2016 TURNOUT 2020 TURNOUT

+2PP +5PP +7PP +8PP +1PP +6PP +9PP +11PP

74% 75%
70% 72%
68% 67% 67%
63% 64%
61% 61%
55% 57% 58%
51% 52%

BLA CK BLA CK BLA CK BLA CK BLA CK BLA CK BLA CK BLA CK

URBA N SUB I NNER SUB OUTE R RURA L URBA N SUB I NNER SUB OUTE R RURA L

17
 Our Findings | Voters of color are persuasion voters

AAPI voters The data show rising electoral participation from AAPI voters
likely made the difference in top races – including some of the

made a
closest presidential contests this cycle

difference in
● AAPI voters saw significant growth in their electorates with large
increases in registration and turnout

top races
● Over 4 million ballots were cast by AAPI voters in 2020 – a 47% increase
over their 2.8 million turnout in 2016 and a substantially sharper rise than
the 12% increase among all other voters
● Georgia saw the second highest state-level increase in AAPI votes in the
nation with the surge exceeding Biden’s win margin – AAPI votes grew by
almost 62,000 votes over 2016 in a state where Biden won by just 12,000
votes
● Arizona was among the top ten states that experienced a surge in AAPI
voters
• In CA-39, AAPI voters now comprise a quarter of the electorate (a 6-point
growth since 2016)
• While Biden grew support among this community since 2016 (three points
in TX-24, one point in CA-39), congressional candidates underperformed
him significantly in both districts

18
 Our Findings | Voters of color are persuasion voters

In CA-39, the share of AAPI new registrants has been steadily increasing since
2016 and in 2020 – now accounting for nearly 1 in 3 new voters
Although Cisneros was unsuccessful in his re-election bid, the growth in registration and higher turnout among AAPI voters
helped Biden improve on 2016 support

CA-39: Composition of New Registered Voters, 2016-2020

< 30 30-39 40-49 50-64 65+ AAPI BLACK LATINO WHITE
2020 25% 18% 16% 24% 18% 28% 6% 31% 35%

2018 32% 19% 15% 20% 13% 25% 6% 32% 37%

2016 41% 17% 14% 17% 10% 21% 6% 32% 40%

AAPI COL AAPI NON-COL WH COL WH NON-COL AAPI-URB. AAPI-INNER WH-URB WH-INNER
LAT COL LAT NON-C OL OTH LAT-URB LAT-INNER OTHER

2020 20% 8% 24% 12% 12% 19% 6% 9% 15% 15% 17% 18% 11% 15%

2018 22% 3% 31% 7% 10% 21% 6% 9% 13% 17% 19% 19% 10% 14%

2016 19% 2% 33% 7% 10% 22% 6% 8% 11% 19% 18% 21% 9% 13%

19
 Our Findings | Voters of color are persuasion voters

Congressional In TX-24, growth in the AAPI

electorate did not translate into votes
In CA-39, although Cisneros improved
on his 2018 support in Asian American

candidates
for Valenzuela, who underperformed precincts, he still received roughly
Biden 2020 and Clinton 2016 support 9 points less than Biden

underperformed
TX-24: Dem Support CA-39: Dem Support
in AAPI Precincts (>20%) in AAPI Precincts (>50%)

Biden with AAPI

PRES CONG

voters 60% 63%

59% 58% 59%
48%

47% 50%
39%

'12 '16 '20 '20 '16 '18 '20

Pres Pres Pres CD

20
2. Republican attempts to brand
Democrats as “radicals” worked

21
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

Messaging
Win or lose, self-described progressive or moderate, Democrats
consistently raised a lack of a strong Democratic Party brand as a
significant concern in 2020

mattered – in ● Campaign public discourse – in paid media, earned media, and direct voter
contact – and message and opinion research indicated that voters recalled
attacks related to “radical socialists,” “Defund the Police,” and related

good ways messages

○ In the districts we looked at, those in which “law and order” or “socialism” was

and bad
a continued drumbeat also saw a higher share of Latino/AAPI/Black voters
who supported the GOP

○ Districts with significant rural populations required in-person voter contact to

effectively push back on these attacks, but most chose not to do so

● Campaigns that were able to reach constituents with local messages –

especially incumbents who could point to specific legislative wins – fared best
● Anecdotally, messages about jobs, the economy, and rebuilding post-COVID
were most effective – but the GOP successfully branded Dems as the Party
that would keep the economy shut down
● Republicans (Trump) successfully raised concerns about election security and
VBM – potentially dampening their own voters’ enthusiasm
● GOP media organizations and overall structure made them more effective
than Dems

22
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

Race was a primary focus throughout the cycle – Campaigns and practitioners widely described
most notably as Black Lives Matter protests and 2020 messaging as nationalized with Trump at
the Defund the Police movement took off in the the top of the ticket
summer of 2020
• One theme that arose was Party messaging that leaned too
heavily on "anti-Trump" rhetoric without harnessing a strong
• One candidate’s internal campaign polling in summer 2020
economic frame
showed that voters didn't believe they "shared their values," a
notable shift from earlier research that they interpreted as a need • As a former Member of Congress described the issue, Democrats
focused on arguing that Trump was bad, not why a Democratic
to address the Defund the Police movement – which they did in
majority would help voters, “It was the lack of an economic plan
ads featuring a former Republican law enforcement officer as a
that really hurt”
validator for the candidate’s values and character
• In NY-02, when Jackie Gordon was asked if she supported • Some campaign teams we spoke with felt that the Party didn’t
have a message beyond “Donald Trump sucks,” and this void led
defunding the police, she responded by pointing to her decades
to split-ticket voting for Biden at the top of the ticket and
long career in law enforcement, as a military police officer,
Republicans down ballot (leading to split outcomes in 16 districts
working with the police as a member of the Babylon town
nationwide)
council, and then as a guidance counselor, inviting police into her
o In this absence of strong party branding, the opposition
school – internal polling indicated that constituents didn't buy this
latched on to GOP talking points, suggesting our candidates
line of opposition attack
would “burn down your house and take away the police.”
• In NE-02, Kara Eastman refuted the claim that she wanted to
defund the police both in debates and on social media, and cited
o Dana Balter described NY-24 not as a swing district, but a
split ticket district, where "we vote Democrat for president
her past experience working with police departments as a social
and Republican for congress"
worker, but she did not feel her approach blunted the attacks

23
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

Candidates The top of the ticket ultimately mattered the most, but the
personal record and bio of federal candidates mattered, too –
especially where campaigns had early funding

mattered ● Consistent with previous cycles, presidential performance had the largest
impact on down ballot races – just 16 of all federal races in 2020 saw split-
ticket outcomes

● Most federal candidates that had the resources to offer up their bio earlier
– and to maintain a bio track in their paid media for the duration of the
cycle – fared best, winning by an average margin of D+1.8 (for Senate
candidates) and D+5.4 (for House candidates) and some outperforming
the top of the ticket
● Candidates who won districts that President Biden did not relied on local
knowledge, a personal bio that resonated with voters (often with a
background in national security), and disciplined campaigns that stayed
on message and seamlessly shifted its operations in a tumultuous
campaign cycle

24
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

o The campaign was disciplined and focused on his bio for the
Candidates with strong, homegrown biographies
duration of the campaign, weaving in positive and negative
that had the resources to stay on that message messaging closer to Election Day
throughout their campaigns tended to win

• In MI-08, Congresswoman Elissa Slotkin’s team developed their

messaging in January of 2020 focused on her biography and
used it to build a bridge to the voters in her district who weren’t
predisposed Democratic base voters
o Rep. Slotkin messaged on health care, sharing the personal
story about her mother’s cancer and lack of health insurance
• In NY-19, Congressman Antonio Delgado spent 2019 doing 33
town halls across all 11 counties in his district, ensuring that the
message of being available and working hard for his constituents
was front and center
o When the pandemic hit, the campaign stayed focused on Positive messaging and branding worked –
meeting the needs of constituents, which served as a way to especially early on. Later in the cycle, the
inoculate themselves from outside narratives narrative was too cemented to break through no
• In the Arizona Senate race, Senator Mark Kelly’s campaign went matter how much a campaign spent in its closing
up early – the first full week of September – and focused heavily days.
on his bio and appealing to multiple demographics needed to win
statewide
o As one strategist noted, people responded to different parts
of his story in extremely positive ways: “Seniors loved that
he is an astronaut, Latinos loved his commitment to family,
and non-college educated white men loved that he is a
combat veteran”

25
“This election, Republicans tended to reach
people and connect at an emotional level and
Dems tried to connect with people at an
intellectual level.”
– National media consultant

26
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

Incumbent Democratic candidates focused their But Republicans successfully painted Democrats
COVID messaging on constituency services as the Party that wanted to keep the country shut
down
• In NY-19, the Delgado team called constituents to see what they
needed rather than to ask for their vote
• According to a leader of a national nonprofit organization, “We
• In NJ-03, the focus was on being in this together, and
Representative Kim was able to visit small businesses (while
were right to shut down the economy and mandate masks, but
we also needed to make clear that was because we stood for
following COVID safety protocols) getting the economy, schools and stadiums open ASAP. There is
• In MI-08, Representative Slotkin utilized messaging about PPE a line between being the adults in the room and being the nanny
being “Made in America,” which was specific to her district state, and during the election cycle, we did not make clear that
because of its industrial nature we were the party of solutions [in the way President Biden has
• In GA-06, Rep. Lucy McBath’s campaign research showed that done so successfully since taking office.]”
messaging on how the Congresswoman has delivered for
veterans by sponsoring the HAVEN Act to “stand up for our
• In FL-26, where a substantial portion of the population works on
cruise ships, in hospitality, and in the tourism industry,
veterans by amending current bankruptcy law and supporting constituents didn’t want to hear the message of “stay home,” but
disabled veterans’ eligibility for relief during financial hardship” looked to leaders for answers on when they would be able to get
resonated with voters back to work

When Democrats focused their COVID

messaging solely on health care without talking
about the economy, it exposed them to GOP
attacks.

27
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

Republicans universally attacked Democrats with

the “socialist” label, and moderate Democrats
struggled to counter since some within the Party
identify as Democratic socialists

• The socialism attack was called out by many Members of

Congress and candidates who were not successful who believe it
hurt campaigns in states and districts with immigrant populations
that fled socialist governments, including among Venezuelan,
Cuban, Vietnamese, and Filipino voters
• In NJ-07, the Tom Malinowski campaign focused their messaging
on issues and their Republican opponent’s record in the state
senate – but challenger State Senator Tom Kean’s campaign
promoted QAnon conspiracy theorists and connected Malinowski
to pedophiles, drowning out any conversations about policy The “Dem potpourri” of attacks meant to brand
• As the Malinowski campaign reported, “We were talking issues
and our opponent's record in the state senate. They were talking
Democrats as “radical” was effective –
especially where there wasn’t enough positive
about pedophiles. QAnon became a huge presence in our race,
bio early and where campaigns failed to
And drowned out conversations about policy. It felt like
respond to the lies.
messaging became a wash. In a district as tight as this turning out
the base is what mattered, and their motivator was partisanship”

28
Our Findings | Republican attempts to brand Democrats as “radicals” worked

29
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

While practically everyone – candidates, • At one end of the spectrum, candidates like Rep. Spanberger
(VA-07) shared widely reported concerns that Republican efforts
journalists, Party leaders, and pundits – has to paint all Democrats as socialists who would defund the police
focused on the impact of Defund the Police as a cost us multiple seats this cycle
solo variable, it was just one of the multiple issues • Similarly, Cameron Webb’s campaign (VA-05) constantly had to
that Republicans used to paint Democrats as bat back claims that he supported defunding the police, even
radicals after running ads that featured local law enforcement as
validators, and after the paper of record, the Roanoke Times,

• This “Dem potpourri” included efforts to tie candidates to debunked the claim. It was particularly difficult for candidates of
color to avoid these race-based attacks
socialism, Medicare for All, the Green New Deal, Nancy Pelosi,
AOC and “the Squad” (in these cases all framed as equally • In other districts, Defund the Police was perceived as a minor blip
among the other “Dem potpourri” attacks. For example, in FL-26,
radical). These attacks were used to a greater or lesser degree
depending on the unique constituencies of a given district attacks against former Rep. Debbie Mucarsel-Powell focused

• There was no one best way to counter these attacks, nor was more on messages of socialism and support for women’s choice –
issues that were more effective with her Hispanic and Latino
there clear agreement even that we should address the issue in
paid communications constituents

• The GOP consistently framed the moment in “law and order” • And in some races, such as the North Carolina Senate race,
Defund the Police "was in the bloodstream because it was part of
terms, which resonated with voters on both the right and the left
– including Latino men and women, as the April 2021 EquisLabs the presidential back and forth," but it was never really a huge
report “Portrait of a Persuadable Latino” called out part of the Senate campaign messaging on either side

Based on interviews, data analysis, polling, and

How much Defund the Police dominated public ads analysis, where Defund the Police had a
discourse – and by extension how much it significant impact, it was as a part of culture-
required campaign resources to address the issue based attack on Democrats that sought to
– varied widely by state and district stoke fears among voters about any candidate
with a ‘D’ after their name
30
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

None of the candidates or campaigns included in Some candidates reported that it was difficult to
this analysis supported defunding the police, but determine whether Republican Defund the Police
nearly all were targeted with paid ads claiming attacks were effectively landing – and that
they did challenge impacted whether and how they
responded
• In our review of the 20 digital ads with the highest number of
impressions run by each candidate in 19 races (760 total ads), 52 • Some campaigns that did not specifically test “Defund the Police”
mentioned Defund the Police – with 2 run by Democrats and 50 in their polling reported their belief that other indicators that they
run by Republicans were tracking – such as “shares my values” – were a proxy for
• Overall, Defund the Police was mentioned in 14% of the GOP how these attacks were impacting their overall favorability
among voters and made messaging decisions accordingly
digital ads with the highest number of impressions compared to
just 1% of the Democratic digital ads with the highest impressions • Candace Valenzuela’s polling in TX-24 showed that attacks
• Even in races where Defund the Police did not appear in paid
advertising, candidates reported that constituents asked about it
related to Defund the Police weren’t any more impactful than
other attacks, but the candidate reported that “people are
during campaign events, indicating that the message was missing the forest for the trees” and suspected these attacks
breaking through in other channels were more pernicious and damaging than polling indicated
• One well-resourced campaign reported that while they believed
that Defund the Police and some of the more extreme “radical
Democrat” lines of attack might have resonated with voters, they
Whether and how candidates responded to ultimately did not use their prepared response ads, because their
these attacks depended on the interplay of opponent did not put money behind any of these attacks
district and candidate demographics and
background
• Some candidates opted against responding directly to
Republican assertions, believing that addressing this complicated
issue would only give more oxygen to a message that voters
weren’t buying, particularly when that push back might alienate
some of the candidates’ activist bases who made up the majority
of volunteers
31
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

• Many consultants reinforced this conventional wisdom that Responses that featured endorsements from local
candidates shouldn’t give oxygen to attacks, and counseled
sticking to their core positive campaign messages – with the law enforcement have been widely hailed as
notable exception of one pollster, who advocated strongly to “effective,” but in hindsight some candidates and
candidates that they should tackle the issue head on. As he put it, Party leaders expressed regret that there wasn’t a
“You have to show up to the debate if you're going to win it.”
more forceful emotion- and values-based counter

Perceptions among voters on Defund the Police • Law-enforcement endorsement ads addressed the substance of
the issue but avoided/missed the emotion of Republican attacks
varied depending on how the issue was framed – and didn’t call them what they were – racist distractions from the
in-cycle and post-election research has continued issues voters cared about most
to show conflicted views across and within • Some Party leadership indicated they wished they had called out
demographic groups Republicans for distracting from ACA and other important issues

• Polling conducted by Brilliant Corners’ Cornell Belcher last

summer for Run For Something and the Collective PAC found
that “defunding the police does not have wide appeal, but Black
voters are divided on the issue,” however, the same poll found
that “Reallocating police department funding is popular amongst
Biden supporters and Black voters”
• Similarly, findings from an Avalanche Research poll in April 2021
found that while 63% of voters disagree with “defunding the
police,” 52% agree with “redirecting resources from police and
funding community programs”

32
“The primary problem with Defund was not
Defund, but the lack of an economic message.
We became the party of shutting down the
economy, the party of wearing masks, the
party of taking kids out of school – not the
party of solutions and science.”
– Major Democratic funder

33
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

Guantanamo Bay as military police, responded to questions about

Republicans used Defund the Police attacks to
defunding the police with “I am the police.” Her ads similarly
best effect against candidates of color in swing reinforced her background as a military police officer, without
districts with large white populations addressing Defund the Police directly
• Both Party committees commissioned message research
• Candidates we spoke to universally shared dismay and deep conducted by longtime Democratic strategists Cornell Belcher
anguish in response to George Floyd’s murder, and supported and Karen Finney, provided these resources to all their
widespread popular demands for criminal justice reform – and campaigns, and encouraged candidates to do follow up polling in
many Frontline members expressed their desire to leverage their their districts. Many candidates chose not to, but likely would
position to support these efforts have benefited from research on messaging that would be
• Sitting U.S. Representatives including Lauren Underwood (IL-14), effective in particularly challenging districts, either due to the
Abigail Spanberger (VA-07), Debbie Mucarsel-Powell (FL-26), demographic makeup or other local dynamics
Andy Kim (NJ-03), Elissa Slotkin (MI-08), Xochitl Torres Small
(NM-02), and Gil Cisneros (CA-39) turned that energy into
legislation and passed the George Floyd Justice in Policing Act of
2020
• Some candidates of color reported feeling particularly wary of
amplifying discussion or debate of Defund the Police out of fear Responding to Defund the Police attacks was
for their own safety most challenging and complicated for
• Some candidates from more conservative districts, or who had
law enforcement-related backgrounds, did respond directly to
candidates of color and they found generic
response strategies extremely frustrating
attacks in paid communications – Sen. Ben Ray Lujan (NM), for
example, ran ads featuring local law enforcement officers
endorsing his candidacy
• Jackie Gordon (NY-02), a combat veteran who served at

34
 Our Findings | Republican attempts to brand Democrats as “radicals” worked

Some districts where “law and order” or “socialism” was a drumbeat also saw
a higher share of Latino/AAPI/Black voters who supported the GOP
GOP Support Among Voters of Color
2016 - 2020

Latino Precincts AAPI Precincts Black Precincts

(>30%) (>20%) (>50%)

60%

50%

41% 41%
35% 36%

2018 CD topline not avail.

25% 25%
21%
2016 Pres.

2016 Pres.

2016 Pres.

2016 Pres.

2016 Pres.
2016 Pres.

2016 Pres.
2020 CD

2020 CD

2020 CD

2020 CD

2020 CD

2020 CD
2020 CD
2018 CD

2018 CD

2018 CD

2018 CD

2018 CD

2018 CD
TX-2 3 FL-2 6 NM-02 CA-39 TX-2 4 CA-39 VA-07
35
“[Republicans] were going to churches and
having pastors spread misinformation that
Democrats were baby-killers, that we were in
favor of killing babies right before a mother
was giving birth, they were showing very
graphic images…[W]hen you hear it enough
times, you start to believe it.”
– Former Rep. Debbie Mucarsel-Powell

36
3. Polling was a huge problem –
even after 2016 adjustments

37
 Our Findings | Polling was a huge problem

Polling was a Campaigns and experts reported that polling was a problem
despite adjustments made after 2016 and was not always well-
understood by campaigns as tools for decision making

huge problem ● Polling failed to reach the right people – especially white non-college
voters – and the methodology changes made in the wake of the 2016
cycle to weight for education level were insufficient to ensure accurate
sampling
● First-time candidates especially needed support in interpreting the
content and implications for taking risks or being creative in a year that
forced everyone to revise their campaign playbooks
● These challenges led to an overly optimistic map and the Party chasing
races that have not historically been in play, such as the Iowa Senate seat

38
 Our Findings | Polling was a huge problem

Polling in the 2020 cycle was widely viewed as a

catastrophic misfire from the top of the ticket • The Party committees invested their financial and staff resources
in statewide races and congressional districts that were either
down landslide victories (where additional funding was not needed) or
wide losses based on close polling (where additional funding
Certainly, there were some examples where pre-election polls couldn’t surmount district-specific obstacles)
aligned with the ultimate electoral performance – polls quite • Declining response rates further exacerbated issues of non-
response bias and led pollsters to miss Republican support
precisely predicted the outcome of the Warnock-Loeffler runoff
election in Georgia and the Kelly race in Arizona. However, • Insufficient sample sizes of Latino voters in pre-election polls
candidates, campaign and Party leadership, and leading greatly contributed to the Party failing to anticipate shifts in
practitioners universally agreed that the dramatic overestimation of support
Democratic support in both publicly-released and proprietary polls
reviewed as part of this analysis had an indisputable impact on While Democratic polling professionals are continuing to evaluate
races in every part of the country: what happened in 2020 – and importantly, are doing
unprecedented knowledge sharing across the industry to inform

• Inflated polling led Democrats to believe some stretch races future work (Source: Democracy Docket post) – pollsters and
operatives interviewed for this postmortem indicated they are
were competitive and to take for granted other races as easy
considering methodology and the unique variables at play with
wins, leading to an overly-expansive electoral map
Donald Trump on the ballot.
• Supporters poured resources into races where historical trends
showed a Democratic victory was unlikely but polling suggested
was winnable What’s clear is the “soul-searching” that followed the 2016 election

• Campaigns made budgetary decisions based on polling – and in and precipitated changes for the 2018 cycle (such as increased use
of mobile numbers over landlines and more multi-modal surveys)
some cases were risk averse in their messaging, voter contact,
did not change the fundamental approach. Current polling relies
and media plans when they believed their races were especially
close – or if they thought they were winning by wide margins heavily on past practices with experienced pollsters weighting
according to their assumptions about the electorate – which
impacts both the sampling and the outcome.

39
 Our Findings | Polling was a huge problem

The myth of the “shy Trump voter” The 2020 election cycle reinforced concerns
about non-response bias – especially among
• Despite being debunked in 2016, the “shy Trump voter” theory white non-college educated voters
that polling issues were due to Trump supporters’ reluctance to
share their honest opinions leapt back on the scene in the 2020
cycle with so many polls showing Democrats with a wide lead
• The success and accuracy of polling is predicated on the idea
that the people who participate in a survey are an accurate
• Following the 2016 election, the polling industry concluded that representation of the larger population – but as survey
issues were caused by a lack of proper educational weights participation falls, concerns about the representativeness of
applied on survey data and late movement among undecided
those respondents increases

•
voters
In 2020, pollsters reported that while Trump voters were not
• Over the course of the 2020 cycle, voters were inundated with
polling calls and texts – particularly in competitive states and
“shy,” they also did not answer surveys and were consequently districts – and declining response rates contributed to growing
massively undercounted concerns of how representative the survey was, especially in
regions with large concentrations of non-college educated white
voters
• These declining response rates further exacerbated the issues of
non-response bias, which led some pollsters into missing
Polling misfires in 2020 laid bare a challenge Republican support (Source: FiveThirtyEight)
that the survey research industry has feared for
decades: there is a systematic difference • In recent months, pollsters and academics have theorized that
the COVID-19 pandemic led Democrats to be overrepresented in
between people who take surveys and those some polls – because Democrats are more likely to work
who do not remotely than Republicans and their anti-Trump energy made
them more likely to answer surveys – further exacerbating non-
response bias issues

40
 Our Findings | Polling was a huge problem

Latino voters are not a monolithic group and Challenges remain in the future of polling,
should comprise a larger proportion of polling despite the adjustments made after 2016
samples
• In 2020, despite the efforts undertaken by pollsters to ensure the
• One of the most surprising findings of the 2020 election was proper educational composition of the electorate (particularly in
states and regions with high concentrations of non-college
Democrats’ underperformance with Latino voters relative to 2016
levels educated whites), these methodological adjustments proved
• While the magnitude of the shift was beyond expectations, it insufficient to correct the continued underestimation of Trump
and Republican support
was hinted at in pre-election polls which showed Trump faring
better among Latino voters than he did four years prior (Source:
Matt Barretto, Latino Decisions)
• Although Latino voters as a whole tend to be more Democratic
than Republican, they don’t vote as a single bloc and should no
longer be targeted under this lens, particularly in pre-election
polls – how Latinos vote in Florida can be very different from
how Latinos in the Rio Grande Valley or on the West Coast vote,
and Democratic support among Latinos can greatly vary by
country of ethnicity Polling errors made some campaigns run as if
they were winning – which led to more risk
• Many polls conducted in the 2020 cycle did not take these
factors into account, nor did they ensure that Latino voters aversion and less responsiveness to changing
comprised a large enough share in their polling samples circumstances.
• These misfires in polling led campaigns and organizations to
overlook gaps and drops in Latino support

41
 Our Findings | Polling was a huge problem

In VA-05, polling showed a consistently tight race for Webb since early
October and significantly underestimated his final margin

Ultimately, Good won by 6 points despite multiple public polls that showed Webb with a slight lead – a ten-point swing overall

VA-05: 2020 Congressional Support Among Likely Voters

Webb Good

(Margin) (-6) (-1) (+3) (+2) (+3) (-6)

53%

48%
47% 47%
46%
45%

47%
46%
45%
42% 43%
42%

8/3 9/14 10/1 10/8 10/22 Final

Election
Result

42
4. A Unique 2020 Challenge:
COVID-19 affected everything

43
 Our Findings | COVID-19 affected everything

COVID-19 COVID-19 dramatically transformed campaign strategy and the

suite of tactics employed by both Senate and Congressional
campaigns

affected ● The decision to stop canvassing and halt most in-person voter contact

everything
activities was weathered well by campaigns that were innovative – making
the pivot to virtual tactics quickly and later finding ways to be physically
present in their districts with socially-distanced events
● Campaigns that were unable to rely on virtual campaigning to reach
voters – either due to broadband constraints, local expectations about in-
person outreach, a dearth of creativity, or a combination of factors –
concluded the lack of canvassing was a critical factor in their loss or
tighter-than-expected margin

● Vote-by-mail education and ballot chase became an exponentially larger

budget priority for independent efforts in particular, with some
organizations investing tens of millions of dollars in these efforts

44
 Our Findings | COVID-19 affected everything

COVID changed everything about health care Many Democrats missed the opportunity to talk
conversations about COVID through an economic lens as the
cycle progressed
• As soon as coronavirus descended upon the U.S. and became the
dominant issue of 2020, candidates’ messaging and campaign
activities focused on addressing the immediate health crisis and
• As the cycle continued into the fall, many Democrats did not
pivot along with public discourse – when they spoke about
the collapse of local healthcare systems as infections surged COVID, it was still in the context of access to PPE, wearing masks,
around the country and trusting science, without further connecting those issues to
• Of 740 digital ads that had the highest impressions for both the critical necessity of re-opening the economy
Democratic and GOP candidates, 12% mention COVID, again
mostly run by Democratic candidates
• And some Democratic health coverage messaging that worked
well in 2018 fell flat despite the pandemic: “People aren’t worried
• Among all references to COVID-19 in the digital ads assessed, the about health insurance when they don’t have jobs” according to a
vast majority (70%) were in ads run by Democrats former national Party official
• Candidates hosted PPE give-away events and shared the latest
information about the pandemic featuring local doctors and
• In some places, the GOP successfully framed Democrats as out of
touch on the economy because they prioritized messages about
nurses staying home while most working-class jobs can’t be done
• But as school and job closures dragged on, much of the public remotely – especially in districts with larger Latino populations
narrative around COVID shifted to the economic impact of the o In Debbie Mucarsel-Powell’s district (FL-26), “Voters didn't
pandemic: When would schools re-open? When would workers want to hear ‘stay home,’ they wanted to hear ‘when am I
who couldn’t work from home be able to go back to work? getting back to work.’”
• Sen. Ben Ray Lujan’s campaign attributes some of their success
to emphasizing the economic impact of COVID, efforts to get PPP
loans out the door to help businesses keep their doors open, and
their plans to get constituents back to work

45
 Our Findings | COVID-19 affected everything

COVID-19 changed the way Democratic COVID-19 changed the way Democratic
campaigns and organizations reached voters campaigns raised money – potentially for the
foreseeable future
• While most Republicans quickly resumed canvassing and in-
person events after a universal pause in March of 2020, most
Democrats never returned to canvassing or in-person events
• Campaigns raised huge sums of money through Zoom fundraisers
that involved all kinds of new donors traditionally left out of this
• Relational organizing and remote voter contact tactics absorbed
the significant volunteer capacity of this cycle
critical aspect of campaigning – both from the perspective of
hosting and attending fundraising events – with young people
• Despite COVID-19, many of the best-performing candidates found and less-affluent voters taking part in small-dollar gatherings that
more creative ways to safely get back in front of voters in the provided added value to the campaigns because they did not
field – from socially-distanced live gatherings held outdoors in require the candidate to travel
town squares and parking lots to drive-thru charity events
• Related, candidates raised money from all across the country
• While nearly every campaign, organization, and political
professional acknowledged that the decision not to canvass was
without incurring additional financial cost to the campaign or
burning through every campaign’s most valuable resource: the
the right call given what we knew about coronavirus at the time, candidate’s time
many of those included in this analysis said in hindsight
Democrats could have been on doors safely in more places – and
• Candidates and campaigns universally expressed excitement
about the possibility of continuing these virtual fundraising events
Democrats would have won several close races if they had gone post-pandemic
back on the doors either sooner or in a more robust way than
they did
It’s impossible to evaluate the
2020 election cycle without
acknowledging how COVID-19
dramatically impacted every
aspect of every campaign, up and
down the ballot

46
5. Year-round organizing worked,
as did cross-Party collaboration

47
 Our Findings | Year-round organizing and cross-Party collaboration worked

Long-term Races that did not have the benefit of longer-term

infrastructure investments – voter registration, continual in-
person organizing and outreach – suffered most

investment ● Where candidates had either started voter registration before the

produced
pandemic or where state-based groups had invested long term in bringing
more people into the democratic process Democrats either won or lost by
narrow margins

better results ● Consistent relationship-building through year-round organizing made a

difference – especially in communities of color, whether the organizers
were Democrats (as was the case in Georgia) or Republicans (as was true
of The LIBRE Initiative in Texas and Florida)
● As one longtime Democratic operative described organizing voters of
color, “It’s not a side deal, it’s got to be core”

48
 Our Findings | Year-round organizing and cross-Party collaboration worked

Georgia saw record setting turnout in both the presidential and runoff
elections – nearly 5M in the general and 4.5M in the run-offs
In 2020, Georgia turnout far surpassed the 4 million who turned out in 2016, and a big factor in this increased turnout was a
substantial increase in registration – 2 million voters were added to the voter file between 2016 and 2020

Georgia: Total new reg by month since 2016 presidential

MODELED DEM MODELED GOP

2018 GUB. CYCLE 2020 PRES. CYCLE

11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12
2016 2017 2018 2019 2020
49
 Our Findings | Year-round organizing and cross-Party collaboration worked

The Party Democratic coordination in the face of a common national foil –

Donald Trump – led to extraordinary collaboration across the
Party ecosystem

coordinated ● Leading Democratic donors and funders took part in collaborative tables

better than
in 2020 organized around supporting on-the-ground work in battleground
states and districts

● Similarly, funders proactively organized others in the space to resource

ever anticipated voter education and voter protection needs – the “Count
Every Vote” movement – in the face of threats to the democratic process
telegraphed in advance by the Trump campaign organization
● The 2020 cycle witnessed the first successful data sharing operation
across organizing efforts on the hard and soft sides, with campaigns and
external organizations consistently touting the value of the Democratic
Data Exchange (DDx) in generating efficient voter contact universes for
persuasion, mobilization and ballot chase, and voter education

50
6. Our hopes for 2020 were
just too high

51
 Our Findings | Our hopes for 2020 were just too high

Our hopes for We heard that Dem expectations for pickups in the Senate and
House – and for retaining seats won during the “blue wave” of
2018 – were too optimistic given historical trends for

2020 were presidential cycles

• Inaccurate polling compounded these assumptions and broadened an

just too high

already expansive map

● Our extrapolations about the behavior of first-time voters incorrectly

assumed Democratic support at proportionate rates to regular voters based
on their demographics

The map assumed ● The key challenge of 2020 – the coronavirus pandemic and the subsequent
shutdown of American life – severely limited the toolkit available to
the midterm Democrats, with the constraints on in-person events and canvassing
electorate was harming first-time candidates the most

predictive for 2020 ● We heard from campaign strategists that the Party drove up expectations
by publicly signaling it expected to win everywhere it made a play

But our miscalculation likely benefited us as much as it hurt us

• The conventional wisdom that high turnout was a positive sign for
Democrats informed GOP strategy as well – so Republicans put resources
and energy into races that appeared close in polling but ultimately did not
require their investment

52
“Our anchoring assumption is that we
should’ve just kept everything that we got in
2018 and that was reinforced by the polls.
But if you go down deeper at what seats we
lost, we’re basically losing seats we shouldn’t
have had.”
– Veteran national political strategist

53
 Our Findings | Our hopes for 2020 were just too high

o Yet despite high turnout and enthusiasm among the

Extraordinary turnout driven by increased access Democratic base, Cisneros underperformed Biden across all
to VBM ballots and excitement about the racial groups, particularly AAPI voters who grew their overall
Presidential election benefited Republicans share of the electorate by 6 points since 2016

• In NM-02, a swell in Republican voter turnout in 2020 led to a

Republican victory and the loss of Frontline Member Xochitl
Torres Small
o Turnout in NM-02 jumped 9 points with almost 89,000 more
votes cast in 2020 than 2016 – however, it was the
Republican base that showed higher levels of enthusiasm
• While personal scandal was likely a deciding factor for NC Senate
candidate Cal Cunningham given his steady strong performance in
internal and public polling ahead of the news of his extramarital
affair, increased Republican turnout in red areas far exceeded
expectations and eclipsed the margins the campaign anticipated The 2020 elections made clear that turnout
in its pathway to victory efforts don’t just benefit Democrats
o In the NC Senate race, although Democrats gained with
White urban voters, GOP improvement among rural and
exurban Black voters gave Democrats a narrow edge but not
enough to flip the state
• Former Rep. Gil Cisneros (CA-39) reported that because of the
demographic makeup of the district, his team focused significant
resources on turning out the voters of color who had been key to
his 2018 victory – particularly AAPI and Latino/Hispanic
communities

54
“We always think that if turnout is up, that’s
going to be beneficial for us. That’s not the
case, Republicans vote too and they’re doing
a very good job at it. So, when turnout is up,
we need to ask who instead of just being
excited about it.”
– NC political operative

55
 Our Findings | Our hopes for 2020 were just too high

In 2020, two-thirds of the voting eligible population voted – more than in any
other election in 120 years

Turnout was especially high in many battleground states and districts and grew by an average of 27% from 2016
across the Tier 1 races analyzed. Both parties benefited from this explosive growth in turnout.

Percent Increase in Votes Cast, by Party*

2016 – 2020
DEM GOP

63% 61% 62%

59%
55% 56% 55% 56%
52% 50% 52% 51%
49%

51% 50% 49%

48% 48%
46% 44% 44% 45% 44%
39% 41%
37% 38%
21%

Win/
AZ CA-39 FL-26 GA IL-14 ME-02 MI-08 NC NE -02 NJ -03 NM-02 TX-23 TX-24 VA-07
Loss

*Based on modeled party in most states

56
 Our Findings | Our hopes for 2020 were just too high

o Frontline incumbents expressed a range of feelings towards

Campaigns, strategists, and pollsters alike said
the DCCC – some felt it wanted too much influence in on-
polling blind-spots and methodological the-ground decision-making, while other freshman
difficulties led to an overly optimistic map Members expressed the desire for more support and
guidance from the DCCC
• Polling difficulties, especially in rural districts where many
Frontline and Red to Blue candidates are based, made these
districts look more likely to flip in 2020 than the post-cycle
data show
o Election forecasts predicted the TX-23 race to be in Gina
Ortiz Jones’ favor – with FiveThirtyEight forecasting a 6-
point advantage for the Democratic candidate
• These polling difficulties filtered their way down to the Red to
Blue list and shifted Party resources to races that were
ultimately unsuccessful, some by large margins
• Some strategists and campaigns reported concern that the
Party publicly telegraphed a significant focus on expanding the
House majority instead of positioning the election as an effort
to protect Frontline members, including those seats gained in
2018
o While the DCCC spent $3 million more on the Frontline
program than the Red to Blue program – investing early in
these districts and adding field organizing staff in spring of
2019 – several Frontline candidates and their staff
expressed their belief that protecting incumbents was not a
top priority for the DCCC

57
 Our Findings | Our hopes for 2020 were just too high

The expansive map and inaccurate polling also

hurt Republicans, who put money in non-
competitive races and diverted resources and
attention from races they could have won

• With such a large battleground map, campaigns and strategists

indicated that Republicans “followed us” to districts like TX-23
and TX-24 that the GOP already held, instead of investing in
other, more traditionally vulnerable areas for Democrats, like
Michigan House races where they might have picked up seats
• Given heavy resourcing for these unexpected campaigns
reported in numerous media outlets from late summer into the
fall, Republicans spent less elsewhere, allowing a number of
vulnerable Democratic Frontline incumbents to win close races, The map had important implications for
including Rep. Abigail Spanberger (VA-07), Jared Golden (ME- fundraising, staffing, and messaging that
02), and Rep. Elissa Slotkin (MI-08) ultimately shifted the outcomes in several key
• Even in NY-19, where Rep. Antonio Delgado ultimately won by
nearly 12 points, lack of Republican attention was a significant
races across the country. While the Party
secured the Presidency, we lost Frontline House
factor in a district with demographics that could have made the seats and were unable to close the gap in
Congressman vulnerable several high-profile Senate campaigns that were
o As one Party executive said, “A can of coke with an R next
widely perceived to be pick up opportunities.
to its name could have won against Delgado,” but rosy
predictions about Democratic gains in 2020 kept
Republicans from investing in the race or recruiting a
strong challenger

58
Conclusion
Parting thoughts from the authors on the 2020 cycle

59
 Conclusion | Parting Thoughts from Marlon and Lynda

A Final Note from the Authors Republicans would have found another racist dog whistle or fear
mongering tool to paint Democrats as out of touch radicals who
Throughout this analysis, one of the things we reflected on most are trying to destroy the America we know and love.”
was the impact and the implications of race in our political
At the highest levels of the Party ecosystem, cultural competency
discourse. During the course of nearly 150 interviews that included
elected and Party leaders, candidates and donors, campaign was a continual concern in 2020. Campaigns shared anecdote after
operatives and leading practitioners across the Democratic anecdote of the most well-intentioned Democratic supporters
displaying a stunning lack of cultural competency in an electoral
ecosystem, the single most-cited challenge apart from the COVID-
19 pandemic was our inability – or unwillingness – to address race cycle where “culture” drove practically every outcome – from the
head-on. suburban white women who berated a Black woman candidate for
not attending the Black Lives Matter protest those women had
The Democratic Party needs to be unapologetic about race. And organized, to political operatives who advised another Black
we need to explain to all Americans why doing so is beneficial to woman candidate not to be overly concerned with responding to
everyone. attacks on her “Defund the Police” stance, to Party Committee
staffers pushing candidates of color to pump their “networks” as if
Those of us who have worked in and around politics know in our the same rolodex-driven approach to fundraising translates in a
bones that what we witnessed in 2020 was not new. Sadly, race non-white, non-privileged world.
baiting and fear mongering have long been part of the political
maneuvering and cultural dynamics of this country. But 2020 We’ve seen the same tactic before.
witnessed a revival of “dog-whistle” politics and an overt racism In 2008 and 2012, the messages were centered
that increasingly flourished under the candidacy and presidency of on government takeover of health care and class
Donald Trump. And where voters probably regularly default to warfare. In 2016, the message was “Build the
voting based on culture over policy issues, 2020 whipped up Wall,” which morphed into claims in 2018 that all
stereotypes and caricatures to a level we haven’t seen in recent
Democrats wanted to “Abolish ICE” and “open
electoral cycles and undoubtedly had an impact on voting
borders” for caravans of migrants.
decisions around the country. As Dr. Cameron Webb’s campaign
“Defund the Police” was 2020’s dog whistle.
manager, Ben Young, put it, “If it wasn’t ‘Defund the Police,’

60
 Conclusion | Parting Thoughts from Marlon and Lynda

The good news is that the 2020 experience has shined a spotlight communicate across these many-splendored communities.
on areas the Democratic Party and Democratic leaders can make
gains in the future – from defining and sharing a vision on race and Commit to early investment and year-round organizing. Victories
racial justice to grappling with the Democratic brand to shoring up in Georgia, Nevada, Arizona, and other states this cycle were built
our small “d” democratic institutions for the future. At a minimum, on years of organizing and community servicing led by some of our
this past cycle should make clear the urgency around combatting best political and labor organizers. This orientation is especially
GOP voter suppression efforts, so often a thinly veiled attempt to critical in communities of color that we know value relationships
exclude communities of color and in 2021 an open backlash to but whom we often treat in a transactional way, showing up every
expanded early vote and vote-by-mail that made it easier for more 2-4 years instead of being continually present and engaged. If
voters than ever before to make their voices heard. Democrats fail to organize communities of color – making the
continual case for how we’re standing up for justice, jobs, health
The opportunity coming off of 2020 is in how we deal with and care, a quality education and more – we can be sure Republicans
define race issues in the future. Here are some ideas: will fill that void.

Reimagine our Democratic Party message and narrative. The Reallocate our spending to reflect our commitment to the
Republican Party has long had a collective gospel about having communities we rely on to win. It’s not just about investing in key
small government, low taxes, and a strong military. Our gospel communities for Democrats – it’s also about ensuring a more
should be about championing all working people – including but equitable allocation of our resources overall. As one longtime
not limited to white working people – and lifting up our values of Latino elected leader put it, “It’s great that we spent more in 2020
opportunity, equity, inclusion. We need to make it exciting to be a on Latino voters than ever before. But what did we spend before?”
Democrat and to ensure voters believe we not only mean well, but Or as Leslie Small, the Executive Director of America Votes in
we understand their priorities and they trust we are actually Georgia told us, “We have, for years as the Democratic Party, have
fighting for them – something that the series of focus groups with treated these mythical white voters as the panacea, and so if you
swing voters recently conducted by Third Way shows continues to have ten dollars, they’re going to give nine and a half to try and get
be a challenge. The only way to do this well is for the Party to do this white voter...For once, Georgia and voters of color were not
more, better research with voters of color and, based on what we under invested in. And by investing in them at the level that they
learn, to be more differentiated and targeted in how we should be invested in, you saw the results.”

61
 Conclusion | Parting Thoughts from Marlon and Lynda

End the “persuasion vs. organizing” debate and focus on getting In 2021, we are more aware than ever of the disparate challenges
votes, period. Campaigns typically think persuasion needs to start for communities of color that are continuing to play out in the
early and mobilization should happen right around voting time. But aftermath of the Trump era – including the rising violence against
that approach doesn’t work, and it isn’t helpful for long term Party- the AAPI community that is impossible to separate from the former
building. When it comes to persuasion and turnout, it’s “both-and” President’s insistence on calling COVID-19 the “China virus,” “Kung-
but needs to happen all the time. flu,” or the “Wuhan virus.” Despite the hope sprung by the verdict
in the Derek Chauvin trial, Black men and women in every state are
Too often campaigns view persuasion as a tactic to get white continuing to lose their lives to police violence and racial injustice
Americans who may not be with us on our side, and turnout as a that is far too often forgotten when today’s headlines line
tactic to get our base out (which implicitly means people of color). tomorrow’s waste baskets. And Latino and Hispanic people remain
We should be trying to persuade all voters to support the ideas we a political football, alternatively portrayed as an electoral prize or
care about – and that means engaging year-round, especially with the infamous “Other” from which American communities must be
communities of color. shielded.

Where we do need to focus on persuasion, it’s to make the The question is, what are we as a Party willing to do about it.
unequivocal case that ending systemic racism is good for
everyone. Right now, people think addressing racial injustice means
they are losing something, or that someone is taking something “Diversity for diversity’s sake is not the goal.
away from them. Last year, when the conversation became about It’s about the country being a better country,
“Defund the Police,” we were stuck on defense instead of telling a about the government being a better
proactive story about necessary systemic changes to policing that government.”
would stem the violence and still prioritize and provide public – Quentin James, Founder and President, The
safety. There’s a way for a “Defund the Police” advocate and Collective PAC
someone who isn’t convinced we should divert police funding to
agree to disagree but to lift up the value of what everyone needs.
We need to be on offense on what we care about, why this is
important, and what we should be doing differently.

62
Appendices

63
Process & Approach

64
 Appendix | Process & Approach

Our approach
1. Live interviews with candidates and staff, Members, Party leaders, external
organizations that were major players in the 2020 cycle, major funders, and
top strategists and vendors

2. Review of publicly available polling and proprietary polling where possible

3. Content analysis of TV ads for priority races aired by the campaigns and
independent expenditure efforts and broader online paid media spending
by the campaigns

This study relied on 4. Independent assessment of the data analytics and modeling using the final
four categories of voter files released by the states in early 2021

analysis

65
 Appendix | Process & Approach

TIER 1
Tier 1 campaigns were selected by Third Way, Latino Victory, and
the Collective PAC to reflect a broadly representative set of the
most competitive races in the 2020 cycle, with candidates, District,
and campaign dynamics designed to provide a better understanding
of what happened across the political landscape. The groups aimed
for regional and demographic representation, Districts with
anticipated voter performance and electoral issues that were shared
with other races beyond this list, and politically-significant
campaigns with a cross-section of both wins and losses.

1. AZ Sen - Mark Kelly 11. NJ-03 - Andy Kim

2. GA Sen - Raphael Warnock 12. NM-02 - Xochitl Torres Small
3. GA Sen - Jon Ossoff 13. NY-02 - Jackie Gordon
4. NC Sen - Cal Cunningham 14. NY-19 - Antonio Delgado
5. CA-39 - Gil Cisneros 15. NY-24 - Dana Balter
6. FL-26 - Debbie Mucarsel-Powell 16. TX-23 - Gina Ortiz Jones
7. IL-14 - Lauren Underwood 17. TX-24 - Candace Valenzuela
8. ME-02 - Jared Golden 18. VA-05 - Cameron Webb
9. MI-08 - Elissa Slotkin 19. VA-07 - Abigail Spanberger
10. NE-02 - Kara Eastman

66
 Appendix | Process & Approach

TIER 2
For Tier 2 races, this analysis relied on a broader survey of the
publicly-identified DCCC Frontline races and the DSCC Priority races
to gather additional information and capture common themes across
the Party Committees’ highest priority campaigns. In addition to the
multi-layered research for Tier 1 races, this postmortem includes
findings based on information received from another 18 races. Their
responses covered messaging, strategy, tactics, and the candidate’s
personal observations about the races.

1. AK Sen - Alan Gross 11. IN-05 - Christina Hale

2. NM Sen – Ben Ray Lujan 12. MN-02 - Rep. Angie Craig
3. CA-10 - Rep. Josh Harder 13. NC-08 - Patricia Timmons-Goodson
4. CA-21 - Rep. TJ Cox 14. NJ-02 - Amy Kennedy
5. CA-25 - Christy Smith 15. NJ-07 - Rep. Tom Malinowski
6. CA-48 - Rep. Harley Rouda 16. NY-22 - Rep. Anthony Brindisi
7. FL-15 - Alan Cohn 17. OH-01 - Kate Schroder
8. GA-06 - Rep. Lucy McBath 18. UT-04 -Rep. Ben McAdams
9. IA-03 - Rep. Cindy Axne
10. IL-06 - Rep. Sean Casten

67
 Appendix | Process & Approach

EquisLabs, Catalist, and Priorities USA. Where the findings in this

The broader Democratic ecosystem postmortem used information from other research or studies that
The team spoke extensively with other Members of Congress across were conducted post-election, those sources are directly cited.
the spectrum of Democratic ideology from some of the Party’s most
progressive voices to “pragmatic moderate” leaders – either 1:1 or as
part of listening sessions organized in collaboration with
Congressional Caucus staff. We interviewed members of the
leadership team at the Party committees, leaders at key outside
organizations that ran 8- and 9-figure independent programs in
2020, state organizations that were major players, and leading
practitioners across the Democratic political space – including
pollsters, data scientists, digital strategists, and message and media
experts. In total, this analysis involved interviews with 143
individuals. The Members from the New Democrat Coalition, the
Congressional Black Caucus, the Congressional Hispanic Caucus,
and the Congressional Asian Pacific American Caucus and staff of
the CBCPAC, CHC BOLD PAC, New Dem Action Fund, and ASPIRE
PAC were especially instrumental in helping to shape this study,
connect the research team to candidates and their campaign staff,
and provided important data and feedback about what they saw in
the field.

Additional informative research

Finally, the team reviewed major postmortem analyses by other
organizations as they were released into the public domain –
including the Texas Democratic Party, Higher Ground Labs,

68
Unprecedented Ad Spending

69
 Appendix | Unprecedented Ad Spending

Spending and raising money early created significant benefits for many
candidates

• Inpolitical
Michigan, one longtime statewide
operative posited that it was
difficult for Republicans to recruit strong
Off-the-charts fundraising didn’t always ensure blowout victories – or victory
candidates to run against Rep. Elissa at all – even if Democrats substantially outraised their opponents
Slotkin (MI-08) or Rep. Haley Stephens
(MI-11) because these incumbents had Dem Candidate GOP Candidate
already amassed formidable war chests
early in the cycle 2020 Total Cash on Total Cash on
Race
• Multiple candidates reported their ability Outcome Receipts Hand Receipts Hand
to go up on TV early in the cycle to
highlight their bios and frame the VA-07 D+1.8 $8,494,948 $582,133 $3,670,266 $29,323
narrative was hugely valuable
• Strong campaign cash allowed Rep. IL-14 D+1.4 $7,862,425 $537,898 $3,297,857 $90,059
Lauren Underwood (IL-14) to hire field
organizers early in the cycle and begin
moving a locally-focused program led by
both in-district and out-of-district staff
who developed a deep knowledge of
their turf

70
 Appendix | Unprecedented Ad Spending

Senate Campaign + IE Ad Spend

Across the ecosystem – including candidate, coordinated, and issue groups - Democrats outspent Republicans in the Senate

Senate Ad Spend, Candidate + Coordinated +

Issue Group
(in millions)

DEM GOP

+ $41M + $40M + $112M

$182
$153

$112 $114

$74 $70

Win/Loss NC SENATE AZ SENATE GA SENATE

71
 Appendix | Unprecedented Ad Spending

Congressional Campaign + IE Ad Spend

The following information regarding ad spend was available for the Tier 1 races [source: AdImpact report]
Dems outspent Republicans in all races analyzed except for NY-24, where the GOP invested just as heavily as Dems

House Ad Spend, Candidate + Coordinated + Issue Groups

(in millions)

DEM GOP

+$2M +$3M +$2M +$4M +$5M +$4M - +$4M +$3M +$6M +$4M

$12

$10 $9
$7
$6
$7
$2
$2 $16
$3 $2
$12 $12 $11
$10
$7 $7 $8
$5 $6
$4

Win/Loss FL-26 IL-14 ME-02 MI-08 NJ-03 NM-02 NY-24 TX-23 TX-24 VA-05 VA-07

72
 Appendix | Unprecedented Ad Spending

o Senator Raphael Warnock outspent Kelly Loeffler by over a

Digital ad spending increased significantly in 2020,
4:1 ratio, with over $4 million in spending to Loeffler’s
with Democrats typically outspending their
$814,697, with no candidate spending outside of the
Republican opponents – but a higher spend did not
2020/2021 cycle
always mean Democratic victory o Rep. Lauren Underwood outspent her Republican opponent
by a 14:1 ratio, spending over $500,000 to Jim Oberweis’s
• In the 2016 election cycle, 2-3% of the total political ad spend
$35,318
across the coordinated and independent expenditure efforts up
and down the ballot went to digital media – in 2020, that spend • Although there are cases where Democrats were outspent online
o Gil Cisneros (CA-39) was outspent by Republican opponent
moved up to 18%. Approximately $700-800M was spent on digital
Young Kim 2:1 on Facebook platforms, with Rep. Kim
ads in the 2018 cycle, while in 2020, the spend was $1.6B [Forbes].
spending $834,337 to Cisneros’ $492,290
The 2020 election saw digital ad spending on Facebook reach
o Dana Balter was outspent by her opponent $157,612 to
$1.07B and $520M on Google [AdImpact report]
$136,569
• The Democratic candidates we evaluated ran hundreds – and in
some cases thousands – of digital ads ranging in spend from less • Several Democrats who outspent their opponent online did not
ultimately win – including Xochitl Torres Small, NM-02 and Dr.
than $100 per ad, up to $250,000 on a single ad [Warnock]
Cameron Webb, VA-05 – but nearly every Democrat who was
o Every Democrat included in this review invested a minimum
outspent in digital ads lost
of $125,000 on Facebook advertisements alone
o Although it is impossible to compare that figure to previous
cycles since 2020 was the first time Facebook and Google
have made full access to political advertising data available,
digital spending widely increased in 2020
• Many Democratic candidates outspent their Republican
opponents in digital advertising in the House and Senate
o Senator Jon Ossoff’s campaign Facebook page ran more
than $2.7 million in ads to David Perdue’s $537,686 in the
2020 cycle

73