Though some current clients say they believe Cambridge Analytica's core data product is of a high caliber and can be applied to ad targeting to produce better results than other types of
voter data models, others who have met with or worked with the firm say they were not impressed with Cambridge's data product, and called it too expensive.
In Advertising Age, a political client said the embedded Cambridge staff was «like an extra wheel,» but found their core product, Cambridge's
voter data modeling, still «excellent.»
Without at least some outreach via channels with a «fuzz factor» to catch
voters your data model missed, you may find yourself with a dangerously open flank.
«But the core product is still excellent,» he stressed, speaking of Cambridge's
voter data modeling.
Not exact matches
Some news accounts indicate that his campaign stopped using the firm's
data after the South Carolina primary in late February 2016, though federal campaign records show more than $ 670,000 in payments to the firm for «media /
voter modeling» or «
voter ID targeting / web service» in March and June, plus $ 218,000 for «media» and «digital service / web service.»
Through utilizing their vast troves of existing
data on individual
voters, constructing highly advanced
data models, and prioritizing
voters by their likelihood to vote and feelings of favorability towards each candidate, Cambridge Analytica created a unique «principal audience» of
voters to target.
In doing this we used a suite of
models produced by the
data science team, which outlined profiles such as undecided
voters or inactive supporters, and matched these audiences to online cookies, mobile devices, and social IDs.
The project is detailed in the contract as a seven step process — with Kogan's company, GSR, generating an initial seed sample (though it does not specify how large this is here) using «online panels»; analyzing this seed training
data using its own «psychometric inventories» to try to determine personality categories; the next step is Kogan's personality quiz app being deployed on Facebook to gather the full dataset from respondents and also to scrape a subset of
data from their Facebook friends (here it notes: «upon consent of the respondent, the GS Technology scrapes and retains the respondent's Facebook profile and a quantity of
data on that respondent's Facebook friends»); step 4 involves the psychometric
data from the seed sample, plus the Facebook profile
data and friend
data all being run through proprietary
modeling algorithms — which the contract specifies are based on using Facebook likes to predict personality scores, with the stated aim of predicting the «psychological, dispositional and / or attitudinal facets of each Facebook record»; this then generates a series of scores per Facebook profile; step 6 is to match these psychometrically scored profiles with
voter record
data held by SCL — with the goal of matching (and thus scoring) at least 2M
voter records for targeting
voters across the 11 states; the final step is for matched records to be returned to SCL, which would then be in a position to craft messages to
voters based on their
modeled psychometric scores.
Data Science Leveraging data science and predictive analytics expertise, we built 20 custom data models that could be used to forecast voter behav
Data Science Leveraging
data science and predictive analytics expertise, we built 20 custom data models that could be used to forecast voter behav
data science and predictive analytics expertise, we built 20 custom
data models that could be used to forecast voter behav
data models that could be used to forecast
voter behavior.
That's one reason «addressable» advertising has become popular in the past five years, particularly at the presidential and statewide levels, where campaigns more often have the capacity to build the
data models to identify demographic groups and specific
voters to persuade or turn out.
What
voters actually do matters more than what you think they MIGHT do:
data derived directly from
voters» choices quickly supersedes
models that try to predict their behavior.
Of course, these tools will only help if you've first done the work of 1) building a sizable Facebook following, and 2) creating a list or
data model of the
voters you need to reach.
Ted Cruz and Marco Rubio chose opposite strategies in the weeks leading up to the early primaries, with Cruz basing his campaign on
data analysis and
voter modeling.
As you gather these
data, you feed them back into the
voter file and — here's the key part — adjust your
models (and hence your outreach targeting) accordingly.
Commercial marketers usually have a wealth of
data to work with, from demographics to credit history to homeownership, but when he started working in politics, Ghani was struck by the fact that political campaigns are trying to build
voter models based on a handful of
data points.
This
data can also test the
model that I proposed to determine the
voters who will show up in the race to fill George Latimer's senate seat.
In that case, they're likely throw money at poorly targeted TV campaigns rather than spend time and resources building up a robust field operation or investing in
data -
modeling and
voter targeting.
A key element in the Democratic win in Virginia this week was the ability to adapt Obama campaign capabilities to integrate historic
voter file
data and recent
data from field organizers and partner groups with analytics capabilities to develop targeting
models, representatives from NGP VAN, Blue Labs and the Virginia Democratic Party said in a press call Friday.
The idea behind MRP is that we use the poll
data from the preceding seven days to estimate a
model relating interview date, constituency,
voter demographics, past voting behaviour, and other respondent profile variables to their current voting intentions.
Based on
voter data and predictive
modeling, the campaign's 100 volunteers have made over 267,000
voter contact calls, and have knocked on over 14,400 doors, Alcivar said.
Top Trump campaign officials, meanwhile, played down the work of the
data - science company, which was paid at least $ 6 million to do
voter modeling and ad buys for Trump in the 2016 general election.
No one understands the transformative power of
data analysis better than Democratic consultant Ken Strasma, who helped propel Barack Obama into office by devising a mathematical
model that predicts the political behavior of nearly every eligible
voter.
As The New York Times reported on Saturday, that is what motivated the consulting firm Cambridge Analytica to collect
data from more than 50 million Facebook users, without their consent, to build its own behavioral
models to target potential
voters in various political campaigns.
Cambridge's
voter data innovations are built from a traditional five - factor
model for gauging personality traits.
«By combining advanced
data analytics with psychological research based off the five factor
model for gauging personality traits, OCEAN, Cambridge Analytica helped the campaign identify likely pro-Cruz caucus
voters and reach out to them with messages tailored to resonate specifically with their personality types,» stated the press release.
Data Science Leveraging data science and predictive analytics expertise, we built 20 custom data models that could be used to forecast voter behav
Data Science Leveraging
data science and predictive analytics expertise, we built 20 custom data models that could be used to forecast voter behav
data science and predictive analytics expertise, we built 20 custom
data models that could be used to forecast voter behav
data models that could be used to forecast
voter behavior.
Some news accounts indicate that his campaign stopped using the firm's
data after the South Carolina primary in late February 2016, though federal campaign records show more than $ 670,000 in payments to the firm for «media /
voter modeling» or «
voter ID targeting / web service» in March and June, plus $ 218,000 for «media» and «digital service / web service.»
Kogan went on to say that the
data model they found wasn't very accurate at an individual level, believing as a result that Cambridge Analytica was selling a myth of sophisticated
voter - targeting methods to political campaigns.
Cambridge Analytica also bragged publicly about its use of
data and
voter modeling during the election, which sparked the interest of David Carroll, a professor at Parsons School of Design who studies media,
data targeting, and campaigns.
As The New York Times reported on Saturday, that is what motivated the consulting firm Cambridge Analytica to collect
data from more than 50 million Facebook users, without their consent, to build its own behavioral
models to target potential
voters in various political campaigns.
The project is detailed in the contract as a seven step process — with Kogan's company, GSR, generating an initial seed sample (though it does not specify how large this is here) using «online panels»; analyzing this seed training
data using its own «psychometric inventories» to try to determine personality categories; the next step is Kogan's personality quiz app being deployed on Facebook to gather the full dataset from respondents and also to scrape a subset of
data from their Facebook friends (here it notes: «upon consent of the respondent, the GS Technology scrapes and retains the respondent's Facebook profile and a quantity of
data on that respondent's Facebook friends»); step 4 involves the psychometric
data from the seed sample, plus the Facebook profile
data and friend
data all being run through proprietary
modeling algorithms — which the contract specifies are based on using Facebook likes to predict personality scores, with the stated aim of predicting the «psychological, dispositional and / or attitudinal facets of each Facebook record»; this then generates a series of scores per Facebook profile; step 6 is to match these psychometrically scored profiles with
voter record
data held by SCL — with the goal of matching (and thus scoring) at least 2M
voter records for targeting
voters across the 11 states; the final step is for matched records to be returned to SCL, which would then be in a position to craft messages to
voters based on their
modeled psychometric scores.
Cambridge Analytica's goal, starting in 2013, was to use
data modeling to influence
voters based on their emotional makeup.
In total, the personal information of potentially near all of America's 200 million registered
voters was exposed, including names, dates of birth, home addresses, phone numbers, and
voter registration details, as well as
data described as «
modeled»
voter ethnicities and religions.