Data, Ta-Da!
By Craig Johnson (@CraigJPolitics) and Alan Rosenblatt (@DrDigiPol), Unfiltered.Media
One of Craig’s favorite movies is The Big Short. He likes it because does a great job explaining in excruciating detail not only just how badly Americans got screwed by big financial institutions in the 2007 housing crisis, but also how there were experts who knew there was a problem and just watched the world burn–so to speak. That is what we are afraid that we are facing today. What we saw back in 2016 represented just a taste of what big data could do to distort an election by micro-targeting people based on thousands of variables.
What we are looking at today, in 2022, is a non-existent data privacy landscape that sees our wealthy political rivals, such as Peter Theil, in control of industry leading companies that connect every single data set (complete with your personal political, social, and economic information) to Facebook and display ad targeting, to text messaging and OTT video ad targeting, and connecting it all to how and where you spend your money. And while our side has Target Smart and Catalyst writing contracts that prohibit combining their data sets to make a better one, the other side has weaponized machine learning, A.I. and the utilization of B-to-B and B-to-C commercial databases to create a fully integrated data modeling monster. (On a related and, perhaps, terrifying note, Google laid off an employee that claims the company created a sentient AI.)
This is a “Houston, we have a problem” level issue. We have nothing as sophisticated on our side while we face the largest threats to democracy to-date. The problem is monumentally clear that, as a coalition, we need a quality repository of easily usable data synchronized from many different sources to compete with what the other side has at its disposal, especially since the building of Palantir, the data mining and content analysis platform that the Department of Defense uses..
We currently have siloed data companies and some attempts at combining data like The Movement Cooperative’s CivicTech project. But while Civic Tech is doing interesting things with combining publically available data with recently acquired LiveRamp style data sources it does not go to the same level of interconnectedness, content analysis, and advanced modeling as Palantir, and every moment we spend with data vendors refusing to cooperate and allow, nay facilitate, the integration of our data, is another moment we fall behind the data race.
Instead of 20 organizations with 20 different data connections, we need a new body that has the buy-in from everyone, or at least everyone necessary. This new organization should be devoted to taking in as much data from as many sources as possible, transparently synchronizing and developing user interfaces to connect the data to a closed community of progressive and Democratic organizations and campaigns.
The key is to go beyond recreating what many commercial ad vendors offer and make the data available in a format that is usable for the average state-based advocacy group operating in the field, as well as providing the underlying data to larger organizations with the internal capacity to work with raw data.
It is vital to run this project as a non-profit, removing the profit incentives that lead to most of the failures we currently see in data quality, accessibility, and integration. For-profit data business models inevitably prioritize volume over quality. And the wholesale approach to data often means the data vendors do not understand our strategic and tactical uses of their product.
If we are able to set up a data sharing collective that was NOT for profit, we can avoid the incentives that drive for profit data businesses to provide lower quality data not optimized for the uses we need of them. This collective data sync would require staffing with specialized talent and skills, including working with machine learning, database operation and maintenance, data analysts, voting behavior analysis, etc. to ensure the data is clean, useful, and we are learning from it.
We need these specialized staff to start connecting commercial data with voter files and other political data. Unless there are secret campaigns out there with secret infrastructure, no one practicing mainstream politics knows about, this isn’t happening.
If we do not up our data game we will continue to be behind in crucial digital fundraising, organizing, and advertising tactics and we will also fall behind in our traditional organizing edge. If Republicans can more accurately target the Rio Grande valley in Texas using their superior data systems to target voters digitally AND on the ground (i.e. generating walk maps), for example, our ability to persuade and to GOTV will get ever harder as their targeting improves, recruiting gets better, and they more efficiently spend every bit of their resources.