Can we talk about VAN Data?

Maybe the data tools that the Dems have come to rely on are what are doing us in.

After you have had time to re-set your thinking and come to grips that the Blue Wave is not coming in 2024 after all, I’d like you to reflect on my thesis: maybe the data tools that the Dems have come to rely on are what are doing us in. 

You’ve probably heard of VAN

If you’ve spent time canvassing, phone banking, or just hanging around at a democratic headquarters, you have heard the terms, “VAN”, “Mini-VAN”, “NationBuilder”, “Vote Builder”, or some variation on the above. 

Since the first Obama run, when Hillary got out-teched by running a more traditional campaign against Obama’s hip, slick, and cool, data-driven campaign, Democrats have knelt at the altar of screens, sort tools, and so-called “turf cuts”. (For the uninitiated, a “turf cut” is a collection of names and addresses, assembled with a map, and handed to a volunteer for an optimized canvassing experience.)

My ride in the VAN

I spent several weeks and several thousand dollars pursuing the dream that out there, somewhere, was a database that could be the key to winning my seat in New York’s 21st Congressional District. 

The real challenge is that there is no genuine competitor to NGP VAN. On the site, Software Advice, an anonymous user writing in March 2024 said this about reasons for choosing NGP VAN: “There were no products we considered, there are not really serious competitors to the product.” Anonymous had other remarks: “Very positive overall, it maintains all of the lists, data, call, knock, and text data as well as valuable targeting information…The interface continues to be clunky and somewhat difficult to navigate for new users. It functions first as a database as opposed to an active management tool.”

I first encountered this world of Democratic data when I announced my candidacy in February 2024. I talked to a campaign data team that said I “must have” Nation Builder. I created a user account and entered my credit card details. I have no idea where their data came from or what was to be used for. Was this donor data? Door knocking data? Who knows? I exited the relationship (thankfully) with that data team. 

The team that I hired to replace them was able to get me out of the Nation Builder contract. Unfortunately they also got me into a contract with NGP VAN that they told me was necessary for completing required financial reporting. Um. No. The version they got me contracted for doesn’t do that. I needed another tool to do that. And in order to get that tool, I needed to spend another $1300 a month, with a contract until the end of the general election. 

A user named Ace, writing in February 2024 on Software Advice, made remarks that resonated with me. Specifically, Ace found that the “Help Desk was not helpful”. And, as I found, the naming conventions of all of the various iterations of the platform are bewildering. Ultimately, this is what cost my campaign thousands of dollars. And no – once you sign a contract, your campaign is committed to regular billing until the end of the campaign cycle. 

I was left with monthly invoices and no usable tools. 

Compare VAN to a county Board of Elections walking list

The group, Indivisible, made VAN data available to all volunteers across the country. It was an amazing investment. I have no idea what this cost Indivisible. A regular person could input their address, click, and find 10 voters they could go visit.

The weird thing was this: I could do that using the county Board of Elections lists, and they found 10 voters on my street or the next street over whom I could visit. The VAN list through Indivisible had me driving over 10 miles from my house.

Rural Districts

Also on Software Advice, I found that Shelley, in August 2022, summed up the pros and cons of VAN the best: “PROS: As a tool it is far more than any rural county party could do for themselves, so it needs to be used to its fullest (but isn't). CONS:...but when used exclusively by volunteers, we can't get the continuity going. Without everyone using it, the input is sporadic and incomplete. There needs to be better coaching from the state party in how a rural county party can and should use VAN, with regular projects designed FOR RURAL COUNTIES (with county party input on the projects)....then deadlines and followup. Tying the why we do something to the results and analysis of what we learn each go-around (and how to track progress from election cycle to election cycle...in our RED county).”

Thanks, Shelley.

The cost of VAN

I am a Democratic Candidate in a state that everyone thinks is a blue, until they look at a map. 

While it is true, Kamala Harris won New York State, it’s the urban centers that brought in the numbers. Only 16 of the state’s 66 counties voted blue. The remaining 50 counties, including the entirety of my congressional district, voted red. 

Consider that my congressional district, NY-21, spans 15 counties and over 15,000 square miles, taking up an enormous amount of land mass. The district encompasses all 6 million acres of the Adirondack Park, which in itself is big enough to fit Yellowstone, Yosemite, the Grand Canyon, and Glacier National Parks in its boundaries. And that’s not all of NY-21. It stretches from just north of Albany, all the way to Plattsburgh, along the northern border to the St. Lawrence Seaway, and down to the Mohawk Valley. We’re talking huge. See the smaller black-and-white inset map to see how this district dominates New York’s landscape.

I was eager to acquire any data I could to make my race more manageable. The consultants I hired insisted that the state Democratic Party would provide VAN data. In fact, the VAN Help Desk insisted that my state Democratic Party would provide VAN data. Friends and colleagues who had worked on campaigns in other states insisted that my state Democratic Party would provide VAN data.

After numerous inquiries on my state Democratic Party’s website, along with phone calls, and email messages, scattered over a period of 3 weeks, I got my answer: the New York State Democratic Party would provide my enormous congressional campaign with VAN data after I input my credit card details and consented to a charge of $8,500. Thank you very much. 

What did I do? 

I vacillated. I really wanted this data, but I did not want to lose $8,500 of campaign funds on software that I, in my limited exposure, could not figure out and could not access a reasonable help desk support system. 

The real kicker was that I could not even get a demo of VAN in order to see if it would be useful. 

I tried to go “up the ladder”, as we say, to see if someone in the state party could either give me access or could even help me identify whether VAN was useful for my ginormous congressional district. 

A number of my county chairs said that they don’t use VAN. It doesn’t work for the rural counties. 

I moved on. I actually did not think of VAN for several weeks. But as we got into canvassing, and I put together my own walking lists to present to town chairs, many people said, “Oh – I can just cut a turf in VAN.”

There it was again. VAN. Many of the towns and a few of the counties had VAN, but they didn’t know or didn’t communicate it to me until we were way into the canvassing process. By that point I had printed literally thousands of pages of voter lists to wrangle my own canvassing project. And of course, handing a volunteer a stack of paper and a pencil is not nearly as slick as asking them to download a MiniVAN App.

How many trees did I lose on all of those pages that I printed?

But back to VAN. Here it was, in more than a few of the towns and counties where I needed it. What was it? How could it help? 

The real question is: Does it help? 

As I understand it, VAN will take a voter’s history, and in some cases (though I can’t get anyone to give me a straight answer on how or when) they will mix in some amount of consumer data to determine the likelihood of voting for Democratic candidates in the upcoming election. 

Whoa. Stop right there. You’re mixing some undisclosed amount of what you learned about me from my Amazon shopping with my voting history to create a profile about me? Where did you get that data? Did I consent to use of that data? To what extent is the consumer piece factored into the profile you create about me? And do all of the profiles in VAN factor consumer data into the scoring? And what about my basic right to privacy?

Nobody could answer these questions for me. 

After I choked on the data that was factored into the voter profile, I asked about the number that was assigned. Someone could be given a number of 85. Or 97. Or 43. The higher the number, the more likely it was that they were going to vote for Kamala Harris and the other Democrats on the ballot. 

Ultimately, a canvassing organizer was to find out where a volunteer wanted to door knock, and together they would decide to visit voters who had a profile or score within certain numbers; say, 80-90. Higher than that, you were somewhat wasting your time; those voters would vote for you anyway. Lower than something around 70-75, and you were also wasting your time, as they were an unlikely voter. 

I was able to do some canvassing with VAN. The questions were clunky. They told me nothing about what the issues were. Instead the scripts that had been prepared for this particular turf wanted me to rank, 1 to 5, the voter’s likelihood of voting for certain candidates. 

In other words, the script was trying to find more data about the people on whose door I knocked. 

But how and when do we reach those people with low VAN scores? Do we ever reach them?

Where we miss the mark is that we aren’t talking to those people with low VAN profiles. 

We need to talk to the Independents, and to the Republicans. We need to find the people who need persuasion. 

And therein lies the biggest fault in the system. 

We simply are not reaching them through VAN. 

I heard it over and over through this election cycle: “We hit 1500 doors today.” “We phonebanked and called 3,000 voters.” 

Great. But did you really even scratch the surface of the persuadable voter? 

If you did use VAN to reach those low-scoring voters, please let me know. I want to hear your story. Did it work? I’m open to suggestions, but I only have until around August of 2026 to figure this out.

What will I do differently next time? 

I will have more presence in the hyper-local newspapers. Here’s why: I had a person show up to a Q&A in Fulton County who happened to see my banner ad in the local paper. She decided to show up to my event and find out more about me. I had similar stories in Malone, NY. I knocked on doors in St. Lawrence County and the voters said they had seen my ad in North Country This Week. And I picked up a few attendees to my events in Glens Falls by running a last-minute banner ad in the digital version of The Post Star. 

Focus Groups

Ideally, we would have focus groups, like we see on the television networks before a debate. People from all sides of the political spectrum would convene, listen to information, and share their views. 

Sounds great. It’s really tough to do. 

To pull it off, you have to be able to reach out to a voter, convince them to show up, and have them speak their truth. Easier said than done. The norm is that Democrats show up to hear Democrats, and Republicans show up to hear Republicans. 

Issues-based focus groups might work but they are still super tricky to shape. There were a few info sessions on NY’s Proposition 1 (the Equal Rights Amendment) in my district. Some of these got downright heated. And it was interesting to find out how people learned about it. For some, it was as random as seeing a note on a bulletin board at the library. And I’m not sure that anyone attended who was truly unpersuaded, or that anyone left feeling any different from the way they felt when they walked in the door. Bias runs deep. 

I’ll make more door-to-door visits with no regard for voter registration.

Here’s a Washington University in St. Louis map from 2020 of the red and blue divide. 

It will be interesting to see an updated map for 2024. 

We are not going to add more blue to this map if we only talk to other blue voters. 

The best voter data system will be one that helps groups of volunteers, such as town committees, share information as to which streets have been visited by canvassers, who actually answered and who got the “canvassing light” touch of simply leaving literature on the door knob. Ideally, we would make a couple of other visits to actually have a face-to-face conversation. 

The value of the technology would be to track whether a house had been visited. Dare I say a Google Sheet could accomplish this? 

What I learned when I went door-to-door without regard for party affiliation

Since I had no real working data tools, I went door to door blindly. Literally, my team and I drove to a neighborhood, said, “this looks good,” and got out and started knocking on doors. 

What I learned was that, other than meeting voters who explicitly wanted to see me defeat my opponent, most voters (usually 6 out of 10) thought the economy was top of mind. Others were worried about their kids’ schools (not really a congressional issue, but I can listen and hope to troubleshoot). Veterans were disturbed by my opponent’s handling of the PACT Act, and Trump’s call that vets were “suckers and losers.”

In other words, I had real conversations. 

I convinced a few voters. I should have used more data tools to be able to tell you exactly how many. 

I may not have the answers, but I do know that what we are doing is not working. 

Our current ways of organizing field work on campaigns are not working. I mean — isn’t that why we are all hurting this morning?

It’s not working.

Don’t even get me started about fundraising; that’s for another column. 

What I do know is that the vast majority of my voters in this huge rural district have not seen a candidate in decades – if ever. It’s prime territory for Democrats (or Republicans) for a canvassing team to go door-to-door and build relationships. 

Have conversations. Ask people, “What’s important to you this election?” Most people will give you an earful. 

And then I need everyone on that canvassing team to have some easy tool in their phone (or on paper for those Adirondack areas where we don’t have cell phone coverage, along with the commitment to transfer that information to whatever tech tool we have when they are back at home with wifi) to keep us all informed about who was visited and what their top-of-mind election issue was. (Sidebar: I heard of voters in Pennsylvania expressing fatigue because they had been visited 4-5 times by teams, or called an equal number of times. Really? At some point we are pushing voters away with our zeal. Update the data and move to the next door!)

SIDEBAR: More than one volunteer has expressed fear of retribution or violence. For that reason, they will not go door-to-door blindly. I get it. We have had some horrible incidents. I can tell you that I have had some of the most amazing conversations with voters by going door-to-door. The die-hard MAGA supporters were usually curt, but ultimately polite. I never experienced a feeling that I was in physical danger.

Finding out what the voter’s top-of-mind issue was can inform me, the candidate, of what I need to be talking about in the press and at voter events. 

VAN fails in that it creates silos of pre-selected voters. By the time we are all settled in for watch night parties, we are convinced that our candidate will win because we had such positive responses when door knocking. But the basic criteria on which those doors were selected indicated that this voter was already leaning towards choosing your candidate. 

And we never got around to knocking on the doors of those persuadable or undecided voters whose VAN scores never landed them on our lists.

I don’t want a political tech tool that rates my voter based on how they voted the last 2-3 cycles, and their consumer choices. Ew. Leave their privacy rights intact, please. 

I just want a tool that will allow me and all of the people on my team to walk the blocks, knock on the doors, have the conversations, share some of the issues with the rest of us. 

Should I just use a Google Doc?