I learned quite a bit while researching my May Campaigns & Elections column about Democrats’ list management practices. I talked to IT Directors for state and local parties and what I learned was discouraging. On the other hand, it made me feel better about Virginia.
The software that the DPVA uses (named “Prevail”) is ill-suited for its application, and really needs to be replaced, but the party’s data practices are great. The DPVA’s IT Director, Brenner Tobe, works hard to make sure that the data in the central database is as current and as accurate as it can be. This is accomplished by working with campaigns, to make sure that their data gets back to the party database, and by working with localities to get them using the same system. What worries me is the localities that are not properly using Prevail, which hinders both those localities and candidates who would benefit from that data.
One local party, which I won’t identify, is a fine example of how not to use Prevail. Because of the limitations of the software, this locality is using their own software to manage their lists, which is certainly understandable. But they’re doing it wrong.
Strike 1: The software is based on Microsoft Access, and, unlike the use-anywhere Prevail, can only be used by an individual sitting at the computer on which it is installed. District and precinct chairs are, consequently, out of luck. They can’t look at or alter the data, meaning that the people with the best knowledge of the area’s population — who’s a Democrat, who’s a Republican, who needs a ride to the polls, who will put up a sign, etc. — are prevented from recording that data so that it can benefit the party.
Strike 2: The substitute software being used by this locality has no concept of a “precinct” or a “district.” So if a precinct or a district chair needs a list of who’s in their district (which is the #1, no-questions-asked, single most important thing for a chair to have), she must provide the locality party’s chair with a list of the name of every street in their precinct. Well, not every street, because that’s too much trouble. Just a list of a bunch of streets. That list is used to extract a bunch of names from the database, though it cannot limit the lists to only the portions of the streets within that district or precinct.
Strike 3: Synchronization of data between the locality’s Access-based list management software and the DPVA’s Prevail is delayed significantly. While the Kaine campaign was gathering voter information, data was available to everybody the moment it went as it went into Prevail. Any data gathered by this locality, on the other hand, sat on their own computer until such time as they exported the data to the DPVA’s system. Candidates unlikely enough to need to campaign in this area will always be behind.
I know a district chair in this locality. She’s been trying to get the list for her district for six months. She still has no idea of who lives in her district, who is willing to volunteer, who will put up signs, or whether who is a Democrat, who is a Republican, and who is an independent. She’d like to have a district picnic, but she can’t figure out to whom she’d send invitations.
Good candidates, good policy, smart strategy — these are all important elements in winning campaigns. But good lists trump all three. We Democrats don’t deserve to win back Virginia if we can’t even list the members of our party. Here’s hoping some of these localities will listen to the DPVA’s Brenner Tobe and get their act together.
Have you ever used the Fairfax County voter file- and do you know what data it has that Prevail does not?
I would no more have cause to review the Fairfax County voter file than I’d have cause to peruse the list of incarcerated felons from Greenville County, the list of USGS daily-monitored groundwater wells in Galax, or the life list of Jane Doe, Westmoreland birder.
In case there’s any doubt from my blog entry, I must point out that I think it’s great to see local parties using software other than Prevail. The DPVA functions best as a data repository, not a software distributor. If Prevail works best for a locality (it could happen), fine, use that. But if they’d rather hack something together in MySQL and PHP, that’s fine, too. As long data is exchanged with the both the DPVA and districts/precincts regularly and accurately, that’s what’s important.
I sort of like Prevail… What would you change? Besides Nationalizing it, becuase the DPVA keeps Prevail very up to date.
Waldo, I am a huge fan of Brenner’s too. He is a joy to work with, and he is an absolute data-wrangler on behalf of Democrats. The party is exceptionally lucky to have him.
Like you, I have little patience for the data-hoarding and luddite tendancy exhibited by different pockets of the commonwealth, and worst of all, by certain individual electeds.
Prevail is extremely difficult for most people to use. The biggest reason is because Prevail is a desktop application that must be used, awkwardly, over a Citrix connection. So when you export data and save it to your hard drive it is, bafflingly, not on your hard drive, but on the hard drive of some computer in Richmond. Then it has to be exported via Citrix to your own desktop. It confuses me every time I use it, and I’m no slouch with computers. Because of that Citrix interface, Mac users find themselves faced with a Windows program running on their Mac, which is bizarre. But it’s weird for everybody to have a desktop in a window on their desktop.
If there were some way to have people run Prevail on their computers without piping it through Citrix, I think that’d be a huge improvement. But Prevail is a desktop application, and that can’t happen. Instead, we should have a client/server relationship — we run the client of our choice on our system, and it sends packets of data to and from the DPVA’s server…rather than screenshots of a Windows desktop in Richmond at 10fps.
I’m glad we have Prevail. But it’s definitely time to move on. I wish that the DNC would establish a standard data storage protocol (they did create an XML DTD for Demzilla, but who knows where that stands now?) along with a standard command set for am XML-RPC / REST interface. That would allow software developers like me to start making client and server software, knowing that they’ll all be able to speak to each other when we’re done.
Prevail does have a huge learning curve, but I think those who are forced to use it regularly, as I was last year whill working for the Dem Coordinated Campaign, do tend to adapt to the program for their own uses. It is actually quite versatile, but unfortunately also very cumbersome.
I’ve found this year that the DPVA date for Charlottesville/Albemarle is freaking awesome. Canvasses this year are taking half as long as last year because all of our data from last year is included in our universe. No knocking on a door of a strong voter who moved out in 2004 only to find a UVA student who’s registered to vote in New Jersey. At least, not until we tackle UHall precinct later this year!
It has been very gratifying to know my hard work was useful past the 05 cycle, and that it makes it that much easier and more efficient to make the data even better.
Waldo, I figure you were talking about the Fairfax voter file since it is a county file, and access based. The Fairfax file is the best one in Virginia (although not perfect). Here’s why:
First, (again assuming you are talking about Fairfax) whoever told you that the file can not be divided by district or by precinct is incorrect. The first screen in the Query allows for dividing by precinct, by magesterial district or by legislative district.
Second, it only updates to the main file when you click on the button to upload the updates. A user can update once a year or everyday. Why is this important? Say you are running a House of Delegates campaign and you phone bank some voters. Do you want to upload voters that are voting Democratic for State Senate but voting Republican for House of Delegates? Well, if you are the House candidate, you probably don’t want to share that info as the Senate candidate will put them on the list to turnout. It would be great if all voters were straight party, but that’s just the reality of the situation.
Third, the other thing I like about the Fairfax file is it goes back longer than the state file. Incumbents are around a long time, and it is helpful to see history. Would it be useful to know for an opponent of Virgil Goode which voters came out for him in the 1994 Democratic primary against Chuck Robb- of course.
The fourth thing I like about the Fairfax file is street sheet formats. It shows long vote histories of voters, while prevail only shows a few recent years on walk sheets.
Finally, the Fairfax voter file also has donation history to the party and candidates, it has caucus vote history and it allows for wild cards- i.e. in prevail you can sort for people who voted in the 04, 05 and 06 Dem primaries, in the Fairfax file you can sort for people that voted in ANY three Democratic primaries.. and when the primary history goes back to 1970 that can capture a lot of data.
Again, I apologize if another county uses a access based file, and you were not refering to Fairfax but that’s what I like about the Fairfax voter file.
Though that’s all interesting, Ben, your comments now constitute the sum total of my knowledge of Fairfax’s database — I definitely wasn’t talking about Fairfax. I’m glad to hear that the database is so good.
Dan, that is awesome news. So often the story is precisely the opposite. The purpose of our lists is to function as our institutional memory. The fact that the Charlottesville/Albemarle data in Prevail is spot-on is great news.
So, at the end of the day, it is not data-hoarding luddites in the localities but rather some localities understanding that Prevail was designed by the State organization and serves the needs as determined by the state organization, which may or may not serve all of needs of the local grassroots organizations. As a result, local committees often will look to fill the technology gap by maintaining their own database. In the finger-pointing process, this might be kept in mind.
Maintaining a separate local database does not detract from the state database. Not spending the time to update databases, local or Prevail, does detract from the overall goal of comprehensive voter data access.
As a developer, I agree with Waldo in that there are many ways to improve the user interface. Ease of access is not the only concern. Security is an even greater concern – statewide and locally. In many ways, it is Prevail, not local databases that provide the road blocks.
Local database efforts revolve around capturing the same data as prevail but also maintaining additional data, as Dan mentioned. Additional data fields could also be used to identify Democratic activists. This type of data is very much sought after by both the State organization and candidates. Campaigns can get a leg up on primary opponents by having this data as they kick off their campaigns. A new-comer may benefit from this data while a veteran candidate may rather his or her potential opponent not have the same level playing field.
A State organization may find this data helpful especially if they wish to by-pass the local organizations and go directly to activists. This is usually a mistake. Those that focus on the technology aspect of political organizing often make the mistake of not realizing that politics is really local – this means that there are private, personal networks that are energized by personal relationships.
The state effort will be more successful if the fundamental design is built from the perspective of the local organization – bottom up and not top down. Local folks, often without a technical background need to see how all of this works for them. For sure there have been improvements in the data. Where did that data come from? It can from local boots on the ground. Improved ease of access, ease of data manipulation and output as well as the ability to capture data the locality finds important will get them more interested in participating in the “big picture”.
It is likely that the state will never be able to satisfy all needs. The state organization should recognize that folks have their own databases for valid reasons. The state can take the Microsoft “kill the competition” approach or they can work with localities to help redesign the database and access to it. If local committees are going to have their own databases, it may make sense to have some kind of batch processing so that the local folks do not have to re-key. Now I do not have access to Prevail, some of what I am assuming may already be the case, such as batch up dating. However, my main point remains the same – make this more appealing to the local folks and more local folks will support it.
Local Democratic volunteers that have, on their own, taken us from 3×5 cards to local databases are not the problem. They should be thanked for the hard work they have put in. Critics may want to keep that in mind.
As time goes by, more and more local folks will become more proficient at database development. Skill sets and technolgy aids continue to race forward. The state will have to improve its services just to remain the same….and it needs to do better than status quo. Technolgy will provide powerful resources for the local organizations. The state could provide leadership here…or the local organizations can continue to provide alternative local solutions. If I remember the old Star Trek episode, they underestimated the luddites didn’t they?
I have messed with Prevail for many years, and I have used local databases as well. The biggest failing — that makes the local database-to-Prevail link a bottleneck — is that it is apparently still not possible to have some batch merge of data from the locality to Prevail. We would love to be able to tell Prevail about what we have done, but the only way to do it seems to be to code a data entry sheet and scan it with a hand-held pen scanner.
There is another reason why local databases are still critical. We don’t store fundraising data on Prevail, but we do locally. So if we are sending out a letter to raise money for a City Council race, it is using our information.
In a better world — at least better from the standpoint of maximizing data sharing, which to some is a worse world — every Democrat who raised money would enter the names and amounts of all contributors into one database, to which all fundraisers or others with a need to know would all have access. Knowing how jealously political types guard their lists, though, the odds of that happening are about zero.
I have to wonder what the best way would be to establish an XML standard, for storing voter lists, and what the best way would be to establish a list-synchronization standard, for up-streaming local lists to state and national lists. There’s no DNC data standards body to whom an appeal could be made. Somebody may simply have to invent a forward-thinking standard, release some decent software based on that standard, and push it hard.
Or we could continue to flounder. I’m thinking that’s more likely. :)
she can’t figure out to whom she’d send invitations.
*totally* offtopic, but
“she can’t figure out who she should send invitations to.”
is the right way to write it. “To whom” is the bastard child of 18th century grammar books.
I’m forever torn on “whom.” On the one hand, I want it to go away. On the other hand, to most educated people, using “who” as a direct object is a sign of a lack of proper education.
“To whom” is right, but “to who” is better.
Access can be used by multiple users on multiple sites simultaneously. It is nearly infinitely customizable, but it is not easy to use or set up. However, once it is built and customized for the job at hand, it works fine.
When on a host, absolutely, Access can be used in the same manner as MySQL, PostgreSQL, Oracle, etc. But when used for these purposes, it’s almost always put together as a desktop database, able to be used only on a single system by whomever happens to be sitting in front of it.
As an open source bigot, I’m not inclined to see Access used as a database for anything more serious than a recipes database, but maybe that’s just me. :)
I wish I had paid more attention to what we used in the Roemmelt campaign last year, it was pretty useful data. We had excellent street maps and address list sheets for walking around canvassing with bar codes under each likely voter.
The maps were for drivers (making them for walkers or cyclers might help in dense neighborhoods). The barcode on the printed page was easy to highlight answers on.
Back at the office the interface was a bit of a kludge; scan the sheets into a Windows(pc) application with a hand held pen. It updated some database somewhere.
Early in the summer I was predicting the party leaning of the likely voter about 60% of the time. My the end of the campaign the data sheets and my experience had improved such that we were guessing correctly 95% of the time. So the system worked.
So far this year in an unamed campaign; I am appalled at the lack of organization. I can only hope that enthusiasm more than makes up for total chaos.
Phone bank call lists should be sorted on phone number and then last name, and if possible by district, then phone number, then last name. With the last name sorted lists we have now long distance calls are mixed in with the local calls; we could district localize the calling. Additionally, you can call the same house two or three times when people with different last names live there.
Canvassing lists should be sorted by side of street in cul-de-sac or one-way road situations; and climbing numbers when you are making a loop. Grouped by walking distance in densely packed neighborhoods, cycling distance in suburbs, and by car in rural areas.
Hey, all of this is enormously more organized all ready when compared to my GP/LP days (grin).
Scott, you bring up a good point.
There’s one group in political campaigns that annoys me to no end, and that is people who make their naive idealism the entire basis for their campaigning. I’ve seen this anywhere from new volunteers to the candidates themselves.
Not to say that idealism in politics shouldn’t be encouraged. It’s how most people get started in politics and what provides a lot of the passion that is so necessary. But politics is a bloody game, and the sooner a political activist wakes up to that fact, the better.
Too many times I’ve heard someone say “We’re using the grass roots to win!” That’s fine, but it all too often comes at the expense of things like fundraising, or a field program.
The grassroots, netroots, etc., are all fine, and if they do their part, are a great help. Retail politics look good, and buttons, pencils, fans and balloons are fun. But these don’t win campaigns.
When Tim Kaine was running for governor, he spent AT LEAST six hours every day on the phone raising money. As a result, he wasn’t vastly outspent by Kilgore and his campaign workers had the ability to build a great infrastructure.
Political activists tend to live in a bubble with other activists, and it’s sometimes hard to remember that the general electorate isn’t as obsessed as we are and are probably far too busy to watch the full debate or investigate the details of this or that policy initiative. That’s why it’s the campaign’s job to contact these people, briefly talk to them, find out what’s important to them, give them some positive information about the candidate and ID who they will potentially support in the election. An experienced canvasser will know the structure of the family, their political leaning and their most important issues before they even knock on the door.
The identification comes into play during the GOTV plan. Those voters that are supporting your candidate, you get to the polls. You know which ones vote every time, come hell or high water, and you don’t bother them. But you also know which voters tend to forget, and you make sure they get to the polls. You call them, you visit then, if you have to, you drive them to the polls in your own car that very minute. I remember screaming through Dover, NH to get to a polling place after picking someone up at their house two minutes before the polls closed. We made it.
But, if you don’t have a structure, who are you going to turn out? The more ID’s you have, the more votes you will turn out and the better your chances are of winning. It takes money, it takes infrastructure, and it takes an overwhelming amount of manpower.
The Kaine campaign was identifying voters as early as January of 2005. By the beginning of the GOTV phase of the campaign, we had identified more voters than any campaign in Virginia’s history. We had a well oiled GOTV machine, we were prepared and able to respond to every conceivable problem, and things went perfectly in the governor’s race.
But, you know, if we had raised a few more dollars or hit a few more households, Creigh Deeds would be attorney general. That’s why it’s so important to have a strong field and GOTV program. It really does make the difference.
Dan – excellent points.