The People's Audit - Measure First!
Never Prescribe Before You Fully Diagnose
I credit my years as a mechanic more than my MBA when I approach any new organizational challenge. Even as a mechanic, I would watch people jump to conclusions and prescribe solutions before fully diagnosing a problem.
A common car problem, in the Southern US heat, is a dead battery. If you bring your car to a shop, 9 times out of 10 they simply replace the battery with a new one. 95% of the time that would be the right solution. But hopefully, your family does not get stuck someplace dark and dangerous if you fell into the 5% where it was something else. For example, something in the rest of the vehicle’s electrical system.
To check the 5% chance of cases takes about 5 more minutes for a mechanic with the right tools. Don’t you think they should all make that extra effort if it means YOUR FAMILY’S SAFETY?
Well, with our country’s future on the line, I want to share my experience and show you all the ways you can diagnose your state’s voter rolls.
Whenever I approached a new organization that I had to scale, I would immediately start looking for things to measure. And most importantly measure those items over time. For example, weekly and monthly financials were typically available.
However, I would sometimes get blank stares back when I asked questions like:
How many sales are won AND LOST each week/month?
How long does it take to complete a sale?
How many support calls do you get daily and weekly?
How long does it take to fix a support issue?
Sometimes the managers would ask, “Why would you have to know that???”
That is because most people only focus on “How can you do MORE?”
However, what I focus on for scaling companies is “How can you do it BETTER?” A natural by-product of doing something better is the ability to do more of it.
So how does that all relate to the Voter Rolls?
The same kind of precise troubleshooting is needed here. We already saw how both ERIC, working with the county SOEs, were all about let’s register MORE people. How about if we focus on registering voters accurately? How about we do elections BETTER?
If you were running a business where an important part of your service is sending mail to your customers, you might have a little interest in that mail arriving to them. If you think of the importance of chain of custody in elections, you would also think having a valid mailing address would be CRITICAL to providing mail-in ballots to your constituents.
So the FIRST thing we recommend is measuring the quality of those addresses and what changes in them over time.
And how can you measure the quality of mail addresses? Do you think online retailers like Amazon and eBay that survive on their products arriving through the mail may have a clue on what to do there?
Early on, online retailers were subjected to growing losses surrounding bad addresses entered by innocent mistakes and intentionally being used to perpetrate fraud. Can you imagine mailing a $3,000 Apple computer and then the customer claims they never got it? Would you just send another package to the same address?
We were told by one Florida Supervisor of Elections that “per the statute” they can send up to 3 ballots to any person who claims they haven’t received their ballot. However, if you read the full statute that they were citing, even a high school student would understand the statute was referring to mistakes during in-person voting.
Over the last two decades, online retailers worked with the US Post Office to create the Coding Accuracy Support System (CASS). CASS checks are performed by the US Post Office to ensure the address is correct and deliverable. It will even correct typical data-entry issues like incorrect zip codes and misspellings, and most importantly, it will standardize the address so it can be quickly recognized in the US Mail Sorting systems. As an incentive to use CASS, users were even offered lower postage rates and guaranteed delivery since the addresses were verified as deliverable.
Best of all addresses can be checked by the UP Post Office for FREE and in bulk with a widely available web service. You do need some programming services, but this tool has been so widely available for so long, that it is built into most online store platforms and bulk mail software.
I was always surprised when communicating with some Florida Supervisor of Elections that they never heard of this CASS check and worse complained back to me that using property tax records or the 911 systems was superior to it.
How could a group of election officials promoting mail-in voting everywhere be so lost on how to effectively mail items using widely adopted, proven standards for sending mail? Based on what I heard from them, I don’t think these election officials could get a job at a company that was sending cheap junk mail, let alone be trusted to send out hundreds of thousands of mail ballots.
How can you spot a CASS Address? The most obvious indicator will be an extra 4 digits at the end of the zip code. HOWEVER, you still have to check it against the USPS Web API to know for sure if the address is deliverable. When passing an address to the API, it will return the following important information:
DPV Confirmation Codes are as follows:
Y = Confirmed Deliverable Address.
N = Undeliverable As Addressed (UAA) This will bounce back.
D = Secondary Address is Missing (Possible UAA). For example, the apartment number is missing. Up to the mail carrier to decide if they can deliver it.
S = Secondary Address is Incorrect (Possible UAA). For example, the apartment number is incorrect. We have observed some addresses showing “Apt 8”, but the correct address is Apt 8A so this incorrect address may not be delivered. Up to the mail carrier to decide if they can deliver it.
Error Code = In some cases you will get errors back like “Address Not Found.” Again these will fall to the mail carrier to decide but this code is how we found Red Belly Rd. The address may be a new development that has not been registered with the Post Office yet. The carrier will attempt to deliver it and may report the address back as Code N or Y. If you remember the Red Belly Rd story, the mail carrier did recognize the names of the couple who did get their Voter Information Cards, but the rest of the street was returned as undeliverable.
CMRA Code. If “Y”, this would indicate this is a private business acting like a mail-receiving agent. For example, this address may be a UPS Store. Per FL law, it is illegal to use a UPS Store as a Residential Address. However, these mailing services can be used as a mailing address. We recommend looking closely at these addresses.
Business Code. If “Y”, this would indicate this address is a commercial business. Keep in mind, in some cases, Nursing homes and some Retirement Communities will show as a business address.
Vacant Code. If “Y”, this address has been unoccupied for 90 days or more. From the feedback we have heard, the UPSP is a little slow in switching this code back to “N” when someone new may move into this address.
There is also some good information in the “Footnote Codes” for troubleshooting undeliverable addresses. But all of this information will give you tremendous detail on the quality of the addresses in the voter rolls. As well you will also get back the address as a CASS formatted address with corrected abbreviations and zip codes.
If your voter rolls are currently missing the zip+4 in their addresses, there is no way the election officials can tell you with any verified accuracy how good the addresses they are recording are…But you could!
Also one of the small, overlooked details that will come back from the CASS processing will be the correct abbreviations for the address. We typically all assume a digit state abbreviation for example, but in some parts of the voter rolls, primarily the mailing address that may take an international address, the states may be typed out like “Florida” as opposed to “FL.”
If you want to count how many voters are in each household with any accuracy, it is SUPER IMPORTANT to make sure ALL your addresses match in formatting.
We found a yacht club in Clay County Florida, that appeared to have a little over 2,000 people registered there initially. But when we looked closer, the address had many variations with the city spelled differently, and that was expanded with the street being shown either full “Street” vs “St”. Here are some ACTUAL examples of the city variations pulled from last month’s voter rolls in Florida (NOTE: THIS IS LISTED AS THE SAME ADDRESS):
GARDEN COVE
GHREEN COVE SPRINGS
GR COVE SPRNG
GREEEN COVE SPRINGS
GREEN COVE SPRINGS
GREEN CAVE SPRIGS
GREEN CAVE SPRINGS
Green City
Green Coe Springs
Green Cove Prings
Green Cove Sorings
GREEN COVE SP
GREEN COVE SPERINGS
GREEN COVE SPGS
To a computer, these will be considered distinct and different addresses because each version of the address is different and unique. Another important point. If you see examples like the misspelled cities and missing zip+4 in your voter rolls. THERE IS NO WAY ERIC CAN EFFECTIVELY MATCH THESE ADDRESSES TO THE NATIONAL CHANGE OF ADDRESS (NCOA). The US Post Office’s NCOA stores both the former and new forwarding addresses in CASS format. If the address you are trying to compare is not in CASS format, the chance of it matching in the NCOA will diminish radically. So you miss people who moved.
Most 3rd Party NCOA software used by the printers has to convert the addresses to CASS and see if it can correct addresses BEFORE it can check against the NCOA list. This check is required by the Post Office to qualify for the reduced postage for pre-sorted bulk mail. A good question for election officials is why do 3rd parties like the printers using this software know more about our voter rolls than they do?
Once we standardized our addresses we found an astounding 4,363 people registered at this single location. Double the 2k voters we first counted without standardized addresses. As you can see from the picture below, this appears to be an abandoned commercial building.
Once you have your addresses standardized, we recommend you count the following:
Count how many registrants within the same address
Count how many of those registrants voted in the last election
Count how many of those registrants voted BY MAIL in the last election
There is one extra step that will be very helpful in spotting suspicious registrations. If you remove the “secondary address,” for example, the apartment, unit, and suite, you can then count how many people are in that entire building. This is very helpful to see how many people are registered in high rises and nursing homes. This approach will ignore the apartment, suite, and box numbers that are in the distinct address.
From these counts, you can then calculate the following percentages:
Turn Out = Total Voted divided by Total Registered
Mail Usage = Voted By Mail divided by Total Voted
You can then use these percentages to filter your counts to find:
Unusually high volume of registrations. For example, a yacht club that has 45 voters per boat slip.
Unusually high turnout. For example, anything over 60% in a popular election is very suspicious. Historically turnout for even the most popular elections was 45-55%.
Unusually high mail voting usage. By itself, this may not be unusual. For example, we expect mostly mail-in ballots from people using government buildings and nursing homes, however, if you see it along with a high turnout percentage, this should be investigated.
Here is an example of our Same Mail Address Report before the standardized address:
This information would be a great place for canvassers to start investigating as all these locations should be vetted before jumping to conclusions.
Next, since you have checked the address with the US Post Office and have the DPV Confirmation codes, you can count the addresses that don’t have at least one DPV Confirmed Address. The first time we ran a report on all the UAA Residential addresses, the election officials immediately rebuked the report since in some cases a Residential address will be an Undeliverable As Address (UAA) if they are in a rural location. That is why we have a Mail Address for the Voter Registration so a valid mailing address can be given.
So we counted the voters that had both Residential and Mail Addresses as UAA and then ranked them per county. In Florida, we have 67 Counties so this report was split into 34 Counties that had the Most UAA Addresses vs 33 with the Least UAA Addresses.
This approach will help spot the worst offenders which also have the most volume of UAA potential to “influence” an election. As well as reward the counties that are doing a good job at keeping their voter rolls clean. Flagler County, for example, is the 35th largest county in Florida by population but consistently ranks in the top 10 of Least UAA Addresses. Flagler also has a large rural population, so it makes it hard for election officials to blame rural areas for UAA addresses.
Another place to measure is the number of registrations per month. You can use the Registration Date and simply count how many voters fall in a specific month. Highly recommend also splitting this number out per Party so you can look for trends. Here is an example of our Monthly Registration Report.
I’m calling the unusual spike around October 2020 “The Trump Tower” although I feel that most of the voters there were probably stolen for Biden.
If you watch the trendlines, somehow Florida elections officials registered over 153,309 new voters IN A SINGLE WEEK! On the last day of the deadline, October 5th, 2020, 46,644 were registered. Then officials used Covid to add another 24hrs on to the deadline which is normally 29 days before the election. So on that last day, October 6th, 2020 an additional 21,895 were registered.
Our weekly average for registrations in Florida is about 21,340. So by some miracle, election officials were able to register over 7 times that number all during the typical busy time to prepare for our elections. Another miracle is how motivated these last-minute registrations were to vote because their turnout was 79%.
You can also add another layer to count these per county to watch for unusual patterns per county. One of the most haunting issues we see in the line graph is how perfectly proportional our registration counts per party seem to be. Somehow the number of Democrats getting registered dances in harmony each month with the number of new Republicans getting registered.
As Dr. Frank would say…
So I hope I have demonstrated how much information can you measure just from a single instance of your voter rolls. Everything listed is exactly what we started with in The People’s Audit system. What was a game changer was when we did this analysis each month in Florida and carefully monitored what was changing on each voter roll each month. If you followed our Red Belly Rd story, we were on a mission to find these address flippers statewide.
What we finally created was a monitoring system that could monitor the voter list maintenance in every county every month as well as spot Red Belly Rd type address breaks…THE MOMENT WE SEE THEM APPEAR. We no longer have to wait for them to switch back to the original address to find them.
We’ll have a separate Part 2 of the “Measuring First” discussion just to cover what you can do with monthly monitoring.