Telephone Appending
Telephone appending refers to attaching a telephone number to a voter at his or her residence. Accurate telephone appending is paramount to conducting a credible phone banking operation, or absentee ballot drive.
This area of campaign technology is little understood by candidates, and it has resulted in the greatest number of questions concerning product performance. Typically, candidates want to know why there are no cell phones, or why there are dead lines when dialing some phone numbers, or why some vendors tout "70% appends" instead of 50%.
Instead of referring you to one of our many "white papers" devoted to campaign related topics, we have chosen to include detailed information on telephone appending in this product page. We feel it is too important an area to not describe it in detail here. You can choose to read this technical dissertation or not, but know that we take telephone appending very seriously. For many campaigns, effective phone banking is truly the difference between winning and losing.
Where Do The Telephone Numbers Come From?
First, understand that the telephone companies in the United States do not give out telephone numbers to telemarketers or to vendors for resale. In other words, no one can ask their phone company in Florida for the telephone numbers of a county or municipality within the state, or for that matter, any geographic area.
So, how is a master list of telephone subscribers created? The answer will surprise most everyone: there is a small number of companies in the world that purchase local telephone directories for every county or municipality in the United States, tear out every page, run the pages through a sophisticated scanner, and then employ optical character recognition software to transfer the subscribers and their telephone numbers to a database table which becomes a master telephone subscriber list. This master list is kept current by repeating the same process. Primitive perhaps, but it is the only way to do it.
Since the basic numbers for a master list are taken directly from telephone directories, only published telephone numbers are generally available. No non-published numbers are available except for a few telephone numbers that are added based on a hand full of additional proprietary public record sources that augment the list.
As recently as 2001, the one or two companies creating and updating telephone subscriber lists operated out of China and the telephone directory subscribers where actually entered and updated by hand. The technology had not advanced sufficiently to use scanners and software. Master telephone subscriber lists are now being created by a handful of companies in the United States using the techniques we described.
There are dozens of append companies on the internet, but we use only companies that directly process their own data through the method previously described or companies who buy their data from the manufacturers of data created as described above.
Append Ratios and Append Accuracy
There are two ways to judge the effectiveness of telephone appending. First, there is the percentage of households in a voter file to which you can successfully append a telephone number. This is the append ratio. Secondly, there is the accuracy of the append based on the percentage of telephone numbers that will be valid numbers for an actual voter in households you are calling.
Append Ratio
To reiterate, the append ration is the percentage of households or individuals in a voter file to which you can successfully append a telephone number. The append ratio for households will be lower than the append ratio for individuals. That is because once a telephone number is uniquely identified with a household or a specific voter in that household, the telephone number is spread to every member of the household who is a voter. For the sake of simplicity, we will base our explanation on the append ratio of unique, or householded, telephone numbers.
Some append companies or data suppliers purport that they can provide a 70% append rate. Try not to buy into this, and here is why:
A few years ago, I read that the number of US published telephone numbers versus the non-published numbers was equal; i.e., each represented 50% of all the US telephone numbers. So, based on my explanation in the section above of how master lists are created, how can a political operative or telephone append vendor claim a telephone append ratio of 70%? Well, the explanation is that a 70% append ratio can readily be obtained by matching only the voter's address with an address in the master list. If you are charging a per thousand rate, this is better for the data vendor, but lousy for the candidate.
The most sensible, economical, and time efficient method of telephone appending is to to match both the voter's last name and address to a master list of telephone subscribers. We have supervised over a half-million polling and advocacy phone calls over the last 8 years, and if we had used address only matching, it would have cost us a lot more money for the data and more money in salaries to our professional callers because of bad numbers and frustrated callers.
Because of our match strategy, and our insistence on using the most professional append companies available, our best append rates have never exceeded 60% with an accuracy rate (see below) of 90%. Generally, our append rates using both last name and address matching will range between 35% and 55%. The lower rates occur when appending lower socioeconomic neighborhoods or very affluent neighborhoods. The prior has fewer telephones and the latter has a much higher rate of non-published numbers.
We are not in the business of padding our profits by selling products that we know to be inferior, so if you are a campaign manager or data processor looking for an unreasonable append rate, you will probably need to look elsewhere. We are committed to providing the best data available to our candidates for a fair price.
Accuracy
The second consideration is the accuracy of the data used by the telephone appending service. This is dependent on the company that came up with the master list in the first place, or more pointedly, the currency of the list that was used to append your voter file in conjunction, of course, with the method of appending in the first place.
Telephone vendors that do not have the capability of updating their data will eventually produce more and more errors if they only use address matching. Curiously, even if the data is old or not updated, appending using both an address and last name match will produce even more accurate appends the older the data is. See below under "An Anomaly Worth Noting" for an explanation if you are curious as to why this is true.
To Be Fair
Vendors who match by address only, contend that it is better to have too many appends rather than too few appends and miss a voter. For example, a voter might be at an address, but their last name is not the name of the person subscribing to the telephone number. In this case, we would not get an append to this person if we used a last name, address match. We would, however, have the potential of contacting the voter if we used just the address as a match. A valid point.
We evaluate this differently. First, we do not believe that the situation above is that common. There is a very strong tendency for voters to live with other voters, therefore a greater chance that the phone subscriber is also a voter. Now, if we are paying for 20% (70%-50%) more telephone appends to capture 2% to 5% more voters, it is a poor tradeoff in our opinion. The cost in appends and caller time does not justify the hand full of additional voters that you may be able to reach.
An analogy involves your mailing program. It may be better to send a mail piece to every household in your district, but for anything other than a small district, this would be cost prohibitive. That is why most campaigns mail to only quality 3 or better voters. The same thing applies to telephone appends: There is no reason to pay more for little potential return on that investment.
Availability of Cell Phone Numbers
Cell phone subscriber information is not supposed to be available. Period. If a vendor says they can get you cell phone numbers for voters anywhere in Florida, you should take a pass. Don't waste your money.
Although some cell phone numbers are floating around in telemarketing lists, legitimate phone data companies do not have access to cell phones. Take a look at information published on Wikipedia if you doubt their unavailability.
Availability aside, we also doubt the efficacy of using cell phones to generate voter support for a candidate. We know of few voters who might appreciate a political pitch on their cell phone. That goes for polling calls as well.
Telephone Numbers in Apartment Buildings
If a vendor tells you that that they can consistently or accurately append telephones by address and apartment number, consider going elsewhere for your appends. I mentioned how the best companies use the white pages for creating and updating their master telephone list. How many times have you seen apartment numbers in a telephone book? Very rarely.. So apartment numbers are only occasionally available as a legitimate field for telephone appending.
As it turns out, as long as you append by both the address and the last name, not having an apartment number for those voters living in high rises is not much of an issue. Anecdotally, in the past, we appended apartment buildings differently than single family residences. We appended by address, last name, and the initial of the first name. Out of 10,000 voters, we probably appended no more than an additional 50 or so telephone numbers to voters by using this additional match criteria. It turned out to be not worth the hassle for the small number of extra numbers that were appended.
The reason for this is that, except for a few glaring exceptions like Smith or Gomez, surnames are so diverse that you will generally miss only a tiny number of voters, even in a large complex of high rises.
Disconnects, Faxes and Answering Machines
Disconnects
A lot more disconnects are showing up as a result of the foreclosure debacle, especially in hard hit areas of Florida. Disconnects were usually accompanied by a recorded message that goes something like "We're sorry. The number you are trying to reach has been disconnected... Etc." We are now experiencing just a dead line after the call is "picked up." Dead lines are disconnects. Disconnected telephone numbers are generally recycled after 90 to 180 days, so even if the appended telephone numbers were current, the dead lines indicate recent disconnects.
Until the foreclosure rate subsides, possibly in late 2010, we are going to have more problems than usual with disconnects. We are sorry, but there is nothing we can do about the economy.
In connection with changed telephone numbers, we recently spoke with a source from a major telephone company. He said that they are no longer playing a recorded message when a telephone number is transferred. It is just dead on the other end; again a disconnect. We are told that this is because some people are switching to cell phones, and many of them are leaving the company, whether using cells or land lines, for other carriers.
Fax Machines
Obviously, fax machines are worthless telephone numbers. They rarely occur as an appended number, but we have seen it happen. Usually it is because the subscriber inadvertently left their fax machine turned on. Sometimes the voter has two telephone numbers, one devoted to a fax machine, and we unluckily appended the fax number instead of the actual contact number. The appending software would not know the difference.
Answering Machines
Encountering answering machines is inevitable. Actually, they can be an opportunity short of actually talking to a voter.
We understand that there are auto dialing companies that specialize in answering machine telephone numbers, so that they can leave a recorded message. People with answering machines many times just let the message run until they get to the next message. At least, that is how this service is marketed. At any rate, if you are advocacy calling and get an answering machine, it is advisable to have a prepared, short verbal pitch to the machine ready in hopes that ultimately, the person listening will be a voter. In other words, don't just hang up on an answering machine. Give it your pitch.
Election Department Phone Numbers
Some vendors sell State of Florida generated voter data which contains a telephone number field furnished by the submitting counties. Before you assume that you have valid telephone numbers after you get your data from some vendors, you need to make sure the vendor is not using Election Department telephone numbers. When the various County Election Departments asks the voter for a contact telephone number, they almost always request a daytime telephone number. After all, the Department of Elections will most certainly never call the voter after 5:00 PM, so what is the point of requesting a home telephone number from a voter who works.
Daytime telephone numbers should never be used for polling or advocacy calls. Always insist that your data be appended by an outside source.
An Anomaly Worth Noting
The most accurate, productive, and economical telephone append actually comes from old telephone numbers, provided you append using a match on both the address and last name.
Suppose we submit a list of all the voters in your town to a company who has been requested to append by last name, and address. The company has not updated their data in four years, so no one who has not lived in the community at least 4 to 5 years will have a telephone appended to their name and address on your list. All of the appended numbers will be valid and consist of only those people who are long-time residents who, by the way, are frequently very good voters.
My anecdotal testing indicates that if you append old telephone numbers, your accuracy rating will be 90% or even greater even though the append ratio may be reduced substantially.
Yes, you will miss voters that just moved into the neighborhood, and yes, you will miss all of the voters that just moved into the new residential development built two years ago on the other side of the highway. But you will get the best bang for your buck with the numbers that you do have.
Also, remember that new voters to the district will take a long time in getting on the master telephone list of those vendors who actually do updating, so they will not surface for a year or so after they move into your district.
This is not to say that you only need old numbers for your appends. You should have as up-to-date appending as possible. This topic is simply an observation.