Traditional Compiled Business and Consumer Lists:
1. We strongly suggest your primary source to obtain your current-month, retail, compiled business data is from the largest compiler, and your primary source for current-month compiled consumer data is Epsilon/Equifax. We are certain their accuracy, conservative models, and (in most cases) coverage better meets the needs of most marketers than do their competitors’ lists.
Nearly all of the nation’s largest and most sophisticated mailers start with data from these two sources. The very largest then supplement data from these sources with data from other sources. With certain caveats, InfoUSA business, but not consumer data might also be considered as the primary data source for certain industries that lend themselves to Yellow Pages ads (InfoGroup’s primary source of data).
2. No matter which is your primary source it is essential to obtain the most appropriate of their available data.
As an example, there are many, many places to obtain the largest compiler’s data, with some offering over 22,000,000 domestic records, and some of their reps include “second class” records, yet their retail file contains less than 14,000,000 and well over 10% of these are undeliverable addresses that fail even minimum postal requirements.
There are also examples of businesses gone nearly 10-years in the Equifax and ALC business files.
As important as selecting the most accurate sources, the biggest secret in the list industry is that the compilers sell their current-month “retail” data at premium prices with their own reps, and they also sell “wholesale” out-of-date partial copies to list resellers at a fraction of the price (28-cents for a typical telephone record from the original compiler versus 5-cents to a reseller (who sells it to you for 12 to 15-cents)).
Most wholesale copies are updated only quarterly or semi-annually (and are all too often overused and even boot-legged), so are always 3 or 4 (depending on the compiler) months old, to up to 8+ months old.
Records in most lists are considered “current” if they have been even minimally acknowledged within 12 to 24-months. Similarly, even the largest compiler sells weekly hot lists of “brand new businesses” even if they are more than 2-years old.
Buying data that is even 3 to 8 or more months older than “current-month” data, no matter how low the price, is compounding your already low deliverability and appropriateness percentages.
Do not buy your data from a website or wholesaler or reseller unless they can prove it is being obtained directly from the compilers’ current-month, retail database the day you ordered it.
You’ll know first because it will be very expensive, and second because they can’t allow you to download it yourself as they obtain it from a wholesale website or FTP. We will allow you to download your data directly and will even provide you with a copy of our invoice (sans pricing).
The websites and resellers advertising monthly “update” dates on their databases are starting with (3 to 6-month) old wholesale data and while they may run it through postal processes (such as CASS, LACS, DPV, and NCOA) monthly, the data is only refreshed by the compiler on a quarterly or semi-annual basis (again, always using out of date data).
Even the largest list brokers, resellers, ad agencies and others that obtain data monthly from the compilers are receiving the same, already out-of-date data. Updates are not compiler refreshes and, again, the salespeople at these firms often don’t know they’re selling out-of-date data.
3. Beyond starting with the most current, retail data from the most accurate sources is to know what data is modeled and what level of accuracy to expect.
A quick comparison of sales volume counts on various business lists, and income ranges on various consumer lists, immediately demonstrates that these two compilers have the most conservative (and we are certain, more accurate) models. Please see the charts below, “Compiler Counts Compared”.
4. Beyond the most common models for sales volume and number of employees on businesses and income on consumers, models are utilized to generate or supplement the populating of many, many data elements and selects, all too often with (in our opinion), unacceptable results.
Lists are already unacceptably inaccurate (the USPS reported 25% of direct mail is undeliverable as addressed), and reliance on questionable data elements and models can greatly exacerbate the inefficiency of your lists. Many lists are not what they claim to be and the list seller and customers are not aware, period.
As an example, Epsilon/Equifax showed only 3,200 20 to 40-year old homes in our county (from court house records) and InfoGroup had 52,000+ (from a model that assigns all households to a given parameter if 51% fit). Needing 10,000-names, the client bought the InfoUSA list and soon found the bath remodeler mail piece was going to whole neighborhoods of brand new homes.
5. Buy the day after the file is updated to obtain the most current data. Run fresh counts to verify the update took place. Plan your mail drops according to the update dates.
Databases don’t change 15% per month because businesses go in and out of business or 15% of the populations turns 18 and 15% die, they change because the sources feeding the file change. New sources are being added as regularly as former contributing sources become unavailable.
6. Be sure to understand the available selects to obtain the most appropriate data for your campaigns. There is so much inaccuracy in all lists (again, 25% of all direct mail is undeliverable according to the USPS), compounded with inappropriate selects and inaccurate “modeling” (please reread #4, above), that the variable outcomes (successes and disasters) are endless.
Start by understanding the available postal selects (because wasted postage and printing are far more expensive than list selects), like Zip+4, DPV, LACS, NCOA, CASS and similar.
Beyond out of business and other inaccurate records, even the largest compiler’s Retail file contains 10% nearly undeliverable records that don’t even qualify for Zip+4 as the USPS doesn’t acknowledge the addresses (and so fail to qualify for postal discounts). We’ve seen many lists where nearly 40% do not meet DPV (Delivery Point Verification) minimums.
Continue by understanding the accuracy of given selects and how completely and accurately they’re populated.
7. Be sure to use every appropriate select on each new campaign. Where applicable (more commonly when using subscription, specialty, or response lists), buy the 30 or 60-day hot list names in addition to narrowing your list as much as possible when testing any mailer for the first time. Buying the most current names as well as selecting the most specific audience available assures your first test is your best test.
If your mailing is successful, you can test less specific, less expensive lists against that benchmark whereas if you are not successful you’ll have to buy the more targeted list and run the test again.
8. Consider matching your list to a second and even third-best source and concentrating on those that match more than one database. You can easily knock out 20% or, if you need more not fewer names, add 20% not being mailed as frequently as the 80% that match more than one source. Again, be quick to mail them after obtaining your list the moment after the update.
There are ways to match databases at extraordinarily low cost (1-cent or less), and you can usually obtain the match rate for free without an obligation to buy, before you make your decision to buy. Match parameters (tight or loose) are critical here. Please read more about match parameters below.
9. What level of record do you need? Many firms acquire more expensive products than needed to obtain certain important data elements. In many cases a given data element can be added to a less expensive record or a separate list containing just the desired (or not-desired) data element(s) can be merged into or purged from the primary database at very significant savings.
It is imperative to know all of the product options to determine the most efficient route to obtaining what is needed.
Unfortunately, few compiler representatives and list brokers/salespeople even know the various combinations. This is one area where we save our clients a tremendous amount of money.
As an example, with far less than 1% of all business or consumer records showing a bankruptcy, those clients wishing to have them excluded from their lists can often simply buy a separate list of those in bankruptcy to purge from their list at far less cost than paying for bankruptcy (yes or no) information on the entire database. There are many examples similar to this.