Mailcodes

   The Theory and Practice

mail·code

def., mailcode, n.
pronunciation: mA(&)l - kOd
a : something sent or carried in the postal system +
b : a system of signals or symbols for communication +
c : a system of symbols (as letters or numbers) used
to represent assigned and often secret meanings.

copyright © 2002 by Merriam-Webster, Incorporated, their courtesy

In the idiom of direct-marketing parlance, we have ascended to the conjoining of the two wordlets, mail and code, so as to arrive at a suitable, `tangible' word that fits our communication with it. This we have neither stopped nor fretted to do for other words of particular importance to us, such as datafile, filename, username, mailpiece, on through to our lettercopy, presschecks, and signoffs. We are happy have our `license' for doing this, and so we do.

That said, we wonder where we are with our mailcodes, in our mastery of and skill in their use.

If you have seen as many schemes and varieties as we have over many years, you could but wonder if there is some misunderstand or suspicion for the lack of its entire grasp, or the full grasp of `them,' as the category.

For this cause we will proceed to lay out some of the better models that have evolved, and to disclose some of the shortcomings that have made their way into our retrieval systems.

One defining truth, undeniably, is that mailcodes are in the domain of the director of marketing, the fundraising director, or the particular person whose duty and privilege it is to assign them. It is also certain that he or she has discovered the flaw in the manner by which their predecessor has done it.

Thus, for an organization with a thirty-year history, and four past directors, there are always exactly five matching mailcode theories that have been followed. Because this part of the company's management falls to the prerogative and sanctity of imperial dictate, mailing designers do not generally have a final authority on mailcode setup. They suggest them always to coincide with their packages. And during the same thirty-year period, the organization may have enjoyed the services of more than four of them.

In an archeological dig of the company's mailcode records, the surveyor can demark the comings and goings of the said directors with a date clarity that can render no misunderstand. It's as clear as the strata of the earth's crust, descending for a thousand feet.

And to show finally that the field is marked with uncertainty, those of us who have been on the job for thirty years have had our conversions into new thinking on mailcodes. This is perhaps because it's an evolving science within our community, and we should like to take advantage of the latest newly revealed wisdom, as we see it.

Some Varieties of Mailcodes

The most primal, yet still often seen, manner of mail coding is the simple one-letter model:

This is a wonderfully simple and effective method for short term advice. It is the least obtrusive, fairly easily decoded, and easily set-up and verified in the production stages.

When historical records are kept, their clarity is completely undone and obliterated with every subsequent mailing of their use.

To the single-letter model we will add one of its `enhanced' schema:

And the enhanced variety is seen in as many iterations as its alteration will permit.

From the above we see that the intent was to monitor two aspects in the tracking, the numeric portions and the alpha portion.

What Needs to be Tracked

Well, let's see. The organization could be needing to track the kinds of mailpieces that it has mailed. Within those mailpieces are the varieties of offers they contained. Within those offers are payment options and contact preferences. Applying to all of those are the audience sectors, as to the mailfile list segments that were mailed. And in just a moment's consideration we have scores of combinations that could be construed, imagined or considered for `crucial' after advice to be kept.

Information Age

We must at all times remember that we live not in 1955, or for that matter, 2001. There is a proliferation of information in our day unprecedented in human history. The effect of the uncontrolled onslaught of computer-generated information upon the human mind is understudied. In time, ample evidence will be given finding that a person's quality of life is adversely affected; and, for that reason, barriers to block its approach will become necessary to protect the delicacies of our easily befuddled mental processes from a type of insidious collapse many have termed `overload.'

Before that day arrives, we can all do our part in forestalling information overload by carefully guarding against its creation. When it seems there is never enough time, and you have begun to spend more time interfacing with devices than people, a serious degree of overload has already silently robbed your live. The compulsion to feed bioabsorbing devices with ever-greater attention is a dangerous precipice over which many today are falling.

Matched to a machine's capacity for challenge and stamina, the human can never win. The machinery has enticed the human into a domain of operation much dissimilar to our own biophysical setup. It is at ease with this challenge, and was created to present it. The machine's `master' is really no such a thing. This is because a machine, by its design, has rigidly refused to live a human existence with us, or to `be' human with us in any true way. We accept this shortcoming on its part, and therein lies its trap toward us.

With respect to corporate information brought in by mailcodes, it is of first importance to see that no information is captured that cannot be later acted upon with certainty, error free.

We must refrain from capturing information for the sake of its capture, or for the sake of having the appearance of a voluminous and comprehensive achievement. Some person, some day, may be challenged for the comprehension of what you have started. That person could easily be yourself.

If it takes thirty minutes of your time to mentally explore the scope and latitude of a mailcode's schematic structures, both ruling out and including in, every conceivable possibility of interpretation, that is a huge and demanding load that is being installed for others' future self-destruction.

Because we may not be able to `punish' some machine, its scanner and all, to carefully make these interpretations for us, as a servile instrument acting on our pleasure, and sparing us the difficulty of the task, we have to step in and do that ugly work for the machine, as it stands-by watching helplessly. We may have to bang our knuckles on its keyboard for hours, trying to make in understand what we have taken great pain to ourselves digest. Here again we find its complete unwillingness to help us, to make our lunch schedule, or complete our day at work.

The chief focus of everything we design for computerized assistance must become simplicity. There is no time in our natural lives for anything less. If it can be easily understood after thirty minute's study, then there you have it. A thirty-minute price tag. There are a limited number of thirty-minute time blocks in our daily routine. How we will expend that limited time tomorrow depends on the quality choices we make today. Let us today proceed to make some wise choices about mailcode structures.

Make things as simple as possible, but not simpler. -- A. Einstein

    A3D4066
    A3D4781
    A3D4034
    A3D4216

Some Good Working Models

The second thing that will be needed in the future is a good decipherment channel. The first thing will be knowing instantly, without labor, where to find it. Thus we proceed forward for our design of a simplicity system for our mailcodes to later fall into.

The basic problem we find with mailcodes today are their inherent cleverness. The designer has adopted the notion that we can construct a set of codes that will be human readable and human comprehensible. In any progressive setting, this notion is false. To assume you can make it true is to construct a mental mouse trap to be sprung upon the innocents of tomorrow who neither understand your view, nor remembered the keys by which the codes were formerly simple.

Let's anticipate that the codes are never going to be human comprehensible, and from this the progression toward the simplicity system is launched. We go to programming and say, all those past mailcodes we have, and all the new ones to come, any of those that we'll be needing, you will please make your computers decipher for us. And when that's done, we'll be needing to know where we can point our mouses to obtain the system's findings. That `clicker' needs to be at hand at all the popular places were this information will likely be sought. Thank you. Have a nice day.

Since programming lives to perform the above function, you can consider that part done, and then plan to use either scalable barcodes in the new mailcode generations, or human readable (not comprehensible) display codes.

If the latter are chosen, we will need to adopt some rules. Rule one is that they be unmistakable in character formation. This means no "1" ones; "0" zeros; "O" oohs; "l" or ells; will be permitted. All of them can be human confused at data entry, or in an automated OCR reading. We need to prohibit dual "V" vees, VV; as its combination makes a "W" double-U; in some fonts; and we need to look at our font varieties and decide which "fixed-length" character set will work for us. If it's going to be the font of the day at the imager's form writer, then look out, one of these days an unsuitable font will appear. It's best at signoff time to check the mailcode imaging for its readability, on behalf of your data entry personnel. If a perfect font can be decided upon, that's good. Tell the imager to use that font for the mailcodes, in the point size and pitch that's good for you.

Moving Forward

Assuming all of the foregoing will be ignored, let's proceed to make some codes as we have for many years.

For our basic purposes we'll say that we shall want to know from our mailcodes such things as:

These being basic, we can see a scheme that includes, say, five parts.

    
    part 1      package     P     alpha a-z
    part 2      variable    V     numeric 0-9 or 1-9
    part 3      wave        W     alpha a-z
    part 4      list        LCD   numeric 001 - 999
    part 5      profile     L     alpha a-f

    for a combined appearance of:

        P-V-W-LCD-L

        or, filled in,

        B2C144A

To analyze, we can see that programming can affix a meaning for year and month from its records of when the waves mailed.

For our packages, we have 26 possibilities, perhaps omitting "O".

For our waves we have 26 possibilities.

For our list rentals, we have hundreds of preset list names which should never be duplicated.

For our donor profiles we have as many as needed, with our alphas.

The product or package column can be expanded to two characters, and the donor profile omitted; or a permanent `T'est vs `C'ontrol column inserted just following the package. Many people like the two-letter package name codes because they can closely recall a product, such as "WM" for the `Washington March.'

Then you have a concise

    WMC continuation and
    WMT test all in a wink

The trick is to never exceed eight mailcode characters, and to keep all the mailcode meanings, by type, segregated into programmable vertical columns. By doing this, splits can be accomplished by altering a character of the mailcode column by automated means.

You should not try to embed your list rental names into the mailcodes. If a person belongs to the `Tree Lovers of America,' and sees that `TLA' in that mailcode, the chance that the recipient will confront the `TLA' about the rental is large; and the `TLA' might then be reserved in its welcome to serve your name-exchange needs for the future. It also puts your organization off on a wrong footing with a new prospect from the beginning.

Further Refinements

For rental files of the first sampling, choose a numeric range of, say 100 through 199. Then for the second use of file 125, you can call it 225, and a third use, 325, on up until it's established at, say, 525 as a permanent outside list source. You can also block out ranges of names from a large provider to represent all their exchange files, for 6-month donors, 12-month, 24-month, 36-month, best, expired, etc.

Another important consideration is the sort order of the completed mailcodes, especially within your past codes. Ask DP to sort all your past codes and give you a summary count. From that you will see just how you want the sorts to come up in the future, without having to tell programming to sort the mailcode on column four, and such. This will help you position your new codes within the previous schemes, hopefully keeping all the new ones in a contiguous sort-range block.

The above-suggested scheme puts the mailcodes sorted primarily in package or offer order, which is not a bad start. But for all of this, every organization will have its unique desires; and, using these logical steps for planning ahead, your adopted mailcode design will permit you to make a scheme of long-term use and utility for all current, past and new products and offers as they come on line.

Data Entry

If you cannot scan your entire volume of return inputs, use an advanced keying system that brings up a menu for the mailcodes, such that the operator just hits a couple of keys and scrolls to the exact mailcode and presses enter. This will reduce the mailcode entry errors to almost zero, even if ambiguous characters such as "O" and "0" are in the codes.

Preset keystroke macros, in these same advanced systems, permit many current mailcodes to be inserted, error-free, with a two-key macro execution. For keying systems such as these for the PC environment, Alpha4 from Alpha Software Corp. is a good choice and is widely used by hundreds of thousands of users for direct-mail data-entry tasks.

The best Windows GUI DBF Editor is CDBF, which is powerful and easy to use, and it's available for immediate download at White Town Software.