Column: Lists: Black Boxes or Known Quantities?
I want to talk about logic. The list rental world for multichannel direct merchants is experiencing pressure for a profound change in logic, and I'm not sure the change is logical.
First, let's define our terms. There are two types of co-op databases: list specific and membership.
These large aggregations contain numerous response, publication and compiled lists that have been specified as core continuation prospecting and test lists by the participating mailers. List-specific co-ops also contain the mailers' house files for suppression against their prospecting names. The mailer is not required to make its house file available to other mailers.
The owners of the prospecting lists in these co-ops are paid their data card list rental rates or the rate negotiated between the owner and the mailer. When a name is supplied to a mailer from two or more lists, credit is given on a fractional allocation basis. For example, if three list owners provide the same name, payment will be made at 33.3 percent to each.
Though list-specific co-ops are used almost exclusively for prospecting, participants who furnish their house files to the co-op often find it efficient to extract their own house file mailings directly from the co-op rather than building their customer mailings internally.
Though any list in the co-op is a part of the whole, specific lists still can be selected or suppressed. Custom response models often are deployed for use on a list-specific basis and for selections made across the database. List owners screen the participating mailers' offers in the traditional manner and can refuse rental requests that are considered competitive or inappropriate.
There are two business-to-business list-specific co-ops: MeritBase and Datawarehouse. There are none on the consumer side.
Only mailers who are members can use names from a membership co-op. To qualify, the mailer is required to submit its customer files complete with recency, frequency and monetary data - the transactional data.
The co-op aggregates and integrates all the data and creates a profile or "customer-look-alike" model with which to select names. These co-op are a "blind" source of names; that is, the member who takes names has no idea from which contributing lists the names come. List-specific selection or suppression options do not exist, and members lack the option to screen offers.
The six consumer membership co-ops are Abacus, Z24, Prefer Network, Next Action, Wiland and iBehavior. There are two BTB membership co-ops: Abacus and b2bBase.
In short, a membership co-op is a "blind, black box" option while the list-specific co-op offers known quantities with "open optional selectivity."
Membership co-op operators want all multichannel marketers to join in order to roll up all the customer files and transactional data into one massive model emanating from their "black box" on a blind, single-source basis. There are list costs and fees to mail, but no list rental or list enhancement income is paid to the list owner mailers.
List owners can block certain participants from using their data, but some membership co-ops do not provide a list of participants. To block other participants, each member must keep a list of actual or potential competitors, including small, start-up catalogs.
In contrast, list-specific co-op operators want only lists selected by the mailers to participate by providing their usual list rental files. They want the mailers using the co-op to participate with as much customer file data as needed to suppress existing customers and to assist in reactivation and retention campaigns. List rental income is paid to the list owners with standard commissions and charges. List owners and mailers have complete control and approval.
For list-specific co-ops, these include access to a variety of known lists with the ability to tailor selections with additional data and models unavailable in a traditional list order. Cost efficiencies for both prospecting and house file mailings often can be combined, and firmographic and appended data are obtained and the usual merge-purge process often is eliminated. For mailers who are also list owners, list rental income is a benefit.
A logical benefit of membership co-ops is access to transactional data as a part of the resultant model and selection criteria. There can be an average of 20 individual purchases for a single name when aggregating from all lists. More members mean more buyers and more transactions factored into the roll-up model. Pricing is 10 to 20 percent lower than average list costs of a list-specific co-op or standard rental because list owners get no rental fees.
Both types of co-ops can identify multibuyers. The membership co-op can augment multibuyer creation with transactional data to identify the most responsive multibuyers. The list-specific co-op can select multibuyers from specific lists or groupings of lists.
There have been three concerns for list owners about list-specific co-ops: security of the list, putting the list on the market and the effect of list-specific co-op databases on list rental revenue.
Security, though important, has not proven an issue in the nearly 20 years of list-specific co-ops. Data security measures on incoming files have ensured that files are handled with appropriate safeguards. Automated validation tables control each mailer's access to each list in the database. This is a proven technology used by both major list-specific databases.
Participation in a list-specific co-op does not require that a mailer's list be on the market, and if it is, does not require that the list be available for rental in the database. The list owner can control this. Should a list owner choose to contribute its file to generate list rental revenue, both major databases have hundreds of mailer/offers mailing from them, usually resulting in more list rental revenue opportunity for lists as compared to not participating in the databases.
Concerns about membership co-ops also start with security. Again, the industry is sensitive to these issues, and the experience has been that files are handled with a level of security commensurate with the value of the files.
The other concern involves sharing of the direct marketer's most valuable asset: the customer list with all the recency, frequency and monetary data. The blocking schemes can prohibit data from being used by a competitor and they do work, when used correctly.
Sharing RFM data at all gives pause to many marketers. Sending this most precious of all data out of house to be used in a "black box" is a serious decision. Many strain, as I often do, to see the logic in a model whereby I provide the raw materials for free in order to access rental names later at a rate little different than response lists that would not require me to provide any data at all. Deep thought must be given when contemplating contributing RFM data into a processing model that uses it not only for selecting names for other non-members, but for producing other database products as well, especially when the revenue is not shared with the list owners.
I began by expressing concern about the pressure for a shift in the logic of the list aspects of multichannel direct marketing. The growth of the blind, single-source membership co-op model could have serious and unintended consequences for direct marketing.
Logic tells me that the largest possible choice of response lists from which to choose is more logical than a single source of "mystery meat" names, no matter how good the model. Further, logic tells me that DMers having the freedoms to control their list choices and circulation destinies through a free market is more logical than having their futures dictated by a "black box" oligarchy.
And logic also tells me that giving up list rental income to the oligarchy's black box is tantamount to giving up an important part of my net earnings (read: EBITDA and corporate valuation).
And other potential effects slam into the Great Wall of Logic. Membership co-ops use models that isolate what I call the "Super-Buyer." These are super-frequent buyers of unknown downstream value who dominate category-specific buying, such as computer peripherals or office supplies. However, membership co-ops are less effective at identifying the two- or three-time "evolving" multi-buyers. They tend to be mediocre in identifying niche or vertical market customers, and there is no effective method for targeting new businesses and new movers.
But of greatest concern to me are mailers who find themselves caught up in the cost and simplicity benefits of the membership co-op. They begin to overweight their prospecting in favor of the black box and are seemingly rewarded with a response rate lift. Many of these mailers are oblivious to what is happening downstream to the lifetime value of their customer base. The "Super-Buyers" inevitably disappoint on this front.
This is a question of the future lifetime value of your customer base. The only way to satisfy the logic is to do head-to-head lifetime value comparisons of list-specific and membership co-op sourced customers: known and controllable list selections versus unknown and uncontrollable list selections.
But lifetime value is a bit of a moving target for most direct marketers. Less than 30 percent of DMers do an accurate lifetime value calculation. Of those that can, they are finding that the membership co-op model often produces a group of buyers who are not strong repeat buyers. Why? They are Super-Buyers from many catalogs and tend to be less loyal (or they may be from less traditional BTB sources and have lower demand potential).
Growing evidence shows that the traditional list approach and the list-specific co-op are better at producing high frequency buyers, niche buyers and overall more loyal buyers with a greater lifetime value. The restrictive, black box membership model has an inherent fatigue factor that influences the quality of the customer. Response rates may look good upfront, but lifetime value corrodes over time. Logic tells me to retain the control over the quality of my future. When I hand too much of it over to the black box, I no longer am in control.
Plus, the more clients that mail from the same universe, the more incestuous the pool becomes. As more mailers mail from the membership co-op pool, the percentage of new sites available for productive prospecting steadily decreases. The long-term health of every consumer and BTB mailer requires constant discovery of new sources of names to add to the prospecting pool.
Another issue emerges in this shift of quality control. Aside from sacrificing list rental income, the initial single-source black box pricing may seem attractive for catalogers with large prospecting budgets, but the near-total control rests with the membership co-op operators. All future pricing is in their control because they are managing a group model, not an individual client model. Sorry, Dad taught me never to give up price control.
What leverage would I have in the membership model if the lifetime value began to tank, especially if I were locked into a multi-year contract? And if I did have to revert to controlling my own destiny and return to using individual lists that I select based upon my own list-level analytics, intelligence and common sense, I likely would go through one to two years of intensive, expensive re-testing because I had to give up all of the list-specific response knowledge while participating in the single-source black box.
And that spotlights a serious flaw in the membership co-op model: loss of knowledge and flexibility. First, the source of the names spit out by the model is unknown to you. Is this logical for the long term? If you give up the intimate, hard-won and immensely valuable knowledge of your expensively tested, list-by-list performance, you are effectively giving up the temperature, pulse and respiration as well as all the lab testing that your doctor relies on to maintain your health.
When you know the intimate details of core list performance, through careful list testing and selection and through careful and analyzed continuation strategies, you understand not only what is working, but why it is working. What do you know with a single-source, black box approach? Some of the lists are direct response, some retail and some are from the Internet. How do you specify channel selection to see which buyers from which channel test better? Future mailing and online campaigns will require in-depth channel selection. How will you do that with the "blind" approach to prospecting? And how can you execute "combo" campaigns that alternate postal with e-mail? You can't.
The Proof Is in the Pudding
There are a few things I believe are logical (and fair) for every multichannel direct marketer to ask:
• Who, specifically, is using membership co-ops successfully? Are there really only a small number of large mailers who fit this model?
• What are the short- and long-term results?
• How long do members participate, and why do they stay or leave?
• Can you show me a uniform suite of metrics that fairly compares the results of the membership co-op versus the list-specific co-op or traditional list rental?
• How well did the member companies perform prior to and after joining the membership co-ops? This would be an "oranges and oranges" EBITDA, repeat purchase measurement and customer lifetime value calculation.
My skepticism tends to be fed from observing the results of shifts in the prospecting models of many consumer and BTB catalog companies in the past 12 years. In my experience, more companies try the membership co-ops and abandon them or scale back their use.
I am not flatly against the use of membership co-ops. These names can be a productive addition to the core continuation prospecting pool, but only after they are properly tested and the lifetime value metrics are understood. Then, mailers can deploy these names, along with the list-specific core continuations, each in proportion to the contribution they can be expected to make to the acquisition and retention objectives - short and long term - producing a balanced strategy.
These determinations cannot be based on simplistic list cost benefits or initially promising response rates. They must be based upon the productive yield, over time, of every list source right down to the keycode level. I would approach the membership co-op like this:
• Understand the model, how it was built, the goal and especially the potential rollout universe. If testing 25,000 names in a segment leaves a rollout universe of only 13,000 prospects, that is not a desired result.
• Use membership names that are incremental to the list-specific portion of the mail plan. This provides a clear answer to the question, "What am I getting in return for providing my RFM data?"
These two qualifiers would lead to a more reasonably successful prospecting approach with better longer-term results. The model would produce fewer prospecting names, but they would be unique names better serving my constant need for universe enlargement. This approach also would introduce some customer leverage in pricing that might make up for all of the illogical "give-ups" that the membership co-op oligarchy commands. In other words, maybe there should be something in it for me, the cataloger, too.
If left unquestioned, membership co-ops might come to dominate the long stable, long profitable DM list world, and we might all be mailing long-term less-profitable names that somebody somewhere (without our input or consideration) justifies by saying, "Mail these."