Hitmetrix - User behavior analytics & recording

Maximize Third-Party E-Mail Programs

In the past year, the mounting problem of unsolicited bulk e-mail seems to have taken the luster off legitimate e-mail marketing. A scan of recent months’ newspaper and trade publication headlines would lead one to think that e-mail marketing is on its last legs, that spam is “killing the killer app,” as so many industry figures put it.

So how did my company, CoolSavings, increase overall response to our advertisers’ e-mail programs in 2002 by more than 50 percent over the previous year?

Despite warranted concerns over spam, e-mail remains one of the most powerful direct response marketing vehicles in terms of cost efficiency and unprecedented response rates. But you have to know how to do it right.

The following checklist will help you understand the questions and answers you need to know to identify a proper third-party e-mail supplier:

Consent and deliverability. Consumer opt-in to e-mail from the supplier isn’t enough. Subscribers also should give explicit permission to receive third-party offers from the suppliers’ advertisers. Further, such offers should be sent by the supplier on behalf of the advertiser. Doing so will help create a better open rate (and therefore better response) because consumers are more likely to read messages from a trusted source.

Do not market to names collected for secondary use such as purchaser files or other purchased lists. Doing so contributes to the distrust of consumers in the e-mail marketing channel and, worse, risks a negative impression of your brand and message.

In short, anyone sending third-party e-mail to a consumer should have explicit permission to do so.

List hygiene activities such as bounce processing and subscriber opt-out capabilities are a given. Continuously mailing to bounced addresses will lower the chance that any mail will be delivered to that domain. Likewise, sending to consumers who already opted out or are having difficulty doing so will increase complaints and make your mail more likely to be viewed as spam.

Another factor affecting deliverability is the recency and frequency of mailing. A sophisticated supplier with an attention to its subscribers’ needs will dynamically control recency and frequency based on consumer responsiveness and behavior profiles. Not all subscribers have the same standards for too much or not enough e-mail. At a minimum, the supplier should have fixed standards for mailing recency and frequency that will not dilute the effectiveness of your message.

Data scale and quality. One major benefit of the e-mail marketing channel is the ability to collect large amounts of behavioral and self-reported data for an individual subscriber quickly and easily. Though demographics are obviously useful, behavioral data is generally a much stronger predictor of responsiveness and should be used whenever possible. This brings up the next point.

Your supplier should do more than just collect a lot of data. It should have a deep understanding of the value of this data in order to use a broad array of selects for targeting and analysis. Further, the supplier also needs a data warehouse that is structured to quickly access stored consumer information for use in targeted campaigns.

Data testing and targeting models. Sure, most good marketers have an idea of the demographics, geographics and some of the behavioral characteristics of their best customers. However, this isn’t always the best target to use for a third-party e-mail campaign. Remember, these consumers are the supplier’s subscribers, not your customers (yet); they are prospects who need to be enticed to first respond to your e-mail before they can become customers.

As such, there may be other important variables in your prospect profile that you have not considered. For instance, past responsiveness to offers in your category is a characteristic that often identifies future responders from non-responders within the same demographic profile. The way to uncover and use such information is through testing and modeling.

For simple explanation, the process of data testing and target modeling can be summarized as follows: Deliver a test e-mail offer to a small, random sample of consumers; analyze the results; and build a profile of the offer’s ideal target – those most likely to respond based on similar user characteristics to the individuals who responded to the test offer. The model is then projected across the supplier’s entire database to segment like consumers for a large-scale e-mail campaign.

A random test of all variables and a custom targeting model effectively test every possible combination and permutation of targeting criteria at once. Simple select or cell testing – demos, category interests, etc. – to identify a basic targeting combination that will perform well is inefficient and does not maximize the benefits of the robust data set your e-mail supplier has at its disposal.

Tried and true modeling techniques such as CHAID and logistic regression, which have been used for years offline, often provide response rates that are at least three times higher than standard selects alone, assuming that solid behavioral data is available. You will be left not only with a list of the best possible targets the supplier’s database can provide, but a performance projection for each subsequent target.

Implementation and beyond. Once an effective targeting model has been created, the hunt for performance improvement through targeting is essentially complete and you can focus on other aspects of your campaign, such as frequency and creative testing.

After targeting, creative testing has proven to be the largest opportunity for performance gains. Small changes in subject lines and offer propositions can more than double response, which often means that you can afford to mail to more targets from the targeting model. Keep in mind, creative testing should not be viewed as a one-time or even a quarterly exercise, but rather an ongoing effort to refine and keep creative fresh.

Mailing frequency is another important area for testing, as the optimal frequency may vary based on your message and the composition of the supplier’s list. However, many marketers underestimate the potential for repeated contact and the dynamics of most lists.

A well-built targeting model will take into account previous contacts and will refresh with new subscribers on a regular basis without burning out the list.

Total
0
Shares
Related Posts