Faced with new customer types from search and affiliate programs, lower direct-buyer conversion rates, and higher overall costs, multichannel marketers are continually evaluating new strategies and techniques to maintain and improve acceptable levels of success. In the quest for the “next big thing” in marketing, many companies often overlook proven strategies that have the potential to greatly improve marketing performance.
Nevertheless, many marketers rely solely on recency/frequency/monetary value (RFM). The reason? RFM is seen as more familiar and less intimidating than custom modeling. Also, custom models have an unwarranted reputation as an expensive solution available only for larger, more sophisticated marketers.
In reality, custom modeling can be an affordable yet powerful tool for shedding light on customer behaviors and preferences and offering powerful—and profitable–insight on improving customer relationships.
Here’s a simple six-step strategy for benefiting from custom modeling:
1) Define your marketing goals. Well-defined short-term and long-term goals serve to clarify your priorities and resource-allocation decisions. Specifically, they help you determine which models to use to deliver the needed customer insights.
Each goal should have a quantifiable metric to determine a model’s overall effectiveness. Some of the more common industry-standard goals for measuring the success of a contact strategy include: • increased profit contribution per customer • increased repurchase rates • improved customer acquisition rates • reduced marketing costs • increased reactivation rates
2) Develop a master marketing schedule. A comprehensive master schedule combines all marketing efforts into one document that functions as your blueprint for all marketing activities.
One of the more critical elements of this blueprint is your master contact plan, which organizes by month all of the activities for each campaign across all communication channels—mail, e-mail, telemarketing. The blueprint also requires you to separate prospecting efforts from existing-customer strategies and to determine the available population to target with each mode of communication.
3) Determine lifetime value potential. Many mailers determine list cut-offs based on performance per campaign, with little regard to the ongoing value of each customer. Repeat buying rates differ dramatically by list, and you should consider the lifetime value potential of each name when deciding which ones to mail.
Lifetime value studies consider acquisition cost and subsequent buying behavior to determine the total expected worth of new customers by list source. Mailers can then reallocate their prospecting budget toward those lists most likely to generate the greatest return over time—leading to a more robust database in the future.
Many mailers agree that lifetime value is important. Yet many overlook lifetime value in favor of short-term returns from prospecting. The most effective contact strategies balance long-term gain with short-term returns and the positive the effects on cash flow, marketing budget and corporate goals. The rationale for this shift in emphasis from lifetime value to “long term” value may be simply to break even within 12 months of the first purchase. You can then vary prospecting cut-offs by list source to maximize the circulation plan and reach your growth goals sooner.
4) Segment the marketing database through predictive modeling. With your contact calendar in place, you can begin to allocate available budget to the “contact worthy” population identified through predictive modeling. Depending on the number of broad segments within your marketing database, you may need multiple models. Retailers typically use models to predict behaviors such as response, revenue potential, and likelihood of conversion. To supplement the predictive strength of models, many companies are turning to cooperative databases to provide additional data overlays.
The output of the modeling process yields many valuable insights. Profiles of key segments can create a baseline of your current contact plan success by group. You should also test variations to the current plan to confirm the validity of your results.
5) Define the contact strategy. An effective contact strategy includes multiple components. You can apply the insights from your models to improve marketing performance by segment. Understanding who your customers are, their preferences, and their shopping habits are all key components of an optimum contact strategy. You should select contacts according to purchase behavior and supplement your selection to meet the total consumption requirement for the competitive category.
Cross-sell opportunities should emerge from the combined picture of what your clients buy from you and what they buy from others. Lifetime value potential derives from your own data, supplemented by the ability to spend in your category. Original source, channel preferences, and response history to particular offers and content should drive your choices regarding offer, message, and medium. Channel preferences also guide the cycle of communication, including the number of times to mail and the interval between targeting.
6) Analyze results and return on investment. Effective contact strategies need proper testing. You can gauge what works best for each segment only by monitoring what has worked over time. Historical patterns can guide your initial approach, but segment-level testing is the critical element in determining the optimum strategy.
Baseline analysis can help you determine the effectiveness of you current contact strategy prior to testing and implementing new plans. Variables to examine can include:
• average number of times contacted (direct mail vs. email vs. telemarketing) in the past year • average spending • average number of store visits • interval between purchases
Your analysis should provide the answers to the following questions:
• How are you allocating your marketing budget across major database segments? How many times are you contacting your best/worst customers given their purchase behavior? • Are you sending mixed messages to your customer base? Do your full-price campaigns overlap with your discount e-mails? • Are you hurting your profit margin by overmailing certain segments?
This type of analysis provides some directional information. Granted, some groups merit more mail than others, and modeling should clearly identify those groups. However, testing is the only way to measure the true cause-and-effect impact of different strategies.
Network
