Crawl, Walk, Run With Optimized Data
How does it work? An optimized data guidance system increases the precision of your data at minimal cost, letting you boost response by leveraging the behavioral and attitudinal aspects of your data to individually personalize message and creative. Data optimization provides continually intensifying levels of sophistication that can increase direct marketing ROI. Like learning to crawl before you walk, and walk before you run, a carefully orchestrated plan must follow a logical progression. Constructed and painstakingly followed over six to 18 months, it requires a disciplined approach that aligns all organizations (client, database marketer and agency) under one cohesive plan.
Beginning with a crawl, your organization must be able to articulate a unique business model and differentiated products and services. Be sure you already have identified and stratified your competitive positioning. An outside consultant with expertise in direct marketing may be helpful here.
First steps. Take your first steps by fine-tuning your target data using your customer transaction data and external data to optimize response - for example, "a decrease in customer service inquiries or transactions." This lets you select interactive data elements that could uncover high-response prospects such as "credit score sliding downward" and "increasing credit inquiries" or "change of address" and "assumption of new trade lines."
Because demographic and lifestyle variables change little from quarter to quarter, include another, higher level of information, such as transactional data, or demographic/lifestyle data and transaction data, to maximize optimization.
Gaining speed. An optimized data guidance system lets your marketing campaign take advantage of a broad mix of media and Internet direct marketing channels. Once equipped with the selects and interactions within your variables, consider mapping the density of these individuals by market. Gather comparative information, including market share on your competition, to assess your campaign's competitive performance. Also, compare product differentiators and how these are being articulated by solicitation and media category.
Up and running: response and conversion models. Response and conversion models are fundamental for any direct-to-consumer marketers seeking to scale their business. Though many list brokers claim that models don't work, this argument is largely self-serving. Brokered response lists work on the principle that past buyers are future buyers. But without enough prospects within the best of these response lists to meet ongoing needs of large-scale marketers, the typical list broker solution is to batch multiple lists, diluting quality by including leads that may be irrelevant to the client offer and depressing response rates.
However, research shows that a compiled file correctly modeled on recent responders and converters is a response list. More importantly, this response list can scale in volume and can predict the projected response and conversion rates that will be achieved with over 90 percent confidence.
Moving forward with the optimum mix. So how can you simultaneously develop the optimum mix of: offer, channel, message, package, price and cost across an 18-month test/control plan? Using a multivariate process, you can select one objective to optimize and let your other decisions be driven by this one objective, creating a marketing plan for each objective to be optimized. If your goal is to generate the highest number of customer acquisitions and you are constrained by a fixed budget, you can use a complex set of algorithms to create the optimum mix of solicitations by channel, offer, prospect sources and package. You can create any number of scenarios using any type of constraint and historical performance.
Prerequisites for success. Data optimization necessitates certain prerequisites:
* An established, disciplined marketing plan of test and control tactics across each component of your direct strategies. You must have been executing campaigns with the appropriate cell sizes of: offer, creative, message, channel and data, as well as the mixes of these cells, which must contain statistically representative sample results, not just across solicitations, but for responders and converters.
* Statistically valid historical performance data across all of these cells and combinations of cells.
* Established valid business constraints that you practice on a regular basis, i.e., managing within a budget, a stable product line, and response and fulfillment capacity.
Most of the work is in setting up the data, which often deters analytic organizations from tackling more complex projects. However, the long-term benefits will be worth the investment, delivering (ongoing) higher response rates.
Running smoothly. Here's a pre-flight checklist to keep your data optimization system running smoothly:
Design your marketing strategy (not your 12-month acquisition plan) to include: new product (or features/benefits) introductions, channel expansion plans and multiple test/control tactics over a 12-18 month period. It typically takes organizations two to three months to define new product offers, pass legal approvals, develop marketing collateral, etc., so use the 12-18 month plan broken down into the multitudes of test and control tactics as a launch model.
Establish cell sizes by projecting response and conversion volumes by each cell and combination of cells. Be conservative and ensure that your responder or converter base is representative of your target base. If this is too large, you can accumulate required volumes over time, across campaigns.
Determine what should be included in your test/control plan. Avoid over-complication, focus initially on the four to five highest performance drivers (typically offer, channel, market segment, package/creative and message). Put your best ideas into your test/control strategy to quickly accumulate and analyze performance history. Though response rates are your leading indicator, many other factors can affect conversion, including some beyond your control.
Sprinting into multivariate optimization. For organizations with data warehouses (which can provide a wealth of promotion history), following a disciplined test-and-control campaign strategy requires another tier of decisioning tools to achieve continued levels of performance boosts.
Crawl, walk, run ... fly. Within a real-life set of business constraints, deploying disciplined tactics and rigid test and control scenarios, bolstered by sufficient campaign history, can solve simultaneous equations for a single objective. Though bringing your data-optimized marketing up to full speed can take up to 24 months, you can expect to develop quantifiable techniques and realize consistent gains and increasing altitude along the way.