(Multichannel Merchant) Is your business ready for data-driven marketing? Did you know that data-driven marketing is nimble enough to fit even multichannel merchants with annual sales of less than $5 million? A little known secret is that data-driven marketing works for any size organization.
The giants of our industry are often gazed upon as having all the latest analytical tools plus desktop data access, operational support, and endless marketing dollars. Yet these organizations face the same issues as their smaller counterparts: a lack of keycode (a.k.a. source code) capture, system constraints, and how best to evaluate marketing efforts.
If these issues exist for the giants in our industry, then what can smaller merchants do to renounce data defeat and conquer information? While there are as many approaches as there are companies, perhaps one of these examples will inspire you to look at alternative ways of deciphering data and spend marketing dollars.
A common complaint is that “keycode reports aren’t telling us anything.” In the early 1990s it wasn’t unusual for catalogers to report a 90%-95% keycode capture rate. Today multichannel merchants glumly share 50%, 60%, even 70% default rates — in other words, they have no source codes for as many as 70% of their orders.
Certainly one of the most talked-about ways to acceptably allocate orders back to a mailing, an e-mail campaign, or any other marketing effort is to perform a matchback. As the name implies, all orders during a specific time period are matched back to a particular marketing effort, thus identifying the response to that effort. (See “The latest matchback tools,” April 2005 issue, for more on the topic.) But for smaller merchants that cannot afford to use an external database vendor or do not have the expertise internally, what other choices exist?
The answer often lies in recognizing what you want to learn. One multichannel merchant with sales of less than $3 million stated it simply, “For the money I spend on marketing, is it worth it?” First the company asked its part-time IT person to look at its house file to count how many customers were first-time buyers in 2005, sorting them by month (referencing first-time buyers in January as Group 1, first-time buyers in February as Group 2, etc.) and tallying their sales. Then it looked at the marketing efforts to determine how many catalog and e-mail contacts each group received for the whole year. The third step was easy math: How much money was spent on Group 1 contrasted with how much revenue Group 1 generated?
As with any other analyses, this method had shortcomings. The company chose to look at calendar-year revenue vs. monitoring a rolling 12-month period. Nonetheless, it found actionable information and modified its marketing promotions, particularly in latter calendar months, to increase the number of first-time buyers who made a second purchase. For a giant in the industry, this may be an oversimplified exercise, but for the smaller merchant it provides meaningful, directional information.
A synergistic result of this analysis motivated the merchant to make two small changes to its business rules. Now customers are prompted both on the phone and on the Web in alternative ways to reveal their keycode.
For example, if the contact center representative asks for the keycode but the customer doesn’t give one, the rep follows up with discovery questions such as “Did you receive our catalog last week?” or “What did you think of the product on the front cover?” or “Did you like the catalog and all the new products?” The reps are also collaborating with the contact center manager to help find other, unique ways of engaging customers. And on the Website, after a transaction is complete, an automated prompt in the form of a drop-down box provides several options, one of which is “include keycode here.” Customers are not required to answer, but many do, as the drop-down box visually communicates a quick and uncomplicated request.
Another way the company is improving its handle on customer buying behavior and overall Web tracking is by including a “quick order” box, a feature regularly seen on the e-commerce sites of larger, established catalogers. The quick-order box is featured on every page of the Website along with the words “enter catalog item number.” This helps customers place their orders quickly; it also helps the merchant by identifying catalog-driven Web orders. Standard Web reporting subsequently calculates how many customers use the function — and use the catalog.
Companies that already do this are also able to use Web reporting to evaluate how many items were added to the shopping cart as a result of an upsell prompt such as “If you like this widget, you might want this widget cleaning kit.” Prompts can also cross-promote clearance items.
For smaller multichannel merchants, the value of the data from Web activities isn’t just “nice to know.” When the reality of the business climate forces a decision to cut marketing dollars, the result is all too often a corresponding reduction in revenue from the Web. There will always be the delicate balancing act of cash outlay for marketing expenditures vs. cash flow planning for revenue.
This is why another multichannel merchant decided to test its hypothesis that mailing one fewer catalog during its busy season and replacing the contact with an e-mail would yield equal revenue for fewer marketing dollars. But this midsize company has a few obstacles impeding its ability to conduct such a test. It has no real database to use, just transaction data and customer records. There’s no integration between its Web and catalog operational systems; Web orders are batch-processed and manually input. There’s very little keycode tracking; the keycode default rate exceeds 50%, and there’s no keycode capturing on the Web.
If this were you, would you give up? Never say die. Instead you should do what a giant in our industry would do: collaborate with your vendors.
Start by setting up a meeting with your mailing house, your e-mail vendor, your IT person, and your marketing person. The marketing person presents the idea for the contact test: Of the customers who would ordinarily be chosen for the seasonal catalog mailings, some will receive three mailings, while the others will receive two mailings and one e-mail message.
The mail house indicates that it has a small data programming staff and could help by matching data, suppressing certain names from the mailing, and providing general analyses. The e-mail vendor suggests giving the mail house all e-mail accounts so that it can match the information to the universe of customers. The marketing person says that the test group will be the 12-month customer group. The IT person says that he can provide the mail house with the customers whose most recent order occurred within 12 months.
The next steps are isolating the details and structuring the mailing. Have the mail house determine how many of the customer names from IT can actually be matched to an e-mail address — not just an e-mail address, but an opt-in address that we can legally send a promotional message to.
The marketer identifies which segments will receive three catalogs vs. which will receive two catalogs and the e-mail as the third contact. The mail house will hold the mail tapes and receive response data monthly to identify purchase behavior.
After the season of testing is complete, the data will reveal if the reduced advertising investment affected revenue. The midsize merchant I’d referred to earlier that is conducting just such a test is anticipating that the data will offer at least directional information and most assuredly will prompt requests for more testing, more data, and more analysis…and perhaps the need to invest in a database.
If your company, like the one in our example, doesn’t have access to a database, you’re hardly alone. Smaller catalogers often rely on their mail house or data processing vendor for data-related functions. One smaller cataloger needed to better understand how often its customers were being contacted by its five product managers. Because each product manager asked IT for a mailing list and then used his own favorite mail house, the data weren’t centralized. There was an underlying suspicion that too many customers were receiving every single mailing from the company whereas others were being undermailed.
Certainly the marketing intent was good. Each manager was asking IT for the “best names,” and IT wrote a program to provide them. But no one was looking at overall contact strategies or strategically marketing to the customer file.
The cataloger’s solution was to work with one of the mail houses to use the 36-month customer file as the basis for analysis. Each of the product managers had his mail house send one year’s worth of mail tapes for analysis. The purpose of the project was to find out which customer groups were receiving which mailings. The results were presented in a simple cross-tab matrix with customers (rolled into segmentation groups) down the Y axis and the many promotions across the X axis. This report quickly revealed both undermarketed and overmarketed segments.
More contacts to best customers doesn’t necessarily mean unprofitable; the opposite is often true. But the point of this exercise was to reveal which customers were being neglected. One of the next projects for the cataloger is to overlay sales data, but it’s taking things one step and one expense at a time.
As you ready your marketing strategies for the upcoming season, start thinking about how you’re going to evaluate the success of your efforts. Even if you don’t have all the answers, pose the question to your internal and external teams. Data-driven marketing is for everyone.
Gina Valentino is the owner of Kansas City, MO-based consultancy Hemisphere Marketing.