How to Improve Upon Generic Cooperative Database Response Rates
February 2, 2015
5 Tips from the New Best Practices for Circulation Management
October 5, 2014
Measuring the Incremental Value of Catalog Mailings
March 28, 2015
Matchbacks and the Evolving Science of Attribution
November 5, 2014
A catalog generation ago before the Internet, catalog orders came in through the mail or the phone. Source code capture rates of 85% were the norm and it was easy to read the results from each mailing list. Then along came the Internet and measuring catalog response rates became complicated. People ordered on the Internet and now the percentage of orders taken over the Internet continues to grow. But the Internet is not just a channel for taking orders. Demand is being created by an ever increasing variety of web marketing. So the web has evolved into both an ordering channel and a series of ways to create demand. The growth of the web has made the attribution of orders very complicated. Customers receive a catalog, e-mails, on line ads and a robust web site. Knowing which marketing program should get credit for an order is a challenge.
Here are some of the issues with allocating orders in a multichannel marketing environment:
House list segments are typically allocated too many sales in a catalog mailing match back. The match back methodology is to match the mail file from a catalog against the orders received during the life of the catalog. So if you had previously bought from a catalog and were mailed a catalog, then all orders you placed during the life of the catalog are credited to the catalog. Testing shows that a certain portion of a catalog company’s customers will order whether they receive a catalog or not. And a certain portion of your house buyers are responding to the web and to e-mails. A solution to this over allocation is to have mail/no-mail test panels to measure the sales that a house file segment gets if no catalogs are mailed versus the same segment’s sales when they receive a catalog.
Prospecting circulation sales from rented lists that have never ordered from your catalog are typically under reported because matching the names mailed against orders receive omits catalogs that were passed along and used by people at different addresses. Pass along orders could easily top 20% of the reported sales from a prospecting mailing list.
The default methodology for match back allocation has been to allocate the catalog order to the last catalog mailed. But marketers are also using multiple campaign order curves to allocate orders between campaigns when several catalogs overlap. This best example of the need for a multiple campaign order curve allocation is when a cataloger mails two or three holiday catalogs in November and knows it is not correct to allocate all the December sales to the last catalog mailed.
Match back methodology typically includes the ability to subtract out all orders that come with promotional codes for e-mails, web promotions, on line ads, etc. So if a customer uses an on line promotion requiring a source code then that promotion gets full credit for the sale and the catalog mailing list gets no credit.
On line ads get credited to the on line ad if the customers clicks all the way through and makes a purchase. On line ad providers typically cookie a catalog’s web traffic and claim some percentage credit of the sales from that household. Only testing can zero in on the attribution percentage and tell whether the on line ads should get 5%, 10% or 30% or more of the credit from on line ads to households that have had multiple touches from catalogs and web visits.
Match back metrics also show the number of orders not matched to a mail file and this is a good indication of the demand that is coming purely from the web. This is a very useful metric over time because it will show how much demand is building from various web programs over time. Catalogers are seeing the orders that can’t be matched to the mail files growing as a percentage of total sales as the web creates a larger portion of the total demand.
Can you estimate web sales from promotions that don’t have promotional codes for tracking? Review the increase in web traffic and sales in the hours and days after an e-mail blast or a web promotion and, if you see a spike in sales, allocate a reasonable proportion of those sales to that touch of the customer.
Mail versus no mail holdout panels are becoming a more accurate measurement of the incremental sales coming from a catalog mailing. Mail versus no-mail holdout panels can also tell a truer picture of sales coming from all varieties of web promotions from on line ads, to e-mails, to multichannel campaigns. Marketers should know that using hold out panels is the default answer for measuring campaigns where allocation of the orders is an issue.
Know that if multiple channels are claiming more credit for sales than the total for all sales that occurred, then you have some double counting. Know that the web and the catalog can’t claim full credit for the same order. Drill down to the order level and see where the overlap occurs and set some business rules to allocate credit for orders claimed by more than one channel or promotion.
Look closely at where your new to file customers are coming from. Catalogers are looking for any and all ways to bring in new customers. Prospecting for new customers with catalogers remains the primary way to bring new customers. But on line ads have shown success at finding new customers. Know which web programs can profitably bring new customers because it is critical to find new customers.
The biggest lesson multichannel marketers are learning from a granular allocation of their orders is that catalogs remain the primary driver for sales for most catalogers. The web is a great vehicle for harvesting demand and getting incremental business from your house file. But the catalog is typically still the main driver of sales and the most cost effective means for a catalog to acquire new customers.
Please don’t make the mistake of assuming that because the percentage of orders that are placed using the web is ever increasing that the web is creating that demand.
The web is an efficient and inexpensive way to take orders. So the percentage of orders taken on the web continues to increase. But the web still has a limited ability to create demand. So it is critical that marketers understand both the match back tools and their limitations and all the metrics available to measure web traffic in order to know how to accurately allocate credit for orders in a multichannel world. Order allocation will only get more complicated as more and more profitable web programs evolve and are layered on top of catalogs, your web site and e-mails.
Understanding the impact, profitability and interaction of a variety of off line and on line marketing programs is the key to maximizing profitability.
Here are the takeaways from this article:
Match back results from house lists of buyers are overstated because your house file gets more contacts than just the catalog.
Match back results from prospecting lists are understated because of the pass along value of the catalog.
Use a multiple catalog allocation to allocate sales between books when in home dates are tightly bunched together like in Q4.
Subtract out the orders that come with promotion codes for web marketing from your catalog results.
Credit on line ads with click through orders but test to determine the correct attribution percentage for view based orders.
Use your match back’s orders not matched to the mail files as a metric for growth of your pure web business.
Estimate the sales from on line promotions from the increase in daily sales.
Mail versus no-mail holdout panels tell precisely how much incremental sales a catalog delivers.
Know where the overlap occurs when multiple marketing channels claim the exact same orders.
Know where your new-to-file customers are coming from. If the web is delivering new customers, learn how to scale those web programs.