Wednesday, May 20, 2009
Multichannel Marketing Class May 19: Direct Mail and Print
This class on direct mail offers is always one of my favorites to teach. It is intriguing to go over different direct mail offers and see why and how they were constructed the way they were. An offer should always ask for a specific action, have a specific deadline and tell me what I am going to get if I respond by that deadline. As direct marketers we are direct RESPONSE markters. We use a database to identify and track our customers, communicate with them in a way that asks for a specific action and we measure the results so we can expand the communications program if it is successful. We looked at two direct mailers that I had received, one from Travellers insurance which made an offer to lower my costs for auto insurance. We thought the package was reasonable (it had a letter with an offer restated several times, a P.S., a lift slip and a response card) but maybe not personalized enough because it did not express how much I could personally save. Our guest Cyndi Greenglass from Diamond Marketing Solutions reinforced that we use "you" language in persuasive communication because it does involve the readers and get them closer to taking the desired action. The mailing from Mazda was attractive but the offer was buried and we were given a choice, which can be a death knell for offers. People start thinking and then they stop buying. Both our guests talked about the importance of mathematics in modelling. Michelle from Abbott Labs demonstrated how Abbott isolates the results of various media through response modelling. Cyndi Greenglass from Diamond talked about testing principles for direct mailings, including establishing a control and testing one variable at a time to improve on the results of the control. She gave an example from Pitney Bowes where they kept testing to try to improve the control package. I asked students at the end to think about the problem in the Newport News case in the text. With 40% of orders coming from the web, Newport News was having a hard time isolating the effects of their marketing because they could not use the 'keycodes' associated with catalogues to track response. What should Newport News do to handle this problem?