Imagine That a Multivariate E-Mail Test Has Been Conducted by the American

Imagine that a multivariate e-mail test has been conducted by the American Savings Association (ASA).The results are shown below.The five manipulations appear as the column headers while the data beneath each column notes the contribution (to click-through)made by that manipulation when it is present or absent. $$\begin{array} { l c c c c c } \hline & \begin{array} { c c c c } \text { Recipient } \\\text { Personalization }\end{array} & \text { Images } & \text { Subject Line } & \text { From Address } & \text { Length } \\\hline \text { Positive } & \text { (present) } & \text { (yes) } & \text { (short) } & \text { (personalized) } & \text { (short) } \\\text { Contribution } & + 7.6 \% & + 3.4 \% & + 2.2 \% & + .2 \% & + 2.9 \% \\\text { Negative } & \text { (absent) } & \text { (no) } & \text { (long) } & \text { (ASA) } & \text { (long) } \\\text { Contribution } & - 10.2 \% & - .2 \% & - 2.2 \% & - .1 \% & - 3.1 \% \\\hline\end{array}$$
Mary and Peter look at the results.Mary says that "Our next e-mail should not have recipient personalized,but should have images,a short subject line,a personalized from address,and be short in length." Peter says that "Our new e-mail should have recipient personalization,and it should not have images.In addition,the e-mail should have a long subject line,a personalized from address,and be long in length." Who is making the better recommendation? Why?