Using Design of Experiment Procedures for Direct Marketing

Improving E-Mail Response Rate Based on the GLM analysis, the company can conclude that emails sent with detailed headers and text bodies are opened with greater frequency than emails sent with generic headings and HTML bodies. General Linear Model Between-Subjects Factors HEADING Detailed Generic OPENED HTML Tests of Between-Subjects Effects Dependent Variable Type III Sum of Squares df Mean Square F Sig.

Partial ETA Squared Corrected Model TRIAL1 TRIAL2 Intercept TRIAL1 TRIAL2 HEADING TRIAL1 TRIAL2 OPENED TRIAL1 TRIAL2 TRIAL1 TRIAL2 HEADING * OPENED TRIAL1 TRIAL2 HEADING * BODY TRIAL1 TRIAL2 4.500 1 4.500 1.000 OPENED * BODY TRIAL1 1 1.000 TRIAL2 1 1.000 HEADING * OPENED * BODY TRIAL1 66.125 1 66.125 1.000 TRIAL2 32.000 1 32.000 1.000 Total TRIAL1 12943.000 TRIAL2 16140.000 Corrected Total TRIAL1 TRIAL2 7 a. R Squared = 1.000 (Adjusted R. Squared = .) Estimated Marginal Means 1. HEADING Estimates Dependent Variable HEADING Mean Std. Error 95% Confidence Interval Lower Bound Upper Bound TRIAL1 Detailed 35.750 Generic 37.000 TRIAL2 Detailed 45.750 Generic 36.750 Pairwise Comparisons Dependent Variable (I) HEADING (J) HEADING Mean Difference (I-J) Std.

Error Sig.a 95% Confidence Interval for Differencea Lower Bound Upper Bound TRIAL1 Detailed Generic -1.250 Generic Detailed 1.250 TRIAL2 Detailed Generic 9.000 Generic Detailed -9.000 Based on estimated marginal means a. Adjustment for multiple comparisons: Least Significant Difference (equivalent to no adjustments). Univariate Tests Dependent Variable Sum of Squares df Mean Square F Sig. Partial Eta Squared TRIAL1 Contrast 3.125 1 3.125 1.000 Error .000 0 TRIAL2 Contrast 1 1.000 Error .000 0 The F tests the effect of HEADING. This test is based on the linearly independent pairwise comparisons among the estimated marginal means. 2. OPENED Estimates Dependent Variable OPENED Mean Std.

Error 95% Confidence Interval Lower Bound Upper Bound TRIAL1 No 31.750 Yes 41.000 TRIAL2 No 33.750 Yes 48.750 Pairwise Comparisons Dependent Variable (I) OPENED (J) OPENED Mean Difference (I-J) Std. Error Sig.a 95% Confidence Interval for Differencea Lower Bound Upper Bound TRIAL1 No Yes -9.250 Yes No 9.250 TRIAL2 No Yes -15.000 Yes No 15.000 Based on estimated marginal means a. Adjustment for multiple comparisons: Least Significant Difference (equivalent to no adjustments). Univariate Tests Dependent Variable Sum of Squares df Mean Square F Sig.

Partial Eta Squared TRIAL1 Contrast 1 1.000 Error .000 0 TRIAL2 Contrast 1 1.000 Error .000 0 The F tests the effect of OPENED. This test is based on the linearly independent pairwise comparisons among the estimated marginal means. 3. BODY Estimates Dependent Variable BODY Mean Std. Error 95% Confidence Interval Lower Bound Upper Bound TRIAL1 HTML 21.750 Text 51.000 TRIAL2 HTML 28.750 Text 53.750 Pairwise Comparisons Dependent Variable (I) BODY (J) BODY Mean Difference (I-J) Std. Error Sig.a 95% Confidence Interval for Differencea Lower Bound Upper Bound TRIAL1 HTML Text -29.250 Text HTML 29.250 TRIAL2 HTML Text -25.000 Text HTML 25.000 a.

Adjustment for multiple comparisons: Least Significant Difference (equivalent to no adjustments). Univariate Tests Dependent Variable Sum of Squares df Mean Square F Sig. Partial Eta Squared TRIAL1 Contrast 1 1.000 Error .000 0 TRIAL2 Contrast 1 1.000 Error .000 0 The F tests the effect of BODY. This test is based on the linearly independent pairwise comparisons among the estimated marginal means. 4. HEADING * OPENED Dependent Variable HEADING OPENED Mean Std. Error 95% Confidence Interval Lower Bound Upper Bound TRIAL1 Detailed No 28.000 Yes 43.500 Generic No 35.500 Yes 38.500 TRIAL2 Detailed No 35.000 Yes 56.500 Generic No 32.500 Yes 41.000 5.

HEADING * BODY Dependent Variable HEADING BODY Mean Std. Error 95% Confidence Interval Lower Bound Upper Bound TRIAL1 Detailed HTML 20.500 Text 51.000 Generic HTML 23.000 Text 51.000 TRIAL2 Detailed HTML 32.500 Text 59.000 Generic HTML 25.000 Text 48.500 6. OPENED * BODY Dependent Variable OPENED BODY Mean Std. Error 95% Confidence Interval Lower Bound Upper Bound TRIAL1 No HTML 23.500 Text 40.000 Yes HTML 20.000 Text 62.000 TRIAL2 No HTML 29.500 Text 38.000 Yes HTML 28.000 Text 69.500 7. HEADING * OPENED * BODY Dependent Variable HEADING OPENED BODY Mean Std.

Error 95% Confidence Interval Lower Bound Upper Bound TRIAL1 Detailed No HTML 22.000 Text 34.000 Yes HTML 19.000 Text 68.000 Generic No HTML 25.000 Text 46.000 Yes HTML 21.000 Text 56.000 TRIAL2 Detailed No HTML 32.000 Text 38.000 Yes HTML 33.000 Text 80.000 Generic No HTML 27.000 Text 38.000 Yes HTML 23.000 Text 59.000 1 2. The graphical display charts for this model are estimated marginal means of the two trials runs; the plots are included below. The profile plots show the relationship between the email body type (HTML or text) and the email heading type (generic or detailed).

There is a clean distinction between the estimated marginal means of the two style choices for the emails. TRIAL1 HEADING * BODY * OPENED TRIAL2 HEADING * BODY * OPENED The emails with detailed headers are opened at greater rates than the emails with generic headers. Moreover,.