We have tried to look at this relatively close to reality. Hence, we distinguish 8 phases that tend to - but are not necessarily – sequential, 12 different interaction channels, and 9 different means of interaction (how the interaction happened, e.g., via telephone or online). When considering all these options, basically no two journeys are the same, which calls for a meaningful segmentation of customers with regards to their interaction preferences.

Based on a large qualitative study we ran on this exact topic as well as the results of descriptive statistics, we have identified a segmentation methodology that distinguishes 4 segments based on customer orientations/goals.

What we would like to prove is that this is a decent way to segment customers (from descriptives we know that it is better than observable behavior and demographics). But what would be the best statistical test to do so?

Thanks so much for any recommendations! ]]>

"Can we say that the number of hours of Internet use increases with Internet purchase? This influence of Internet purchase on the number of use can it be explained by an attitude becoming more favorable to as we know Internet? "

my variables are:

- Use of Internet per week (Use in hours)

- Internet attitude (Attitude_Internet 1 = very unfavorable, 7 = very favorable)

- If respondent has already made a purchase on the Internet (1 = yes, 2 = no)

I have to do a mediation in SPSS in 4 steps

(Variable z = mediator = Attitude_Internet

Variable x = independent variable = Purchase_Internet

Variable y = dependent variable = Use_Internet)

Step 1: The IV should predict the DV

Step 2: IV should predict the mediator

Step 3: the Mediator should predict the DV

Step 4: The effect of the purchase on the Internet on the number of hours of use will be increased when taking into account Internet attitude

But how can I do the 4 steps ? Should I use regression or anova ?

Thank you ]]>

Particularly, we have distinguished this process into 8 phases that tend to - but are not necessarily - sequential. Also, not all customers experience all 8 phases, leading to only 1/3 of the participants (total of 3,015) to answer all questions.

For each of these phases we have asked participants for their main interaction partner („channel“, 12 different options) and means of communication used (how the interaction happened, e.g., via telephone or online, in total 9 different options). For some combinations between the two, e.g., website (=channel) and PC/laptop (=means) high correlations exist, so that these two variables cannot be assumed to be independent.

Based on a large qualitative study we ran on this exact topic as well as the results of descriptive statistics, we have a number of strong hypotheses on the factors that influence customers‘ interaction choices (channels and means), such as what interactions customers picked in the previous phase, how they liked previous interactions, the extent of their product/service knowledge etc.

Now, I am unsure about the correct statistical test to use. Any recommendations would be much appreciated! ]]>