Use the following average number of impressions by each tot_impr_category range to figure out the cost to run the ads by each sub-segment:

Now that you have the average impression by category information and can translate that into the advertising cost per category in your analysis, which total impression category ranges appear to provide the highest ROI? Remember to only perform this ROI analysis on impression category ranges where the test group results are significantly greater than the control group results at a 95% confidence level (results of step 4B). Also, just as you did in step 2A, only include $ margin in your ROI calculation that is incremental to the baseline $ margin that would have been achieved without the use of the handbag display ads

\begin{tabular}{|l|l|l|} \hline Tot_impr_category & \begin{tabular}{l} Range of Impressions Shown per \\ User \end{tabular} & \begin{tabular}{l} Average \# of Impressions Shown per \\ User \end{tabular} \\ \hline 1 & 1 ? 20 & 7.84 \\ \hline 2 & 21 ? 40 & 28.56 \\ \hline 3 & 41 ? 60 & 49.25 \\ \hline 4 & 61 ? 90 & 73.31 \\ \hline 5 & 91 ? 130 & 107.50 \\ \hline 6 & 131 + & 227.88 \\ \hline \end{tabular}