Working for a lean organization we often didn’t have sufficient business intelligence resources to determine the effectiveness of our marketing efforts. As the go-to analytics person on my marketing team, I used the following methodology to determine which promotions actually helped to move product.
Step 1: Two months worth of sales data (orders) was pulled for each sku in each individual promotion.
Step 2: Average orders per day was found for two months. The average orders pay day represented the baseline for each sku- it’s expected sales without any promotion.
Step 3: Average orders per day for the promotion was found.
Step 4: Totals for each average was compared and a t-test was performed to determine whether there was significant lift.
Step 5: Sales prices for each sku was multiplied by the total # of skus sold during the promotion dates.
Step 6: For outstanding skus – standard deviation was run. Skus that performed above normal ( 2 standard deviations) within the promotion event were considered star performers for that particular promotion.
Now the above methodology did not incorporate other factors that could impact performance such as inventory levels, pricing, frequency, or traffic levels to the site. But it did give us direction about our best performing skus in a quarter, popular brands, best performing promotions as well as the impact email or a sitewide or limited banner had on the performance of the promotion.