Analytics & Data

Ecommerce merchants typically run promotions throughout the year. Some are seasonal, such as “Free Shipping on Valentine’s Day Flowers” and “20% Off for Cyber Monday.” Others focus on product categories, such as “Free Socks with Purchase of Any Jeans.” Still others are sitewide, such as “Free Shipping on All Purchases.”

But how does a merchant know if a promotion was successful? That’s the purpose of this article.

### Marketing ROI

The most common way to analyze a promotion is to compute the Marketing Return on Investment.

MROI = (Gross Profit – Marketing Investment) / Marketing Investment

Marketing Investment should include the discount itself and the cost of running the promotion.

Here is an example.

Your company just completed a promotion for 20-percent off all t-shirts. Before the promotion, your average t-shirt selling price was \$29.99, and your average t-shirt cost was \$9.99. You spent \$1,000 advertising the t-shirt promotion using Facebook ads. The average discount per t-shirt was \$6.00 — \$29.99 x 20%. After running the promotion for one week, you sold 100 t-shirts.

Gross Profit = Number T-shirts Sold x (Average Selling Price – Average Cost)

Gross Profit = 100 x (\$29.99 – \$9.99) = \$2,000

If you don’t know Average Cost — i.e., cost of goods sold — replace Gross Profit with Total Sales in the MROI formula.

In our example, your marketing investment was \$1,600 — \$1,000 for Facebook advertising and \$6 discount per t-shirt. Thus your marketing return on investment is as follows.

MROI = (Gross Profit – Marketing Investment) / Marketing Investment

MROI = (\$2,000 – \$1,600) / \$1,600 = 25%

If you had run multiple promotions in a quarter, you could compare the MROIs to determine the top performers.

For example, the MROI for the 20-percent off t-shirt promotion (above) was 25 percent. Compare that with the following hypothetical promotions and their MROI.

• “\$5 Off All Ties” = 50% MROI
• “Free Shipping Sitewide” = 120% MROI
• “50% Off All Accessories” = 115% MROI

In reviewing the data, we can conclude that all promotions were successful, as each produced a positive MROI. “Free Shipping Sitewide” was the most profitable, however, at 120-percent MROI.

### Hundreds of Promotions

With a bit of data manipulation, you can expand the MROI analysis for hundreds of promotions. Here are the steps.

• Step 2. Classify each promotion, such as “\$ Off,” “% Off,” “Free Shipping,” and “Gift with Purchase.”
• Step 3. Classify the timeframe, such as “Christmas,” “Valentine’s Day,” “Mother’s Day,” and “Non-seasonal.”
• Step 4. If appropriate, classify the promotions by type of products, such as “dresses,” “t-shirts,” and “jeans.”

You can now run simple pivot tables for each classification to identify the best performing promotion type and product type. You could also classify by seasonality.

Here is an example.

SeasonPromotionMROI
Valentine's Day
% Off70%
\$ Off80%
Free Shipping110%
Christmas
\$ Off40%
Free Shipping90%
% Off90%

We can conclude that “Free Shipping” is the best Valentine’s Day promotion while “Gift with Purchase” is best for Christmas. Classifying in this manner also allows you to plan for new promotional tests in other periods. Additionally, it provides a benchmark for the type of promotion, season, or product.

### Long-term Analyses

To develop multi-year plans, keep track of all your promotions in a database, business-intelligence tool, or even a simple spreadsheet. Create new columns for types of promotions, to easily analyze the data in multiple ways, such as:

• Promotion type vs. product type. For socks, does 50% off work better than \$2.99 off?

For companies with a lot of data — many promotion types, many products, many timeframes — consider these analyses. Some involve adding new classifications.

• Promotion type vs. seasonality vs. product type. Which promotion and product work best for, say, Christmas?
• Product price. Should promotions for cheaper items be for a percentage off? Should promotions for high-ticket items be for a dollar amount off? For example, 50-percent off for items less than \$10 and \$250 off for items that sell for \$1,000 or more.
• Customer type. Consider classifying promotions by new customers, returning customers, high-spending customers, and occasional customers.
• Channel. Classify promotions based on the promotional channel. For example, you may have a 10-percent off promo code for email recipients and \$5 off promo code on Facebook ads.
• Geography. Regions and cultures react differently to promotions.

The key to analyzing the data in multiple ways is to classify your promotions accurately. It may seem daunting. But, in my experience, it will pay off as you run more promotions and develop better benchmarking.

Anna Kayfitz