Web Analytics KPIs for Retail

Often when we first talk with new clients we discover that they’ve done lots of different experiments in marketing, trying print ads, contests, cable TV, sponsorships, and more over a period of moths or years. Often they remember these experiments in terms of how much they spent. Sometimes they’ve been able to capture the ROI, but as often as not, they just have a general feeling that the results were pretty good or not so impressive.

With Google Analytics, you can dive in much deeper and get a much better idea of what works and what doesn’t. While some of the things GA looks at may be unfamiliar, with a little effort you can really find more common ground than you might think at first. Here are a couple of metrics that will sound familiar to CPG companies and retailers — but you might not have thought of them in terms of your web analytics.


  • Cost per acquisition, or CPA: (the amount you spent on marketing)/(# of people who make a purchase)

Imagine that you’re conducting an introductory campaign for a new product over a period of one week.

You conduct an email campaign that leads people to a landing page where they can make a purchase, and set up your analytics goals to show the purchases being completed. 

You can see how many people purchased the item. Total up the cost of your email campaign, including the price of the graphics or template and the cost for content plus the cost of sending them out. Divide the total by the number of purchases.

If, for example, you spent $200 on the email campaign and 12 people purchased, the CPA for that campaign was $16.67.

Say that for the same item you had a paid search campaign that brought you 643 visits at 69 cents per click, for a total of $443.67.

If 26 of the visitors purchase your item, your CPA for that campaign would be $17.74.

Which tactic was more valuable? For that, you need to think about a second important KPI: the Life Time Value of the customer.


  • LTV or lifetime value: (average number of purchases per year) x (average total value of a purchase) x (average number of years a customer continues to order)

Cost per acquisition makes most sense in tandem with the LTV of a customer. $17.74 is a high CPA for a single purchase of a $20.00 item, but not for a customer who makes that purchase every three weeks for 15 years. Chances are you can get the data for this calculation from the back end of your ecommerce site. General web analytics can help you connect this with other information to make better marketing decisions.

Since you can’t track specific individuals with web analytics, you’ll have to extrapolate from web visit data. You can see how frequently visitors come to your site and how likely they are to make a purchase. Compare this data with the information from your customer database or your retail analytics to gain an estimate of your average customer’s LTV. Segment your web data to compare different sources or groups.

For example, if you discover that the average purchase of customers who found you via paid search is $36.43 while those who come to you through email campaigns make purchases averaging $22.15, the LTV of visitors from PPC is higher — and the CPA can also be higher. If the email campaign brought back existing customers and the PPC campaign brought new customers, the values of the two campaigns could be quite different.

Harmonizing the data

Without connecting traditional KPIs like these with your web analytics, you would draw different conclusions from those you’d come up with if you looked at both sets of data together.

For example, that paid search campaign brought in more than twice as many paying customers with just a little more than a dollar higher CPA.  If email is more comfortable for you, you might decide that the ads weren’t worth doing. However, if paid search brings customers with a projected LTV of $9,000 and email customers have an average LTV of $5,500, the PPC campaign was definitely worthwhile.

Our examples use data available for ecommerce, and that does give additional data you won’t get if you don’t have an ecommerce component at your website. In fact, setting up a small ecommerce element at your website can often give you so much good data that it’s worth doing even if you don’t have much in sales. Even without ecommerce, though, you can dig into your website’s additional information. Add those additional insights into the numbers you use for your shop and see how they change your perspective.

Use the additional data your web analytics provide to good effect and you’ll make better marketing decisions.






Leave a Reply