E-commerce Metrics That Matter
by Geoff Hineman
In recent years, the availability of—and preoccupation with—big data has given rise to more questions than answers. Indeed, as Google has taken away information, it has given new ways to slice and dice other information. The end result can often seem like data for data's sake, which sends marketers blindly down a rabbit hole of numbers. Different sites have different key performance indicators (KPIs) for success. At Lett Direct, we primarily work with retailers who actively manage e-commerce sites. To that end, we will focus on the metrics that matter for proprietors of e-commerce sites. Below are what we view as top priority metrics. As is to be expected when dealing with data, there will be overlap in some places, but I will try to delineate that overlap as much as possible. Let's begin.
Quite simply, revenue is your master metric. Beyond question, revenue supersedes all other metrics for e-commerce sites. Why? Because revenue is what goes in the bank. Every effort put forth by a marketing department or agency should be built around increasing revenue. Period. How do you know when you are increasing revenue? I recommend looking at annual month-over-month comparisons (i.e., August 2015 vs August 2014). The reason for this is because holidays have a very real impact on revenue numbers. Since many holidays fall on different dates each year (e.g., Thanksgiving, Easter, etc.), the data for those specific dates don't always line up well. In addition, the day of the week on which some holidays fall—Christmas being the big one here—necessitates orders being placed by certain shipping deadline dates. By backing the microscope out a little bit for a month-over-month look, you get a more accurate representation of revenue changes.
After revenue, sessions will be the next metric to queue up. If your revenue is not increasing to the degree you would like, it's time to see if you are getting more traffic to the site. A decrease in traffic (or growth that is not in line with projections) could be the root cause in under-performing revenue numbers. If the session numbers are lower than desired, you need to look at ways to get more people to the site. It is worth noting here that it is not uncommon for revenue numbers to increase as sessions decrease. This happens quite often at the start of new marketing initiative, as on-site optimizations and/or new PPC ads create a more targeted approach. The lost traffic is, essentially, non-targeted traffic that clearly wasn't converting anyway. As evidenced by the revenue increase, this change in visitor data and behavior places continued emphasis on revenue as the main metric, with traffic as a primary support metric.
Transactions / Average Order Value.
While it might be intuitive to look at conversion rate data next, I feel strongly that, once you have looked at revenue and sessions data, the conversion rate percentage actually means less than the raw numbers when it comes to transaction and average order value. This is because of the sometimes contrary movement of revenue and sessions I mentioned in the last section. I like to look at the raw numbers because they work in tandem to achieve higher revenue. Some clients see a greater number of transactions at a lower average order value. Some see fewer transactions at a greater average order value. While we can't control what people buy, the cost of goods goes a long way towards maximum profitability. Ideally, you are looking for the sweet spot of transactions and average order values that equal maximum revenue gains. This can be done by changing prices on the site or by placing greater focus on certain items that fall in that sweet spot. Remember, however, that the selling price of an item doesn't equal profit. Some items have higher margins than others and you need to sell fewer of them to realize better profits. These products should always be discussed at the start of a program to maximize revenue from the start of every marketing initiative.
Once upon a time, Google Analytics used to provide organic keyword data so marketers could see which keywords were driving the most traffic and sales. They don't anymore. Still, if you look at landing page data, you can discern which products (and by extension) which keywords are working for you. You can dig through this data all day and I wholly recommend that you at least look at the top 25 pages that lead to revenue on a regular basis. At a higher level, however, we know that customers who come to the site on a page that is not the homepage almost always convert better because they come in further down the conversion funnel. They don't have to search for what they want from the homepage. Fewer clicks and less time on site are better for conversions. That's why it is important to regularly look at landing page revenue and traffic changes that are happening on the home page, as well as non-home pages. Ideally, people coming in at a page that is not the home page are getting there because they looked for something other than a brand name. This is what you want. This is how you grow. Revenue generated by non-homepage visits is a good indicator of how well a program is working.
On Site Search.
Just about every e-commerce platform comes with a built in on-site search component; some are more flexible than others. If your system allows you to override default search results pages with custom results based on the search term, you should be taking advantage of it. How do you know if you need to, though? Google Analytics lets you look at the search terms people use on your site when they get there. You can also see the exit rate for a term. High exit rates mean that customers are not finding what they are looking for when they search on your site. Often this is caused by less-than-optimal default search results provided by the platform. In many cases, you have exactly what the customer is looking for, either at the product level or at the category level. If this is the case, and you have the ability to control the results, you should override the default results and serve customers the product or category page that best matches the search term. If you have put the effort into getting customers to the site, this can be a very useful step in getting the conversion and boosting your revenue.
Ad Spend ROI.
If you have enough budget, you can buy traffic all day long. Of course, as we have mentioned several times, if you aren't seeing revenue increases, you might as well be setting your money on fire. A well-managed pay-per-click (PPC) program operates at a maximum return on investment model, except on rare occasions. If you spend $10 on advertising and see $100 in revenue, that's good. If you reduce your spend to $8, but still see $90 in revenue, that is even better. While the overall revenue might be a little lower, your profitability is better. It also works the other way, too. If you increase your spend to $11 and see $125 in revenue, you should keep increasing your spend until you nail down the optimal spend-to-sales ratio. Revenue is the guide, for sure, but you want to attain it at the best return. I mentioned there are some occasions where you can forego this mindset for a short while. This would mainly be during your company's busy season—the holidays perhaps—when you want to increase your overall visibility, as you know your potential customers are searching on higher volumes. After a while, you can reign it in and, ideally, the visibility at that time can pay off at a later date.
Let me be clear, keyword rankings are very much a secondary (perhaps even tertiary) metric. In the age of Google's Hummingbird, the keywords you track have become an increasingly smaller sample size of the potential keywords that could be driving traffic to your site. If your revenue is down and your traffic is also down, however, it's worth taking a peek at your keyword rankings again. While fluctuation of up to three positions in either direction is very common, a large-scale downward trend in all of your keywords could be a sign that there is something very wrong. In such cases, you'll want to check any recent changes in the Google algorithm and review your site to see if this could be the cause. Common culprits for large-scale declines often include spammy content on your site, or even a lack of content. Your backlink profile might also be suspect. Also, check to see of recent programmatic changes have been made to the site. In some cases, a seemingly simple change can inhibit a search engine's ability to crawl the site properly and gather all the information it needs to reflect your site in a more accurate manner for searchers.
Being able to have access to an endless supply of metrics can be a blessing and a curse. For most e-commerce marketers, however, the metrics listed above will be the ones you need to follow for the vast majority of performance reporting and, when needed, troubleshooting.