4 mistakes e-commerce marketers make when measuring ROAS

The return on ad spend metric is an important tool for showing the efficacy of your message—but mistakes can render the number meaningless.


Most e-commerce businesses rely heavily on advertising to draw customers to their websites and, in the case of retargeting, remind them of past visits or past purchase intentions. When it comes to measuring ad spend, there are a handful of common pitfalls that every e-commerce business must know how to navigate.

Here are four different mistakes marketing managers make when assessing the return on their advertising investments (ROAS), and the only ways to avoid them.

1.      Double (or triple) counting a conversion.

Often, e-commerce businesses will use different marketing agencies to manage different channels. One agency will run the email campaign, another one will cover social media, etc. There is absolutely nothing wrong with this—that is, unless the measurement is conducted through the use of the engine data of the different platforms.

Google, Facebook and the other platforms each have their own data and they are all too happy to claim success for a conversion. Let’s say your clients have seen a Google ad, a Facebook ad, maybe a retargeting display ad on a website as well. They might have engaged with several of these ads, but only made one purchase. The problem now is that in many cases, adding up the data from the different platforms will show up as multiple of purchases what was actually purchased just once.

It’s only through the use of their CMS systems, or possibly even just their Google Analytics last touch attributions, that businesses will avoid over-counting purchases.

2.      Incorrect growth measurement.

An e-commerce business will encounter unexpected explosive sales during a period of time (or maybe the very opposite), and so the business will naturally want to know what happened. While asking yourself questions about irregularities is highly recommended, jumping to conclusions about the effectiveness of a paid advertising campaign because you see a sudden change is far from prudent.

Sometimes the most mundane reasons can cause spectacular changes in patterns. Take the weather, where a period of cold and rain followed by agreeable weather will have people run to be outside and soak in the sun. The clement weather often sees consumers neglect to purchase things online for a few days. (They might still buy sunscreen and camping gear!) Did this course of events mean that the advertising campaign faltered? No, of course not.

Month-over-month (MOM) measurement is not very instructive for the measurement of success of campaigns, and, for similar reasons, day-over-day (DOD) metrics make even less sense. Businesses will do well to focus on year-over-year (YOY) measurement comparing one January to the next.

3.      Wrong measurement design.

Another common mistake made in setting up measurement is ignoring the differences between channels.

Google Ads cookies used to expire in 30 days. Most e-commerce marketers were very well aware of this, and yet got into trouble when Google’s data was compared with the data from other channels. It is important to be able to compare apples to apples—or when you must compare oranges to apples, to at least know how many apples an orange is worth.

Another mistake pertains to plain errors in setting up the data sets. Google Analytics is a boon to marketers who are looking to measure return on advertising spend (ROAS) if (and only if) they truly understand how to use the tool. One mistake that can have a far-reaching impact is not accounting for the correct source of traffic when customers make a purchase.

4.      Listening too much to ad sales reps.

Sales reps from Google, Facebook and other platforms will deliver extraordinary value in explaining the functionalities of their platforms, but they are not impartial consultants when it comes to advice on the usefulness of platforms and how to measure success.

These same sales reps will, for example, often try to convince businesses that they are “leaving impressions on the table,” that their bidding is suboptimal and making them not rank as high as they could. The fact of the matter is that ranking high is not always what you want to achieve because the loss in margin will often not be compensated by the increase in sales volume.

The last word

Technology makes it possible for e-commerce businesses to measure ROAS in a variety of ways, which is good news for marketers. But at the same time, these tools can increase the likelihood of costly mistakes. Marketers should avoid overcounting purchases by relying on their CMS for attributing sales, stick to YOY measurement, take extra care of correct measurement design, and not listen too much to their sales reps when it comes to measuring success.

Darwin Liu is the founder and CEO of X Agency, an integrated digital marketing agency of growth engineers with offices in Boston, Massachusetts, and Nashville, Tennessee.


PR Daily News Feed

Sign up to receive the latest articles from PR Daily directly in your inbox.