If you ask any marketer the big advantages that digital advertising has over traditional, analog channels, “more measurable” has been toward the top of the list for years. Even now, digital earns nicknames like “the nirvana of quantifiable marketing” from industry leaders.
But your digital advertising metrics aren’t all they’re cracked up to be.
Don’t get me wrong: Digital is leaps and bounds more trackable and quantifiable than a print ad, offering both more and more concrete data. But for all of its advantages, digital reporting isn’t perfectly accurate.
Once you recognize the shortcomings that ad measurement sometimes comes with, you can better analyze your data and take countermeasures to compensate for the gaps and skews. Here are a few reasons why your metrics may not be giving you a true view of your digital advertising.
We all have different definitions
The first problem is that the same metric can mean different things to different companies. There’s inconsistency between what a metric is actually measuring depending on the situation. For example, there’s still no standard definition of “viewability” across different types of businesses in the industry.
There may be some agreement among different groups—like publishers, vendors and advertisers—but no one standard for the whole industry. eMarketer found that while more than 70 percent of publishers use the viewability standard set by the Media Rating Council, less than one-half of brand and agency professionals do so, instead either setting their own guidelines or using none at all.
While we all agree that universal standards would be a good thing, competing goals and priorities make it difficult for different sides to agree on what these standards should be. For example, Integral Ad Science’s Look Ahead Report found that some found the MRC’s standard fine, while others found it not stringent enough.
Bots don’t buy
Another problem with your ad metrics is the rising impact of ad fraud, which Juniper Research estimates will cost advertisers up to $19 billion in wasted spend this year. Unlike something like viewability, which is mostly a matter of decision-making, click fraud is a broader problem in the industry, and it’s a hot-button issue.
Left unchecked, bot traffic and results can greatly inflate your metrics and give you an incorrect view of your campaign. Some activity will be from real viewers, some will be fake, and you cannot be sure which is which.
Plus, it impacts different metrics in different ways. For example, a campaign that’s been subject to click fraud may have an overinflated click rate. But if your end goal is purchases and one-third of that traffic is fraudulent, your sales page conversion rate will appear to be lower than it really should, since bots don’t enter their credit card info and buy.
To protect your brand and ad campaigns against fraud, solutions like ClickCease can integrate directly with ad platforms like AdWords and Bing to identify and block invalid clicks. That way, fraud’s taken care of before it hits the advertiser’s website and ad budget.
Ad lift is misleading
Say you do keep bots out of your campaign and it seems to be going well. Unfortunately, you may still be overestimating the impact any given ad has on your overall marketing, sales and revenue.
This is because once we see that people who were shown our ads have converted, we assume that the ads are the reason. But that’s a classic example of mistaking correlation with causation.
To truly measure the lift an ad provides, we need to be able prove not only that an ad viewer completes a conversion, but that they did so because of the ad—and would not have done so otherwise. You’re likely targeting audiences with a high probability of purchasing anyway, for example because they’ve expressed a certain interest or they’ve already interacted with your website, emails or other marketing messages. Some of these people probably would have eventually converted without seeing your ad.
The most accurate way to measure ad lift is to conduct randomized control trials, dividing your audience into two groups that don’t differ in any important ways but one: One group sees the ads and the other does not. Unfortunately, RCTs can be time-consuming and expensive, and many marketers are happy enough to assume their ads are the main driver of their results and don’t even see the need for this. A wiser perspective would be to see ads as brand amplifiers and sales accelerators.
You’re stuck in last-click attribution
Next, you’ll want to check your attribution models and bring them up to the times.
If you’re still mainly using last-click attribution to measure your advertising, you can also count on your metrics only telling a partial version of the story. This is because full “credit” for the sale is being placed on the last action or “click” an audience member took before purchasing. In other models, such as first-touch and linear, you’re able to take into account the role other messages played in the path to conversion.
There are so many flaws in fully relying on last-click that it’s become outdated to the point that even Google is trying to get us away from it.
Similar to the issue with ad lift, last-click attribution gives credit for conversions and results where it may not belong. Since you credit the conversion to the customer’s last touchpoint only, it doesn’t take into account the marketing efforts that came earlier in the sales funnel.
Given that you’re reaching your customers through multiple channels from advertising to email, an attribution model that tracks multiple touchpoints can help you see how big of a role each one plays in assisting conversions. Tools like Bizible can show you a more complete representation of your buyer’s journey.
Not all results are equal
Finally, let’s say none of the above problems is skewing your advertising data. Even if it’s 100 percent quantitatively accurate, it doesn’t tell the whole story.
The fact is that using the same metrics across different mediums and formats doesn’t account for the qualitative differences in each. Different platforms are easier or harder to measure accurately, the user experience and expectations are different across touchpoints, the quality of leads from one campaign might differ from another and so on.
For example, if you have a LinkedIn campaign getting you a lower cost per lead than your Facebook ads, it’s sensible to assume the LinkedIn campaign is a better performer. But what if you’ve noticed that your leads from Facebook are usually much higher quality and more likely to convert, or that the conversions have higher lifetime customer value?
Or what if you’ve got a YouTube campaign that’s yielding huge results from every measurable perspective? You’re still playing Russian roulette, in a sense, with your brand safety, as you never know when your spot will run as a pre-roll for a white supremacist’s vlog. There’s always another layer to consider.
As flawed as they are, your various reporting tools are still going to provide the best ways to understand your digital advertising performance, and it’s not like you can improve and succeed without any data.
But moving forward, be sure not to take metrics at face value. Instead, aim to understand what exactly they tell you, where they fall short and what you can do about it.