White means not measured with javascript tag in FouAnalytics
Javascript tags are necessary for data collection [1]. Without javascript detection tags, a vendor cannot measure things like viewability, audibility, or invalid traffic correctly. So be sure to ask your current verification vendor to show you what portion of the ads were actually measured with a javascript tag, as opposed to a static .gif pixel. If they didn't measure your ads with a javascript tag, you should assume their answer is wrong, not that there's low fraud, high viewability, or that your video ads were audible.
This article is a short one, to show you how to look at FouAnalytics charts to understand what portion of the ads are measurable versus not measurable with javascript tags. There are occasions where the platforms simply don't allow you to copy and paste a JavaScript tag. For example, video ads (e.g. VAST) don't allow javascript tags, only static .gif pixels, so these are inherently LESS completely measurable than other types of ads like display ads.
Use separate in-ad tags for display ads versus video ads
In the screen shot below, the white is about 20% of the stacked percentage part of the time series chart (middle portion). Is this a problem? In this particular case, it turns out the advertiser used the same FouAnalytics in-ad tag for their display and video campaigns. The recommendation is to separate the display ad measurement from the video ad measurement by using 2 separate FouAnalytics tags, one for display, and one for video. This way the more measurable display ads have a dashboard of their own, for your analysis, while the video ads can be analyzed differently (because there is less data to work with).
CTV measurement is always limited to static .gif pixels
In the chart below, you see an example of CTV measurement in FouAnalytics. The CTV environment does not accept any javascript tags. That is why I have said over the years that CTV is less completely measurable than other forms of digital ads, and therefore easier for bad guys to get away with fraud. In this example, you will notice about 30% white, which is indicative for the part that is not measured with a javascript tags and therefore has no additional data than what is collected via HTTP headers and macros supplied by the DSP (if available). I won't delve into too much detail about how we do CTV measurement, but there are specific combination of signals we can see that allow us to confirm something as a real Apple TV, real Chromecast, real Roku streaming stick, etc. IP addresses are just an ingredient in the detection. Other vendors that rely on IP address alone for detection have severely flawed results and should not be trusted.
Typical display campaign measurement with low white
In the following chart, we can see a more typical display ad campaign, tagged with a FouAnalytics in-ad tag. Notice the amount of white is quite low (like 1 - 2%) in the middle of the stacked percentage chart. This means we have good measurability of the ads. We are getting javascript parameters back from the tag and those data can be used to corroborate the data seen in the HTTP headers -- e.g. did the detected userAgent match the declared HTTP_USER_AGENT. If so, that is good and corroborated. If those did not match, then we know something shady is going on.
领英推荐
So what?
If you see a lot of white in your FouAnalytics charts, it may not be a problem. But, if you know that the campaigns are NOT CTV or video ads (of the VAST kind), be sure to check that the tags are set correctly to "script" and not "pixel." It has to be set to script for the javascript tags can fire and collect data. We always want the most complete measurement possible, which means using javascript tags to collect data. Using static .gif pixels means you're doing not much more than just counting. If you care about the accuracy of your analytics, be sure to insist on javascript tags measurement. If certain vendors or platforms don't allow you to use a javascript tags, consider buying from somewhere else. If you don't have complete measurement yourself, can you trust the data the platforms or vendors self-report to you, especially if THEY say it's fraud-free, viewable, and audible?