Why your metrics are lying

  • 5 of 5
  • next
  • Sign in for single page

Video (YouTube)

The metric that lies: Video views. It seems pretty simple; if a person watches a video, one view is racked up.

The myths:

  • Just because a person started watching doesn't necessarily mean they watched to the end -- or even for more than a few seconds.
  • And video views are cheap to buy. Google the phrase "buy YouTube views," and bask in the preponderance of sites willing to sell views, subscribers, comments, etc. Some of the traffic is obtained via arbitrage, which is the practice of reselling a visitor to a page for more money than you paid. Other views are obtained by more nefarious tactics. "Botting," as it is sometimes called, is the practice of using software to watch and rate videos. Google (YouTube's parent) says, "We take abuse of our systems, such as attempts to artificially inflate video viewcounts, very seriously, and take action against known abusers, including termination of their YouTube accounts. YouTube continues to employ proprietary technology to prevent the artificial inflation of a video's viewcount by spam bots, malware, and other means." Despite this statement, most active YouTubers would agree that botting is still alive and well.

What it actually measures: A video view measures one robot or one person clicking the little triangular "play" button beneath a video. That's it. Video views can be useful numbers when measured across multiple videos or as an average valuation of a large sample size. In other words, this metric can help indicate whether a group of videos is generally more popular than another group of videos. But use caution when evaluating a single video unless you have other tools to support the data.

How to avoid being fooled:

  • If you own the video, then you're going to know whether or not you falsely inflated the views. So this really only applies to other videos trying to convince you that they are super-totally-popular on YouTube. And if that's the case, simply insist on more data to support the claim of popularity.
  • Remember that sometimes videos show high number of views even though virtually nobody made it to the end. So how can you know when they stopped watching? Well...

Tips for next time: Use YouTube's excellent analytics data. It's available for every video on your channel, and it contains loads of great information about demographics, on what devices videos were viewed, etc. It looks like this:

(Image source)

Source: Just Google anything like "buy YouTube views" or "how to buy YouTube subscribers." There are pages of options.

You might have noticed a trend in all of the examples in this article. Problems with analytics, metrics, and reporting often arise from the fact that standalone metrics (that is, metrics that aren't examined alongside other available data) are only one part of the whole. Unfortunately, shortsighted reporting can make a good campaign appear broken.

While direct-response thinking (focused only on getting users to perform a specific action or series of actions) can be useful when it comes to accurate measurement of metrics that matter to you and your company, understand that only pursuing these types of metrics might risk leaving the softer "brand experience" metrics behind. Don't lose sight of your brand's core message in exchange for getting clicks and other short-term wins.

Three quick tips for success:

  • Set a reporting baseline early
  • Measure everything you can
  • Test and refine

Drew Hubbard is a search and social media consultant and owner of LA Foodie.

On Twitter? Follow iMedia Connection at @iMediaTweet. 

"A conceptual image of a business man" image via Shutterstock. 

 

Comments

Jay Sterin
Jay Sterin June 26, 2012 at 12:01 PM

What about online audio? instream advertising.