Today’s post is by CitizenNet Founder Dan Benyamin.

newsfeed ad

Can you guess which of these ads draws the eye?

Imagine the following experiment:  Let’s say there is a friend you have lunch with every day at a favorite restaurant. You talk about a pretty broad range of topics, and you are used to your friend recommending things to you: places to go, people to connect with, movies to watch, amongst other things.

Lets say that friend recommends movies two different ways:

– During a conversation, your friend hands you a phone and says, “Hey, you have to watch this trailer.”

– Or, at that same dining table, your friend places a small newspaper ad of the same movie on the right side of the table – right by the breadsticks – and doesn’t mention anything.

Which advertisement for the movie are you most likely to notice? The video on the phone, right? Now, let’s say you acknowledge seeing the movie advertisement in either case – under which scenario are you more likely to see the movie?

This is the crux of native advertising: the placement and formatting of an advertisement makes all the difference. The same is true of Facebook’s News Feed vs. right-hand side (RHS) ads.

In this post, we’ll describe native advertising, examine the power of News Feed ads, and break down how the different comprehension and consumption of ads between the News Feed and right-hand side should help shape your marketing strategy.

Defining Native advertising

There has been much debate over whether so-called “Native Advertising” is really a new and good practice for the advertising industry. Such practice, where advertisements are placed and made to look like non-sponsored content in a published setting, is becoming increasingly popular with the adoption of mobile browsing and ‘news feed’ style page layouts.

A subtle but often overlooked aspect of native advertising is how people actually comprehend the information conveyed. One of the biggest challenges in measuring online advertising is in knowing if someone actually saw the advertisement. Web servers know impressions, but do they really know human comprehension?

Lots of advertisers and publishers don’t actually care if 99.9% of people don’t see the ad. Much of direct response is geared towards the .1% who click – as long as they are profitable, who cares?

Brand marketers, however, absolutely care! Much more of the advertisement is weighted on the impression, rather than the click.

The power of Facebook’s News Feed

Probably the most visible examples of native are advertisements in Facebook’s newsfeed.

Lots of people have shown great performance stats with Facebook News Feed. This is understandable, as a Facebook user is accustomed to gazing through content in the main column of the site, and are many times ‘blind’ to the right-hand side.

So while this is understandable, given a click on the right-hand side (RHS) and a click in the News Feed, what is the mental state of the user in the two cases? A user sees an image, some text, and generally a call-to-action and decides to click. If someone is consciously clicking, is there any evidence that ad placement and size matters?

Artificial Intelligence and human comprehension

While there is no published study on the neuromarketing effects of Facebook News Feed placement, we have set up a pretty novel experiment that models human comprehension of advertisements.

CitizenNet has, over the past four years, approached Facebook advertising as a recommendation problem: given a thing you would like to market, what types of people are most likely to be interested in it? When a recommendation system like Netflix recommends movies you have never heard of, but end up actually enjoying, it feels like magic – as if a friend who really knows your personality is recommending it.

Under the hood, recommendation systems are large, complex pieces of software, but their operation is intuitive. Just like that friend recommending a movie, recommendation software observes what you do, and learns from past mistakes.

Do you have a friend who has recommended a string of horror films, but you didn’t like any of them? Then maybe you aren’t so much of a horror fan. Much like that good friend, software recommendation systems are trained on your behavior in order to make better guesses.

The team at CitizenNet has set up a similar experiment with Facebook advertising as the earlier friend-for-lunch scenario: is the software better at predicting advertisements when only learning from News Feed placements?

Utilizing over 400 movie campaigns of CitizenNet’s clients on Facebook – totaling approximately 2 billion impressions – the CitizenNet data team trained the recommendation system on just the News Feed ad performance in one scenario, and just RHS ad performance in another scenario (each movie title presented a similar number of data points for either case, so it is a fair comparison).  The results are summarized in Figure 1, below:

Facebook ad predictability

The results are pretty striking: there is a 10% improvement when learning from just News Feed data. In addition, the system guessed at 100% of the actual movies you like, whereas with non-newsfeed information it attempted to guess at only 74% of the movies (and abstained from the others).

Not all impressions are created equal

What explains such discrepancy? Let’s return to the example of your friend the movie recommender. Let’s say that, all things being equal, your friend is more likely to recommend movies you end up watching when they show you a movie trailer on their phone rather than with the newspaper ad on the table – just like in the situation with the CitizenNet system.

That is, your behavior is more predictable when presented a movie trailer on a phone; even though you acknowledge both forms of advertisements, your pattern of movie-watching more consistently follows a pattern when presented with the trailer.

Our theory is that even though you acknowledge both forms of advertisements, you are not really processing the newspaper ad; it doesn’t trigger the same pattern of neural passageways as your other movie memories. In the online case, assuming people’s thoughts tend to be consistent, a software’s pattern detection is increased when people tend to think more about their actions.

Thus, with a significant increase in predictability with the News Feed case, we conclude that people more often think about what they click on in this native format.

While this is a profound statement, it makes some intuitive sense. Just as in the newspaper ad, RHS ads are small, static images. The marketer has to choose an image carefully, and that has to make sense to a human. There is, in fact, much less information conveyed in that static ad, so a thinking person has much less to go on.

Lastly, while creative formats and metrics such as brand recall and lift are many times derided as “soft metrics” by online direct response marketers, there is a reason traditional marketers love large, engaging ad formats: they engage and illicit thought, and this thought (even irrational thought) is what transforms watchers into buyers.

Dan Benyamin CitizenNet– For over 15 years, Dan Benyamin has imagined, architected, and built award-winning products used worldwide in a variety of industries. Before founding CitizenNet, Dan did private consulting for a number of consumer companies, most notably Best Buy, where he was a strategic investment advisor. Dan holds B.S. and M.S. degrees in Electrical Engineering from University of California, Los Angeles, and is an active member of several technical and entrepreneurial organizations in the Southern California region.

Leave a Comment

3Q Digital