How Effective is Interest Tagging on Facebook?
With the release of Facebook’s new Preferred Audience Optimization, the social media giant is declaring that these new interest tags will increase click-through rates (CTR) and that they will not limit the reach your organization has on the platform. In effect, the sky is the limit, says Facebook, with regard to reaching targeted individuals.
While the launching of this new feature generated a lot of initial buzz, there has never really been any statistical evidence which demonstrates that adding interest tags to organic posts had any impact whatsoever on engagement or on increasing the level of traffic directed to your site. However, that has changed now, with the availability of extensive data coming from publishers’ Pages, which does show a clear trend associated with the interest tags.
Based on solid statistical evidence, it has become apparent that the use of interest tags does not cause any corresponding increase in click-through rates, and that it does not generate any kind of increased traffic for the user. This means that audiences which are reached by strategically tagged posts are only just as likely to click on a link share, as those audiences which are reached by posts having no tags associated with them. Because this analysis is based on data coming from several of the biggest news publishers in the country, and which serve quite diverse audiences, the results may be relied upon as being accurate.
Audience Optimization is a feature which replaced Facebook’s original Interest Targeting feature in 2016, and when this replacement occurred, there was a huge amount of speculation about whether or not Audience Optimization would be more effective than Interest Targeting. Facebook generates significantly more than 10% of all organic traffic to online websites of publishers, and it has declared that Preferred Audience Optimization would allow these new interest tags to generate much higher click-through rates, as well as to sustain reader interest in tagged posts.
How the analysis was conducted
To carry out this analysis, it was necessary to indoctrinate a machine-learning algorithm in the use of natural-language-processing techniques which were capable of assigning interest tags to various posts on social media. By analyzing and comprehending the post content, as well as the article being shared, a good understanding could be gained of the content.
This step was necessary in order to quickly and consistently include tags across many hundreds and thousands of organic posts in the survey sample. In order to ensure that the machine learning algorithm actually did acquire high-quality interest tags, the appropriate Page owners were asked for feedback on a sample of posts which had been tagged.
In each case, the Page owners agreed that their own selections would have been tags similar to the ones picked out by the machine-learning algorithm and that these were high-quality posts appropriate for sharing on their Pages. Using this algorithm, thousands of organic posts with interest tags were posted to Facebook Pages for different publications.
After a day of sharing posts which had been tagged, posts without tags were shared for a day, before reverting back to the interest tagged posts. After an entire month of this process, the impressions garnered by posts with tags were compared to those having no tags associated with them, to see what the results would be.
In actual practice, none of the publishers’ Pages experienced any kind of significant increase in click-through rates when interest tags had been applied to their posts. This, of course, suggests strongly that tagging has no effect whatsoever on increasing the effectiveness of a post, and that adding interest tags, therefore, does not achieve its main objective.
A second analysis
In an unexpected turn of events, some Pages actually experienced a significant decline in click-through rate, while traffic remained constant. Because this kind of result seemed anomalous, a second analysis was undertaken, and this one used tags which were randomly generated, rather than selected by anyone associated with the survey. By using the same design as the original experiment, post was shared with random tags and with no interest tags at all, on alternating days of the week.
Again, as with the original experiment, at the end of one month’s time, the median impressions were compared between the posts tagged with random interest tags and those with no tags at all. The results were just as conclusive as to the original experiment, showing that randomly tagged posts achieved no better results than posts which were completely untagged. Click through rates remained the same for both types of posts, except that as before, a few posts which had been randomly tagged, actually experienced a notable decline in click-through rates.
With two very similar experiments being conducted on an exhaustive set of data, and which achieved very similar results, it can safely be said that applying interest tags to posts does not achieve any beneficial advantages, nor any consistently detrimental impact with regard to organic traffic. The only realistic conclusion from this is that Audience Optimization does not truly provide an effective way of reaching subgroups of a particular Page’s audience, with the kind of content which would be more important and relevant to them.
What this means to business people is that you would probably be better off focusing your efforts on producing high-quality messages to be shared, rather than attempting to reach out to a specifically targeted audience amongst your broad group of followers. Rather than applying interest tags, you’d be better off to create Pages on Facebook which have different topics and can be curated via the use of intelligent automation.
This will give users the capability of selecting the specific content stream they’re interested in, for example, a favorite sports team, or content strictly related to the national economy. You can use additional Pages to achieve high engagement rates, by interacting directly with users who have special interests, and who are much more likely to engage with you when customized content feeds are provided for them.