Saturday, June 11, 2016

Facebook's New Policy Towards Text in Ad Images

Until very recently, Facebook had a policy concerning text in images of boosted posts. The policy was no more than 20% of an image could be text. Part of the problem with this policy was that it was not an actual 20%. Instead, an image was divided into a five by five grid and if text was in more than five grid squares it was judged as having more than 20% text. It did not matter if the font size was huge or small, or the amount of text was used so long as it occupied less than five grid squares. Placement played a great deal in whether text occupied more than five grid squares. One could sometimes just rearrange the text on the image so that it occupied five grid squares or less. Thus, the policy was sometimes not so much a judge of how much text there was, but a judge of where the text was placed. Unfortunately, the grid was laid out in such a way that such text placement did not always make for good graphic design. Read my previous post on the topic here.

Effective June 8th Facebook changed its policy. The 20% rule is no longer going to be used. Instead, they are classing images as Okay, Low, Medium, and High. An Okay image has no text. A Low image would be one with some text. A Medium image has more text than a Low image. And a High image would be one with a lot of text. A Low image will not get as many impressions as an Okay image, a Medium image will get a lot less while an image classed as High may not run at all.

I tested two images using the new testing tool and found both images to be in the Okay range despite having text. I suspect these images would be classed as Low in reality. Both images though may have ran afoul of the 20% rule due to text placement. This is much more acceptable to me as even though a Low image may get fewer impressions, it may be more effective than an Okay image at getting a reader to read the post's text. As I said in a previous post it has been my experience that oft times people look at the image, perhaps "like" it, and never read the post so you wind up not getting your message across. However, if an image has just enough text to pique a Facebook user's interest they may go ahead and read the post's text.

I can definitely work with Facebook's new policy. The most desirable policy though would be to only penalize images with a lot of text, and allow Facebook users via reactions and clicks to determine how many impressions a post may get. That is  images with no text and images with some text should both be classed as "okay," and have their number of impressions determined by how users react to them. Still though, the new policy is a step in the right direction,

Thursday, June 9, 2016

Facebook and the 20% Text Rule

I occasionally boost posts for a not for profit I volunteer for on Facebook. It is a great deal as you can reach several thousand people at a low cost.  Recently though, I had a sponsored post rejected. Why? Because text covered more than 20% of the image.

This, at first seems reasonable. Pictures catch the eye better than text, and people do not like large blocks of text. Those two things are advertising basics. Sometimes though text is called for, in the form of logos and basic information. Anyone who has ever tried to design a print ad making the most of text will soon realize text covering only 20% of an image is not very much.

After all there are many things that can be done with text that do not involve large blocks of it. The large blocks of text people don't like aren't logos or slogans that are big and bold, but the blocks of print like you see in a newspaper newsstory. And that is where Facebook's 20% text rule can cripple a sponsored post. Sometimes, a bit of text needs to be used to convey what the post is about. Just using the three Ws (when, where, and why) can give you text covering more than 20% of an image if you want the text large enough to be read.

Of course, Facebook probably thinks people will read the post if the image catches their eyes. Alas, people do not always read the text of a post without prompting by text in the image. Otherwise they just look at the picture, judge whether they like it, hit "like" if they do, and move on without ever reading the post.

The fact is many classic print ads would be rejected by Facebook if used in a sponsored post for text covering more than 20% of the image. Take for example, one of the greatest posters of all time, this one by the U.S. Army from the early 20th century:





When checked using Facebook's Grid Tool, this classic ad has text that makes up 40% of the image. Yet, would anyone deny this ad has not been effective in its aims? Would this image if turned into a Facebook sponsored post be more effective with just the picture of Uncle Sam, and then in the post's text "I want you for the U.S. Army?" Or would it be more effective as it is with the text big and bold in the image? I personally think the original would be more effective. Another example of a classic ad that would not cut it for Facebook is this one from Smith Corona:


This classic ad's text comes in at 24%. It does not look like it covers that much, but going by Facebook's Grid Tool it does. Then there is this 1960s ad from Kodak:





Ignoring all the fine print, and just focusing on, "wondering what you want for Christmas," the text covers 20% according to the Grid Tool. That is it is passable. However, if you add the text at the bottom, it goes up to 40%.. 

The fine print in these old ads can be considered like the text of a sponsored post, something you read after the image catches your eye. But with these old ads it is clear that text as well as the image can be used to draw attention.

So what does this say for Facebook's 20% text rule? It would seem to show Facebook's assumption that too much text looks like spam is not taking into account the context of the text. They are right, a lot of fine print, or text not important to the post like, "Only $5" may be spammy, However, text used the way it is in many classic ads is effective even if it covers more than 20% of the image. 

I understand Facebook wants to exercise a sort of quality control. And to be honest looking back, I am glad my boosted post was rejected. However, since then I have been looking at various Facebook page posts that could be boosted, but are not due to the 20% text rule. Most are well done, and the text covers no more than 30% of the images. No one I think would consider these posts spammy. And that is what makes the Facebook 20% text rule a sad thing. There are so many well done posts that could become sponsored posts if only the rule did not exist, or if the percentage of text allowed in the image were higher. I am not sure there is anything Facebook can do about it  though. If they get rid of the rule, they would probably be inundated with boosted posts that look like carnival posters. If they raise the percentage of text allowed the same thing might happen. Perhaps though, Facebook should look into some sort of solution, even if it means manually viewing the posts that get tagged for too much text, and not looking at some percentage, but on the quality of the text used. 



Edit: Since composing this earlier this evening I have read another blog post on the topic called "How to Get Around Facebook’s 20-Percent Text Rule on Ad Images" by Jon Loomer. What he learned was that it was not so much about how much text you had, but the placement of the text. The problem comes down to is the Facebook grid does not precisely measure the amount of text, but only that there is text in a grid square of the 5X5 grid. Thus it does not matter if your text is 12 pt. or 100 pt. both sizes count the same if they are in a grid square. This is even worse than if you were limited to 20% as suddenly your text could cover as little as 10% but if not placed right, but it would be rejected in that it was within 20% of the grid squares.

Update: Facebook has very recently changed its policy. You can read about the change here.