Monday, August 4th, 2014 | 6 min read
A couple weeks ago, I wrote a post about the 10 Commandments of Paid Social outlining the general guidelines that practitioners should follow when incorporating paid social media ads into their social media marketing strategies. And just as there are the best practices for paid social, there are certain mistakes that should definitely be avoided.
You have an ad creative that always outperforms others. You have five years worth of data that tells you a certain line of copy always works best. Certain audiences always respond better for you than others. So you should just keep running them, right? No matter how much you think you know about ad performance, the mantra is always this: run tests. Surprising results will follow. We have seen this time and time again.
The rule of thumb is to use 5-10% of your media budget for testing.
On the flip side, you are so excited to test every single variant that you forget about actually driving results. “What if we ran 37 different images of horses to these 62 audience segments in 87 markets?” Stop!
Keep iterating on your best performing ad creative combinations, but don’t go crazy.
This is a classic! You have analyzed the data and found that the R coefficient is close to one (this is “nerd talk” for a perfect positive correlation between two sets of data). But did you really isolate every single factor and ensure that it was not random or external events that caused the result? Data may suggest that sleeping with one’s shoes on causes headaches. While there is a correlation between sleeping with shoes on and headaches, it is purely coincidental. Perhaps both conditions are the result of consuming too much alcohol?
Takeaway: always ask yourself what other events could have influenced the relationship between data sets A and B.
You are doing great. You have a customer care team that monitors all incoming messages from your Facebook and Twitter pages. They are masters at replying to these messages and you are meeting your SLAs. But now the paid team or your agency is publishing ‘dark’ posts or Tweets, and customer satisfaction is plummeting… What’s happening?
Well, your dark posts and Tweets still get shares, likes and comments, but nobody is monitoring them because they don’t fall under organic activity on Facebook and Twitter.
You probably have the best paid team out there, and they don’t mind working nights and weekends. But this doesn’t mean they should. Automate manual tasks that don’t require human skill — pausing campaigns that are not working is a perfect example of a task that doesn’t need a human touch.
Let your paid platform automate the tedious tasks for you, so that you can save your team valuable time and brain power.
I hate to say it, but please stop doing this. We are all so guilty of this behavior, and it is not funny. We didn’t know any better a couple years ago. Now we do. There is a lot more content vying for space in the Facebook News Feed now, and Facebook has changed its algorithm to try to deal with this.
Now it’s time to abandon the game of buying fans or followers and use your paid social dollars to drive business value.
What are you trying to achieve? Clients have approached us in the past with a burning desire to buy fans, create buzz or get more reach. But before gunning for those big numbers, it’s important to:
It is also a good idea to do a control/exposed test to really see whether your paid social activity actually has a measurable impact.
Not all engagements are created equal. You might have heard this before, but this is very important when optimizing media for business brand objectives. Often, I hear brands say that they care about engagements. Why? What are you trying to achieve as a business, and how does your paid strategy align with that? Different engagements must be given different priority weighting based on what is important to your business. A share on Facebook may be worth more to your business than a like, so why give them the same importance when optimizing?
Test out different weights, and figure out the cost per type of engagement against which you are measuring, so that you can start to optimize your creative and media for these goals.
As we have just learned, not all engagements are equal. If post A has 10 likes, 20 shares and 100 comments, and post B has 10 likes 10 shares and 30 comments, which post should you amplify with paid media? It’s a trick question – it can depend on the sentiment of the comments. If post A has 30 negative comments and 65 neutral comments, but post B has 30 positive comments, you might have promoted the wrong post by basing your decision purely on the number of engagements.
Where does your paid media data live? It’s highly likely that you get weekly excel spreadsheets from your agency, and they get stuffed in a folder somewhere. The point is that your data should not be stored on spreadsheets or on social platforms. Your data should live in an environment that you control, and where the data is easy to find and share with other teams.
If there’s only one thing you remember from this post, please let it be this: the power lies with the people who control the data and know how to use it.