Earlier this year, I tested the use of emojis in Instagram posts. In that study, I discovered that Instagram posts with emojis received slightly higher engagement.
So I wondered: Would using emojis have the same impact on Twitter content?
Existing research is scarce on emojis on Twitter and how it relates to engagement or impressions.
One Wordstream study says using emojis gets 25.4% higher engagement. However, this study used paid tweets and only tested 1 tweet with and 1 tweet without an emoji.
Not scientific enough for us here at Social Media Lab.
This year, Hubspot revealed which emojis get the most engagement and clickthroughs. But they offered no data comparing tweets with or without emojis.
Guess we’ll have to create a solid study of our own, huh?
Here’s my hypothesis — based primarily on what we saw in the Instagram emojis test:
Here’s how I structured this test.
For @Agorapulse, the schedule looked like this:
I posted 10 tweets per day. Every other day, I increased the frequency of emojis. Tweets on Days 1-2 had 1 emoji. Tweets on Days 3-4 had 2 emojis … up to Day 10 where tweets had 5 emojis each.
I kept things simple and put the emoji(s) at the beginning of each tweet such as this:
I used the same emojis on each tweet for that day – meaning that on “3 Emojis Day,” all tweets with 3 emojis had the same emojis as seen above. Below are the emojis used for 1, 2, 4 and 5 emojis.
The links used were a mix of our own content as well as content from friends including Ian Anderson Gray, Steve Dotto, Kim Garst, Owen Video, and Andrew & Pete.
I went to Twitter Analytics to compile the number of Impressions and Engagement.
The highest and lowest performing tweets were removed to remove any anomalies that might skew the data.
And with anomalies gone, I got this for Impressions:
Pretty easy to see 5 emojis outperformed all data sets. Compared to those with no emojis, tweets with 5 emojis received a 9.11% higher number of Impressions.
Tweets with fewer emojis, however, didn’t perform as well. The average Impressions of tweets with 1-5 emojis came out to 480.74 — 27.53% lower than posts with no emojis.
Now let’s take a look at the Engagement numbers.
That’s an average of 2.33 with emojis of any number.
Tweets without emojis had 107.73% more Engagement.
Honestly, this was not the result I expected.
But (and this is a big BUT) I discovered a possible problem with the data — and this is something anyone trying to evaluate results needs to pay attention to.
In the tweets without emojis, many of the tweets had @mentions and hashtags.
The tweets with emojis had no @mentions and no hashtags. Hmm..
So after removing these, I found that tweets with no emojis had on average 578.92 Impressions and 3.78 Engagements. Still higher than the tweets for the test with emojis, but not as much.
With the findings from the @Agorapulse account, I ran a short, second test with 20 tweets without emoji and 20 tweets with 5 emojis on my personal account. (Remember: 5 emojis was the winner in the first test.)
Would I get the same results with a different account and audience? Here’s what the data told me:
So the Engagement was about the same, but Impressions were 31.68% lower with emojis.
Based purely on the data, tweets with emojis underperformed when compared with an equal number of tweets without emojis.
Which makes my hypothesis wrong!
Tweets with emojis on the Agorapulse account had 16.96% fewer Impressions and 38.36% lower Engagement. My personal account saw a drop of nearly 32% in Impressions.
Neither were close to the results I expected. You would assume the friendly nature of emojis would help get more Impressions and Engagements, but the data says otherwise.
So while I think there is still value in including them in tweets at times to show your audience you have a sense of humor, it probably isn’t something you should include on every tweet.
Perhaps emojis and other “silliness” are best for in the replies on Twitter (maybe something to test later).
Being that this study was testing Impressions, there isn’t a way for me to put the numbers into the handy “Statistical Significance Calculator” since we aren’t looking at any sort of conversion. This could mean your results might vary from ours.
The key takeaway: don’t feel like you “have” to put emojis in posts to feel hip and millennial (says the 43 year old guy).
It might be worth running a further test to see if putting the emojis in the middle of the tweet as opposed to the beginning makes a difference or not. Stay tuned!