#Retweet: How to Compose Shareable Tweets

September 1, 2020

Nima Y. Jalali, PhD and Purushottam Papatla, PhD

stock photo of person's finger scrolling twitterIf you want to increase awareness of your brand, then Twitter was made for you. Your followers are some of your best brand ambassadors, as they voluntarily publicize your content to a broader audience. However, on social media, you don’t have much time to secure followers’ attention. Most users will only glance at a tweet for a few seconds before moving on to the next one, and users are only on Twitter for a few minutes a day.1 If users do not recognize the topic of your tweet within a few seconds, they will likely skip it and move to the next one that piques their interest. Brands therefore need to compose tweets so that attention-grabbing, content-rich material is recognized and the topic is understood with a simple glance. Our study draws on findings in psycholinguistics, and we analyzed whether the occurrence of topic-related words at the beginning of the tweet affects the number of retweets. Specifically, we studied sales-promotional tweets from 60 brands and found that concise tweets composed with topic-related words visible at a glance received more retweets. When composing a tweet, less is more, and every word is precious.

How We Read

The few recommendations about tweet composition that currently exist usually suggest putting attention-grabbing words like “WOW,” “LOOK,” or “TODAY ONLY” at the beginning of tweets to create a sense of urgency, but this approach does nothing to familiarize users with the advertised product. Recently, a study showed that most people share tweets when they sense a connection between the content of the tweet and the topics about which they like to tweet.2 Those attention-grabbing words might be enticing users to glance at a tweet, but the content is not understood quickly, and thus, users don’t see the connection for them and do not share them.

Our research is based on findings in psycholinguistics that suggest when we are not using much effort to read, like the everyday twitter user, we comprehend the meaning of a sentence by scanning a subset of the words in the sentence.3 We experience certain paired eye movements when reading this way that we do not consciously recognize. The first movement is known as a saccade, which is a superficial scanning of the sentence that skips some words.4 There is almost no comprehension during this movement. The second is a fixation on the last scanned character.5 Skimming is the experience of a saccade-fixation pair followed by subsequent saccade-fixation pairs until the end of the text. We have not read all the content, but we infer its meaning from the words on which we fixated. Research suggests that we spend about 20-35 milliseconds for each saccade and 150-500 milliseconds per fixation.3 In other words, the goal is to maximize understanding with minimal effort and time.

How We Share

Social media users typically share a tweet when they discern the meaning of it early in the reading process. If it takes too long, they will move on to the next tweet hoping to comprehend the meaning from a few words in a few seconds. Also, people will comprehend the meaning of the tweet based on the words on which their eyes fixate. So, if a tweet is saturated with topic-related words, the likelihood that their eyes will fixate on a word that conveys the right meaning is increased. Increase the chances of fixation on a topic-related word that interests the reader, and it will increase the chances for a retweet. Topic-related words are words that share semantic relationships that people group together. For example, the word “bank” implies finance-related ideas and creates the expectation of other related words like “federal” and “reserve.” If, however, the next word that is fixated on is “stream,” then readers would comprehend the sentence to be about rivers or other bodies of water.6

We find for sales-promotional tweets (SPT) that placing topic-related words at the beginning of the tweet results in more retweets. We also analyzed the effect of words from other topic categories and found a large effect of calls-to-action (CTA) words (although we also find that they should not to be used early in a tweet). The other large effects were of time and brand respectively. If possible, a tweet can contain all three of these topic-related words through creative content, hashtags, and links to other sites. However, less is more. Long tweets get significantly less retweets. If you must sacrifice something, do not sacrifice the CTA topic-related words.

How to Compose Tweets

There are several factors that affect the “retweetability” of an SPT, such as posting time, the number of tweets per day, and punctuation. For instance, tweets posted on Sunday typically receive the most retweets. The number of retweets then decreases daily until Thursday. If a brand posts several tweets a day, retweets will decrease. Also, questions and multiple exclamations decrease the number of retweets. However, brand presence on other social media websites, the number of other brands that are followed, and presence on the Interbrand index have no discernable effect on retweetability.

It remains that the most significant effect on a tweet’s propensity to be retweeted is the early presence of topic-related words. We recommend that brands find the topics that interest their followers by regularly observing their most retweeted posts. Stay away from filler words that are only intended to grab attention or to increase the tweet’s length. Instead, draft concise, content-rich tweets that contain words related to the topics that your followers are interested in sharing. Further, sprinkle multiple topic-related words throughout the tweet so your followers can spot the topic and understand the meaning of the tweet within a few seconds of scanning it.

woman sitting at office table looking at cell phoneOnce you have compiled a list of words to which your followers respond, begin looking at your competitors’ tweets and add successful topic-related words to your word bank. After you have spent time storing up these words, begin experimenting with new words that advertise the same topic and find ones that work the best. Make them short. Make them meaningful. Make them “skimmer friendly.”


Every realtor is in the networking business. You need a client base that you count on for word-of-mouth advertising from your best customers to gain new business. Retweets are digital word-of-mouth advertisements. The customers you have helped are the best ambassadors of your personal brand. The relationships you formed with these ambassadors make them want to follow the work you do. If you are not using Twitter to grow your client base, consider tweeting at least once a day to tell your followers about your work, your listings, or your team. Find topic-related words that result in the most retweets, and use them early when composing concise tweets. Consider advertising your next open house with a call-to-action on Twitter. Slowly build your word bank and watch your brand penetrate new markets with almost no advertising cost to you. Oh, and remember to finish them with the most important call-to-action there is—#retweet. 

. . . . . . . . . . . . . . . . . . .

Recommended Reading

Jalali, Nima Y. and Purushottam Papatla (2019), “Composing Tweets to Increase Retweets,” International Journal of Research in Marketing, 36(4), 647-668.

. . . . . . . . . . . . . . . . . . .


  1. Bennett, Shea (2014), “This is How Much Time We Spend on Social Networks Every Day,” Advertising Week, https://www.adweek.com/digital/social-media-minutes-day/.
  2. Zhang, Yuchi, Wendy W. Moe, and David A. Schweidel (2016), “Modeling the Role of Message Content and Influencers in Social Media Rebroadcasting,” International Journal of Research in Marketing, 100-119.
  3. Rayner, Keith, Alexander Pollatsek, Jane Ashby, and Charles Clifton, Jr. (2012), Psychology of Reading, Psychology Press.
  4. Matin, Ethel (1974), “Saccadic Suppression: A Review and an Analysis,” Psychological Bulletin, 81, 899-917.
  5. Frazier, Lyn and Keith Rayner (1982), “Making and Correcting Errors During Sentence Comprehension: Eye Movements in the Analysis of Structurally Ambiguous Sentences,” Cognitive Psychology, 14, 178-210.
  6. Griffiths, Thomas L., Mark Steyvers, and Joshua B. Tenanbaum (2007), “Topics in Semantic Representation,” Psychological Review, 114(2), 211.   

. . . . . . . . . . . . . . . . . . .

About the Authors

Nima Y. Jalali, PhD
Assistant Professor of Marketing, University of North Carolina Charlotte

Dr. Nima Jalali’s (PhD – University of Wisconsin-Milwaukee) primary research interests include digital marketing, social media, unstructured data analytics such as textual analysis and image analytics. In his research, he develops and utilizes empirical models to gain marketing insights from customer’s activities, which potentially informs firms to enhance their digital and social media strategies and tactics. His research has appeared in International Journal of Research in MarketingQuantitative Marketing and EconomicsEconometrics Reviews, and Journal of Multivariate Analysis. His recent work looked at the role of topic word composition in tweet on retweets, the effect of color composition on click-rates in visual commerce, and the role of peripheral cues on product page visits and conversion from visual user-generated content.

Purushottam Papatla, PhD
Northwestern Mutual Data Science Institute Professor of Marketing, Co-Director, Northwestern Mutual Data Science Institute, University of Wisconsin-Milwaukee

Dr. Purushottam Papatla’s (PhD – Northwestern University) research interests include Bayesian modeling of Big Data to understand how and why Instagram posts by consumers affect how other consumers respond to brands. He is also using Bayesian and Big Data methods with psycholinguistics to investigate why and how consumers engage with brands on Facebook, why consumers retweet brands’ tweets, how Uber and AirBNB have affected consumer response to legacy taxi and hotel brands, and whether consumer ratings affect sales in the restaurant, hospitality, and music categories. His other current research projects also include studying what makes people watch ads on TV more fully rather than switching out of them using ad and program viewing data.  Dr. Papatla works with companies through non-commercial research relationships and publishes his findings through scholarly journals. He teaches Marketing Strategy and Digital Marketing Strategy in the Executive MBA program and Doctoral Seminars on Bayesian Methods and Models in Marketing at his university. He also recently introduced a course called Ideas and Applications of Data Science in Different Fields that was very popular at his university.  Dr. Papatla also has been, and continues to be, the Dissertation Committee Chair for several Doctoral students and serves as a reviewer and Editorial Board member for peer-reviewed journals.