What 488 shipping tweets reveal about posts that land

How the study read four container lines

Surucu-Balci, Balci and Yuen (2020), writing in Computers in Industry, examined how stakeholders engage with the social media output of the largest container lines. The study collected 488 tweets from Maersk, MSC, CMA CGM and Hapag-Lloyd between September 2018 and January 2019, then built a CHAID decision tree to find which features of a tweet best predicted engagement. The method splits the data repeatedly on whichever attribute separates high-engagement posts from low-engagement ones most cleanly, so the order in which features appear in the tree indicates their relative weight.

The sample is narrow. It covers 488 posts from four firms on one platform across a five-month window, so the figures describe an indicative pattern within container shipping, with the caution a single-platform sample of this size demands. Read that way, the tree is informative because it ranks the choices a firm controls when it writes a post, and it does so against observed engagement.

The first split, and the strongest single predictor, was length. Tweets at or under 225 characters scored mean engagement of 2.73, against 1.29 for longer ones. Brevity sat above every other attribute in the tree, which places the most basic editorial decision, how much to write, ahead of topic, media type and timing.

What the tree found about craft

The second decisive attribute was tangibility. Among short tweets, those showing a physical asset, a ship, a container or comparable equipment, scored 3.62, against 2.15 for short tweets carrying intangible content. A photograph of a vessel outperformed an abstract message of similar length, which suggests that the concrete subject matter of maritime work is itself an engagement asset when a firm chooses to show it.

Combining these attributes produced the strongest classes in the study. The highest-engaging group was short tweets that showed a tangible asset and carried corporate social responsibility or celebration content, at a mean of 6.80. Next came short tangible posts framed as company news or advertising with a call to action, at 3.98, then the same form without a call to action and without a firm mention, at 3.29. The pattern is cumulative. Each correct choice, length then subject then content class, lifted the result, and the best posts stacked all three.

The low end of the tree is as instructive as the high end. Long tweets carrying hashtags recorded the weakest engagement in the study at 1.09, and low-vividness, link-less posts of 192 to 225 characters scored 0.82. These are common posting habits, and the data shows them suppressing the attention a firm hopes to earn.

Where the conventional habits failed

Several widely used tactics worked against the firms that relied on them. Hashtags were the clearest case. They appeared in 71 percent of the posts, and among long tweets their absence beat their presence, 2.14 against 1.09. The default reflex to tag a post heavily was associated with lower engagement across much of the sample, which makes hashtag use a habit worth testing before it is assumed to help.

Other features helped, but only in the right place. Among non-tangible posts, an external link lifted engagement to 2.56 from 1.72, giving abstract content somewhere to lead. Video and animated content beat photographs and text, 2.53 against 1.37, yet only 8 percent of the sample used high-vividness video, so the most engaging format was also the rarest. Tagging another company lowered engagement, which complicates the common practice of co-marketing through mentions. Calls to action appeared in 21 percent of posts and advertising made up 37 percent, so the firms advertised often while using the highest-engaging content forms sparingly.

The overall picture is one of attention earned through specific editorial choices. The features that moved engagement, length, a visible physical asset, content framed as responsibility or celebration, a link where the subject was abstract, video over static media, were all decisions made before a post went out. None of them turned on how often the firms posted.

What the pattern implies for maritime marketing

The study points to a single conclusion for any maritime firm building a social presence. Engagement, and the credibility it signals to the people watching, was won by craft. The container lines that earned attention did so by getting the controllable details right, and the gap between a 6.80 post and a 0.82 post was a gap in editorial judgement that owed nothing to budget or reach.

That has a direct bearing on how a maritime firm should treat its content. The sector has a clear advantage hidden in the tangibility finding, because ships, terminals and cargo are exactly the concrete subjects the data rewards, and they are subjects most operators already have in front of them. Acting on that advantage is an exercise in discipline: writing short, showing the physical work, choosing the content class with care, and dropping habits such as heavy hashtagging that the evidence associates with weaker results.

The firms that treat content as a discipline accumulate something the data makes visible. Each well-judged post reads as a small piece of evidence about how a firm operates, and a stream of them builds the reputation and trust that stakeholders in this market form gradually and remember. Louder, sloppier posting does not build that record. A maritime social presence is won the same way the study's strongest posts were, through evidence-based choices made one post at a time, and the credibility that follows is the return on treating the work seriously.