Three Social Media Posts #1 — What Worked? What Didn’t?

Jessie Y. Shi
Social Media Close-up Look
2 min readApr 29, 2019
Photo by William Iven on Unsplash

On Facebook

The post was great because:

  • It featured something unusual and potentially interesting, the snake massage, and therefore it hooked viewers to watch the video.
  • The video was shot in a square mode, which was platform- and mobile-friendly.
  • It had high-quality visuals with a lot of close-ups, making readers feel authentic.
  • The post itself was a question that sparked robust discussion among viewers.
  • It had explainer text that complemented the video as well as subtitles for silent watch, which is common on mobile.

On Instagram

The post could have been great. It needed some improvement:

  • The video was low-quality. It was blurrier than the one from New York Insider. The room where it was shot was not well-lit. Video producers should have lights with them even when shooting with mobile devices to optimize the quality.
  • No subtitles. Mobile views account for over 80 percent of all views on Instagram, and a lot of users consume videos in a silent mode. Subtitles are a must on social media.

On Twitter

The Tweet from Documented stood out because:

  • It was conversational (we/us/you/your), which worked better than “he/she said” on social media.
  • It was community/audience-centric. Instead of promoting content from Documented, the post asked viewers for their insights.
  • It had a clear call-to-action by telling viewers how they could contribute.

It is rare to see posts like this one. Most of the time, publishers only promote their content to viewers. There are still very few two-way conversations on social media. To make social media social, we should initiate conversations back and forth between publishers and audiences on all platforms.

--

--

Jessie Y. Shi
Social Media Close-up Look

Audience engagement editor. I engage the right audiences with the right stories on the right channels at the right times, informed by audience data.