Human Factors and social media: a new lens on known use issue reporting

It’s important to acknowledge that people will often adapt how they use devices, and although this is not always due to misunderstanding, these adaptations can sometimes diverge from safe use. Social media can accelerate the spread of such approaches. But it also provides a seemingly rich source of known use examples that can potentially be drawn on by those looking to improve device use. So how valid a data set could these content sharing channels provide?

Traditional reporting channels and their limitations

Traditional means of searching for known use issues with a device include public databases like Manufacturer and User Facility Device Experience (MAUDE) Database, internal customer complaints records, and clinical literature. These tools are verifiable, document large numbers of issues, and are systematic to support thorough review. However, they have their limitations: they are cumbersome to filter through and can contain duplicate or irrelevant results. Significantly, they also rely on issues being reported, meaning the feedback chain is flawed and a lack of reports for a use issue may not reveal the full picture in terms of the extent of its prevalence.

Why users don’t report issues

So why would people not report errors through traditional means? There are many reasons. Many users may think they are using the device correctly and therefore have nothing to report. Users may also blame themselves for the difficulty, assuming that their condition, not the device, is the problem. In addition, users may think that a use difficulty is not worth reporting, only failures or malfunctions. Many difficulties often result in success of a task, just not in the way the device was intended. As a result, they are often not seen as reportable issues.

Risks of using social media data

However, the use of social media for known use issue collation does carry risks. Content is often unverifiable, and the presence of bots can make it difficult to distinguish genuine issues from noise. A 2017 study published via the Association for the Advancement of Artificial Intelligence (AAAI) estimated that between 9 – 15% of all Twitter accounts were automated, highlighting how engagement metrics can be distorted. High volumes of interaction cannot be assumed to reflect genuine user experience. That said, there are useful signals we can extract: recruiting key opinion leaders, analysing “life hacks” that reveal unexpected uses, and scouting for emerging trends.

A case for integration

Social media platforms like TikTok, Instagram, YouTube, and Reddit have become vast, searchable archives of real-world device use. If combined with traditional reporting, they could provide a richer, more complete understanding of use issues. Should we, as a field, embrace social media as a legitimate data source? Is the verifiability of the data on social media a barrier to its use? And how much would you trust social media to inform your design decisions, and where would you draw the line? These are questions we will continue to debate and reflect on as the known use footage available through these platforms grows ever larger.