Social media and out-of-hours conduct

A post made at home, after hours, to a closed group of “friends” still ends up in front of the Commission more often than people expect — and “it was private” protects an employee far less reliably than most people assume.

What you’ll be able to do

• Apply the out-of-hours connection-to-work test specifically to online conduct, using the multi-factor test the Commission applies to decide whether a post is “sufficiently connected” to the employment.
• Explain why a post shared only with “friends” or in a closed group isn’t reliably shielded from employer action — and why, counter-intuitively, the same privacy setting can also work against a finding of a valid reason.
• Spot when a social media policy actually covers the conduct relied on, and when the employer has reached for the wrong clause.

The connection-to-work test, extended to social media

Module 5 covers the general out-of-hours test from Rose v Telstra: conduct must, viewed objectively, be likely to cause serious damage to the employment relationship, damage the employer’s interests, or be incompatible with the employee’s duty. For online conduct specifically, a Full Bench of the Commission — described in one decision as Sydney Trains v Bobrenitsky (not independently located in this corpus under that citation; check the quote against the original before relying on it) — refined that test further: “it is axiomatic that for conduct to indicate a rejection or repudiation of the employment contract, the out of hours conduct must be sufficiently connected to the employee’s employment. Not every connection between out of hours conduct and employment, will constitute a valid reason for dismissal” [2023] FWC 907 at [116].

Working out whether that connection exists means looking at the whole picture, not just whether a post exists: “the nature of the out of hours conduct and what it involved; where the out of hours conduct occurred; the circumstances in which the out of hours conduct occurred; the nature of the employment; the role and duties of the employee concerned; the principal purpose of the employee’s employment; the nature of the employer’s business; express and implied terms of the contract of employment; the effect of the conduct on the employer’s business; and the effect of the conduct on other employees of the employer” [2023] FWC 907 at [116].

A post’s permanence also matters. A Facebook post “does not stop once work recommences. It remains on Facebook until removed, for anyone with permission to access the site to see” — and “it would therefore be foolish of employees to think they may say as they wish on their Facebook page with total immunity from any consequences” [2021] FWC 3030 at [32].

Public, “friends only”, or screenshotted: what actually matters

A genuinely instructive pair of decisions shows that “private” versus “public” is not itself the test — and that a private setting can cut either way.

In one matter, a swim instructor replied to a stranger’s request for a recommendation inside a 16,100-member private community Facebook group, naming a competing swim school instead of her own employer. A colleague who happened to see the reply through her own Facebook notifications took a screenshot of the entire thread and forwarded it to management [2021] FWC 1952 at [13] — even though both women had their profiles set to private [2021] FWC 1952 at [12]. A private audience is not a sealed one: anyone inside it, including a colleague, can copy and carry content straight to an employer, with or without malice.

But the Commission’s own reasoning then cut the other way on liability. Because the Applicant’s settings were private, “it would not have been apparent to persons reading her post, other than those directly familiar with her employment history, that she worked for the Respondent” [2021] FWC 1952 at [45]. That absence of any visible link to the employer was one of the reasons the Commission found the dismissal was not consistent with the Small Business Fair Dismissal Code [2021] FWC 1952 at [46] and was ultimately not satisfied a valid reason for dismissal was established at all [2021] FWC 1952 at [52].

Contrast a fully public post where the Applicant had identified himself on his own profile as an employee of the Respondent [2021] FWC 3030 at [38]. The Commission accepted it would be “fanciful” to think the public would attribute his personal opinions to the business — but found that naming the employer on his profile nevertheless “associated the organisation with those Facebook postings, such that a person may perceive the opinions to be representative of the Respondent” [2021] FWC 3030 at [44]. A valid reason was found, and the dismissal was not unfair [2021] FWC 3030 at [69].

Read together: what matters is (1) whether the content in fact reached people who matter — which privacy settings do not reliably prevent — and (2) whether, on an objective look, a reader could actually connect the post to the employer and to the employee’s role in it. A truly untraceable private post is a weaker case for dismissal than a fully public one naming the employer, even where both are seen by similar numbers of people.

A third matter shows the same tension again. The Applicant there argued that the conduct in his private Facebook/messenger group “continue[d] not to be in the public domain” and that, but for his dismissal and the ensuing proceedings, there was no reason to think it ever would be [2023] FWC 907 at [83]. He also argued that, to amount to harassment, conduct must be “directed towards a person” and that “the context was a private group of consenting men” [2023] FWC 907 at [88]. Neither argument prevented a valid reason being found — because the post the Commission actually relied on (an on-shift comment disrespectful of management) breached a policy whose application didn’t turn on publicness at all.

Posts about colleagues, managers, clients or the business

Three recurring patterns, regardless of audience size:

• Discriminatory content. Public Facebook posts described by the Commission as “vituperative criticism of Aboriginal people, women, and persons in the media” [2021] FWC 3030 at [43]. Intent to offend is irrelevant to the finding — “whether the Applicant sought to offend by his diatribe(s) is of no consequence” [2021] FWC 3030 at [43].
• Sexual or demeaning content about a named colleague. A sexually explicit meme referencing a specific female colleague, posted to a personal but public Facebook page [2021] FWC 2570 at [256]. The employee argued a “robust” sales-floor culture made it acceptable; the Commission held the conduct was “still sufficiently serious and inappropriate so as to constitute a valid reason” regardless [2021] FWC 2570 at [257].
• Disrespectful comments about management, even inside a closed group. An on-shift post mocking a manager’s new vehicle was found to breach a Bullying and Harassment Policy defining harassment as “any unacceptable action, conduct or behaviour that a reasonable person would find unwelcome, humiliating, intimidating or offensive” [2023] FWC 907 at [120].

None of these findings turned on the post being public. What mattered was the content, and who or what it targeted.

Social media policies: what they can — and can’t — validly restrict

A social media policy survives scrutiny when its reach is tied to the employment connection, not to an employee’s life generally. In one matter the Commission was satisfied a policy “provided lawful and reasonable instruction” because “the Policy’s scope did not overreach. It sought only to regulate the social media use of its employees to the extent where employees made reference to the Respondent, or by inference, the staff member could be seen to be associated with the Respondent” [2021] FWC 3030 at [36]. A policy that tried to control an employee’s entirely unconnected private opinions would not get the same answer.

Equally, the specific policy an employer relies on has to actually cover the conduct. In the private-group matter discussed above, the employer pointed to several policies — a Bullying and Harassment Policy, a Code of Conduct, a Social Media Statement and the contract of employment — but the Commission’s eventual valid-reason finding rested on two particular breaches it identified itself: the Bullying and Harassment Policy’s harassment definition, for the on-shift post about a manager, and a separate requirement for Defence approval before publishing photos of the worksite, for an unrelated photo [2023] FWC 907 at [121]. A broad “Social Media Policy” argument that doesn’t map onto the actual post relied on won’t do the work an employer needs it to do.

What the Commission looks for

• The whole factual matrix, not a single fact like “it was public” or “it was private” — nature of the conduct, where and how it occurred, the employee’s role, the nature of the business, the contract terms, and the effect on the business and on colleagues.
• A real, demonstrable link to the employment — the post names the employer, identifies the person as an employee, targets a named colleague, manager or client, or is otherwise traceable back to the workplace by someone who sees it.
• A policy clause that actually covers the conduct, not just a policy that exists somewhere in the handbook.
• Content and target, not audience size — a defamatory, discriminatory or harassing post about a colleague or the business is assessed on what it says and who it’s about, whether ten people saw it or ten thousand.

How social media dismissals fail

Adviser checklist

Practice scenario: the community group post

Decisions cited

Check each paragraph reference against the published decision on the FWC website before relying on it. The Full Bench authority Sydney Trains v Bobrenitsky is quoted in a later decision without a neutral citation being given and is flagged below — it still needs checking against the original.