Where Enterprise Agreement Ballots Go Wrong — and How to Avoid the Traps
A proposed enterprise agreement can survive months of negotiation and still come unstuck at the ballot box. The vote itself is often treated as a formality — a box to tick before lodgement — yet it is the point at which every earlier assumption about coverage, communication and consultation is tested. When something goes wrong, the consequences are rarely minor: a challenged result, a delayed lodgement, a workforce that no longer trusts the process, or a full re-run under a cloud of suspicion.
The pitfalls below are the ones that surface most often across Australian employers, unions, member associations and clubs. None of them require bad faith. Most are the product of tight timelines, unclear ownership and processes that were borrowed from a previous vote and never revisited. Understanding them in advance is the cheapest form of risk management available.
Pitfall one: an inaccurate or misunderstood voter roll
Almost every disputed ballot traces back to the same question: who was entitled to vote? An enterprise agreement is made with the employees who will be covered by it, and that population is not always the same as the payroll list, the union membership list or the distribution list used for workplace updates. Where an employer has multiple entities, a group of related employers, newly acquired sites, labour hire arrangements or a mix of classifications, the boundary of coverage can be genuinely ambiguous — and ambiguity at the roll stage becomes a dispute at the count.
Common failures include rolls that were extracted weeks before the vote and never refreshed, casual and part-time employees who were missed because they do not appear on a standard full-time roster, shift workers and fly-in fly-out staff whose contact details are out of date, employees on parental or long-service leave, and people who joined or left the business during the access period. Duplicate records are equally corrosive: two entries for the same person, one work email and one personal email, will inflate the roll and distort turnout figures. One person, one vote is only meaningful if the roll behind it is clean.
The practical remedy is to treat the roll as a controlled document. Reconcile it against payroll and human resources records, document the inclusion and exclusion decisions, have a second person review the boundary cases, and freeze the roll before the ballot opens with a clear, published cut-off. Where the coverage question is contested, seek advice early rather than letting the ballot become the test case. It is also worth remembering that union membership and employee status are different things: a ballot of employees is not a ballot of members, and conflating the two is one of the most common — and most avoidable — errors in union-supported votes.
Pitfall two: process, notice and conduct failures
Enterprise agreement ballots are procedural by nature, and procedures fail quietly. Notice periods that are miscalculated, an access period that is truncated because the vote was scheduled around a public holiday, or an agreement document circulated in a form employees cannot practically read all create grounds for complaint long before the count begins. The agreement that is voted on must be the agreement that is lodged with the Fair Work Commission — any change made after the ballot opens, even a minor one, undermines the integrity of the result.
Independence is the second fault line. Employees are entitled to be satisfied that the vote was conducted fairly, and perception matters as much as reality. A ballot run by someone with a direct stake in the outcome, a paper process where completed forms pass through a supervisor’s hands, or a show of hands in a team meeting all invite accusations of pressure, even where none was intended. Secret voting, clear custody of the ballot and appointed scrutineers are the baseline, not an embellishment.
Conduct issues extend to the logistics of the voting window. A single on-site voting session at 10am effectively disenfranchises night shift, part-time and remote employees. Mixing the agreement vote with an unrelated survey or an annual general meeting poll makes it difficult to prove which question each response answered. Declaring a result before all votes are counted and verified — or releasing partial figures to a select group — is a reputational risk that is entirely self-inflicted. Where an employer or union wants an arm’s-length process, engaging an independent returning officer to plan the timetable, review the rules and certify the outcome removes most of these concerns at once. For multi-site or high-volume votes, running the ballot through a dedicated online enterprise agreement platform with a verifiable audit trail gives the same protection at scale and provides a defensible record of every step.
Pitfall three: weak communication, low turnout and poor records
Turnout is the quiet legitimacy problem. A result carried by a small fraction of the workforce may be technically valid and still leave the agreement with little moral authority — and a thin margin makes any procedural complaint far more consequential. Low turnout usually reflects communication failure rather than apathy. Information sessions scheduled only for day shift, dense documents circulated without a plain-English summary, and reminders that go out once and never again all suppress participation. Employees who do not understand what they are voting on tend to abstain rather than guess.
The antidote is a communication plan built around the shift patterns and locations of the actual workforce, not the head office. Multiple channels, staggered sessions, translated materials where relevant, and scheduled reminders — including SMS where contact details allow — measurably improve participation. So does separating the vote from the negotiation: employees who feel the decision has already been made for them disengage, while those who understand the timeline, the questions and the consequences tend to vote.
Finally, records. An audit trail is not bureaucracy; it is the evidence that protects everyone involved. Records should show who was eligible, who was issued credentials, when the ballot opened and closed, who observed the count and how the result was verified — without recording how any individual voted. Retention periods should be documented, and paper ballots, if used, need a secure chain of custody from printing to disposal. Where the result is reported onward, whether to a board, a committee or the regulator, the figures should reconcile precisely to the underlying records. Organisations that can produce this evidence at short notice rarely have to argue about process; the record speaks for them. Those who cannot often find that a single missing document turns a routine ballot into a months-long dispute.
Singapore fintech auditor biking through Buenos Aires. Wei Ling demystifies crypto regulation, tango biomechanics, and bullet-journal hacks. She roasts kopi luwak blends in hostel kitchens and codes compliance bots on sleeper buses.