Political parties targeted Facebook and Instagram ads by gender — and most aimed at men
Between 1 January and 7 August of the 2023 election year, the National Party ran more than 500 ads on Facebook and Instagram. According to an RNZ analysis of Meta advertising data, 92 of those ads were seen exclusively by men. Only one was seen solely by women.
The male-only ads cost National up to $21,000 and were seen up to 3.9 million times. The single ad seen only by women ran on 12 June, cost less than $100, and was estimated by Facebook to have reached between 100 and 999 women aged 65 and older. National denied placing it.
RNZ analysed advertising data made public by Meta — the parent company of Facebook and Instagram — covering the first seven months of the election year. Ads paid for by Parliamentary Service, which funds MPs' communications rather than campaign spending, were excluded. Across all parties, more than $200,000 went to Facebook during the period.
Meta's Ad Library — a public database of all ads running on its platforms — shows who saw an ad but not the targeting settings a party selected. An ad seen only by one gender is therefore highly likely to reflect a deliberate targeting choice rather than a quirk of the platform's algorithm.
A National spokesperson said the party's ads were targeted to a range of demographics, with 80 percent seen by both men and women. The spokesperson also denied Meta's record of the ad aimed at women. Meta's Ad Library page for the National Party shows some content was removed because its disclaimer did not follow Meta's policy for ads about social issues, elections or politics.
ACT placed 63 ads during the period analysed. Fourteen were seen solely by men, reaching up to 518,000 impressions (the number of times an ad was displayed on screen), and none were viewed solely by women. ACT spent up to $4,000 on its male-only ads. An ACT spokesperson said the party does not share details of its digital strategy with the media and that its values and policies are for all New Zealanders.
The Green Party targeted 13 of its 103 ads at women and none solely at men. Those ads cost up to $5,000 and were seen up to 768,000 times. A Green Party spokesperson cited limited funding and resources as reasons for the absence of male-only ads, saying the party must be strategic about where it places resources.
Labour did not run any advertisements funded by itself during the timeframe. NZ First did not run any paid ads on Facebook or Instagram.
The findings give concrete evidence to a transparency debate that has been live since at least the 2020 election, when Stuff reported that Ad Library data showed paid posts from the Greens, Māori Party, TOP and NZ First were slightly skewed towards users identifying with a particular gender. The Spinoff similarly reported on Meta and Google advertising spend during the 2023 general election. An academic analysis published in Internet Policy Review in 2022 considered whether the Facebook Ad Library meaningfully improves the ability of regulators and the public to oversee online political advertising (Internet Policy Review).
The broader context here is that New Zealand's electoral law has not kept pace with what platforms themselves disclose. The Electoral Act sets spending limits and requires authorisation statements on published material, but it does not compel parties to reveal the audience settings they apply to digital ads. The gender-disaggregated data now available through Meta's Ad Library gives journalists and researchers a window into targeting decisions that parties are not required to disclose.
For the parties, the strategic logic of gendered targeting is not new. Micro-targeted digital advertising — delivering different messages to different audience segments at relatively low cost — is well established in comparable jurisdictions. What stands out in this dataset is the asymmetry: National and ACT concentrated their single-gender targeting on men, the Greens on women, and Labour and NZ First made no comparable effort during the period.
National's denial of the single female-targeted ad raises a practical question about how reliable platform-level data is as an accountability tool. Meta's Ad Library records who saw an ad, not who placed it or how targeting was configured. Parties can dispute the record, and the platform's own removal of non-compliant content shows the data is not always clean. For anyone using the Ad Library to hold political actors to account, that is a material limitation worth keeping in view.
