一位软件开发者近期对ChatGPT的付费广告系统进行了为期近一个月的测试评估。1实验期间（2026年8月5日至8月28日），ChatGPT建议的每次点击竞价约为2英镑，但开发者实际支付的竞价仅为0.15-0.07英镑。1然而，这套广告系统的表现令人失望——整个测试周期内的转化率仅为0.46%。1
为了确定问题所在，开发者与广告欺诈专家Dr Augustine Fou合作进行深入分析。1分析排除了点击欺诈和广告创意设计缺陷等因素后，Fouanalytics的检测数据显示，30%的流量存在"多个红旗"或"一些红旗"的异常表现特征。1进一步观测发现，88%的来自ChatGPT的访客会移动鼠标操作，但仅有10%的访客实际点击任何内容。1Dr Augustine Fou的最终结论是，问题的根源不在于机器人流量，而在于ChatGPT的广告定向能力不足，吸引到的访客总体质量较低且商业价值有限。1
A software developer conducted an experiment testing ChatGPT's paid advertising platform and discovered significant performance shortcomings. Between August 5 and August 28, 2026, the ads achieved a conversion rate of just 0.46%, substantially lower than typical advertising sources 1. Despite ChatGPT's recommendation of approximately £2 per click, the developer actually paid between £0.15 and £0.07 per click 1.
To identify the root cause, the developer partnered with advertising fraud specialist Dr. Augustine Fou for analysis 1. Fouanalytics examination of the traffic revealed that 30% of visitors displayed "multiple red flags" or "some red flags" in their behavior 1. Additionally, 88% of traffic involved users moving their mouse without clicking on any content, with only 10% of visitors clicking on anything at all 1. After ruling out bot activity and issues with ad creative, Dr. Fou concluded that the problem stemmed not from fraudulent clicks but from ChatGPT's poor audience targeting capabilities, which resulted in attracting low-value visitors 1.
评论
还没有评论，欢迎留下第一条。