網路廣告點擊率之時間趨勢及其影響因素探討

Abstract

在Web 2.0時代中網路的發展飛快，根據台灣網路資訊中心的統計，2010年台灣網際網路使用人口已由2006年的12,789,114人成長到14,669,915人。在政府以及電信業者的推波助瀾之下，網路使用人口還在持續地向上攀爬。回顧近年有關網路廣告的文獻多數著重在網路瀏覽者的使用習慣、忠誠度、網頁設計與廣告效果衡量等，大部分都是以理論探討廣告或網站本身相關的議題，少有使用實際網路瀏覽者實際線上資料進行分析的文獻。本研究使用酷盈科技的實際線上瀏覽者瀏覽紀錄進行資料採礦，以網站廣告做為研究範疇，並透過階層線性模式中的成長模型，探討網路廣告點擊率受到不同廣告變數而呈現何種不同趨勢。

本研究研究目的為點擊率時間趨勢的探討，試圖了解下列四種廣告變數如何牽動時間趨勢：一、不同廣告分類是否有顯著的點擊率趨勢差異；二、不同廣告型態是否有顯著的點擊率趨勢差異；三、不同廣告投放時間是否有顯著的點擊率趨勢差異；四、不同網站類別是否有顯著的點擊率趨勢差異。結果發現精品美容服飾類的點擊率趨勢表現最佳，而電腦手機相機類的點擊率趨勢表現最差。廣告型態為靜態的廣告點擊率隨時間下降較動態緩慢。廣告投放在寒暑假的點擊率下降較非寒暑假緩慢。廣告點擊率在非入口網站的表現較入口網站優異。所有的廣告變數可以整理出20個排列組合，本研究將所有廣告變數的方程式都估計出來。希望這些點擊率的時間趨勢可以提供廣告主與廣告投放平台做為廣告輪替更換時間點的依據。

Online advertising market has been growing rapidly. According to the Taiwan Network Information Center Statistics in 2010, the active internet universe in Taiwan has grown by 15% during the past five years, and it increase continuously. Reviewing the previous studies, most of them only focused on the browser's habits, loyalty, web design and advertising effectiveness measurement, etc. Hence, there are few empirical studies about this issue. This study use the ADQoQ’s real online visitor browse records as the source date of mining, then using growth model to analyze what factors cause click-through rate of online advertising to have different trends.

This study would like to discuss the time trend of click-through rate for online advertisings, trying to figure out the following questions: Does different types of online advertising have significant impact on banner click-through rate trend? Does different categories of online advertising have significant on banner click-through rate trend? It has significant difference of click-through rete trend whether advertisement releases in vacation or not. Does different websites have significant impact on banner click-through rate trend? The results show that the performance of click-through rate trend of jewel advertisement is the best, but the performance of click-through rate trend of computer advertisement is the worst. Comparing to dynamic banners, the click-through rates of static banners decrease more slowly. The advertisement which is released in vacation period can attract more browsers. All advertising variables can be organized 20 permutations and combinations. This study estimates all of those time trend equations. With regards of these finding, this study can provide insights for advertisers and advertising agencies.