Add Thesis

The true effects of programmatic display marketing

A study on how advertisers could make use of programmatic in the different stages of the customer journey

Written by Stina Hübinette

Paper category

Master Thesis

Subject

Business Administration>Marketing & Sales

Year

2017

Abstract

Master Thesis: Programmatic and real-time bidding As computers become smarter and algorithms continue to take over manual work, all industries must adapt. In terms of marketing, the display business is divided into direct sales and programmatic display. Judging from the trend of Google search term data, interest in this type of advertising has increased sharply in recent years. Programmatic display advertising is intelligent display advertising based on algorithms and big data. (Zhang, 2017) The concept is that advertisers run their banner ad campaigns on the publisher's inventory through an ad exchange platform. Artificial intelligence (AI) technology has algorithms that can help advertisers identify user behavior and demographic data so that they can make more informed strategic decisions, rather than guessing where, when, and how to reach the target group. For advertisers with the highest real-time bids, auctions for each location can take up to a few milliseconds. Compared with traditional direct sales, passive programmatic guarantees or active programmatic guarantees, real-time bidding benefits both advertisers and publishers because the model is based on continuous auctions, similar to the stock market, thus forming a supply and demand market. (Chen, 2017) 2.2. Tracking programmatic results In order to determine the effectiveness of digital marketing, advertisers need to gather insights from already running campaigns. Unlike traditional marketing channels such as television or print, digital marketing allows advertisers to obtain a variety of data. This section will explain some of the core concepts used to track programmatic display marketing and give a specific example to illustrate how to do this. 2.2.1. The publisher’s page for impressions and viewability contains one or more placements that can only be used for ads. When an ad is displayed on such a placement, it is an impression. Displays may not always be visible, for example, they may be displayed just below the fold, or they may scroll too fast to be counted. To count impressions as "viewed" in most analytics tools, at least half of the ads should stay on the screen for at least 1 second. (Miller, 2015) 2.2.2. CPA and CVR Cost per operation (CPA) is a method of calculating the cost of an advertiser’s required actions (such as signing up for a newsletter or purchasing a product). CPA is calculated by dividing spend by the number of conversions (actions). Conversions can be post-view conversions or post-click conversions, depending on the advertiser’s preferences. Post clickconversions is the most common measurement, which only counts users who convert after clicking on the banner. On the other hand, Conversion Rate (CVR) is a way to measure the ratio between the number of "actions" divided by the expenditure. 2.1.6. Attribution model The most commonly used attribution model in the industry is the last click model. In this model, the credit for the action required by the advertiser is assigned to the last channel that the user touched before converting in the required way. The difficulty of attribution is to ensure that each marketing channel gets the appropriate credit line. The multi-touch attribution model will more relevantly indicate which advertising channel should provide credit for each required user action. Such a model would attribute credit to every touchpoint in the customer journey, not just the channel closest to the conversion time. According to Shang He Li’s research, the channel that gets the least credit from the last click model is display advertising. Therefore, using a multipoint attribution model can improve advertisers’ understanding of how much each channel contributes, which will also help understand how marketing budgets should be allocated between channels. (Shang, Li, 2009) Looking at the last marketing channel a user interacted with can create a misleading impression of which channel has the greatest influence on the user's decision. Looking at the user’s several touchpoints with a brand before making a purchase will show that the user’s last touchpoint may not have the greatest impact on the decision. These charts show why the last click is not the best attribution model. Since 99% of the last touchpoint before conversion is made up of some sort of search channel, 99% of the marketing budget should be allocated to search and 1% to impressions. However, looking back at Touchpoint 5, 49% of all users interacting with the brand at the time can be attributed to the display. Universal searches for the same touchpoint accounted for only 16%. (Wiberg, 2017) Therefore, display is a channel defined by the industry as an auxiliary channel. In most cases, display advertising does not drive direct conversions, but instead passes users to other marketing channels, such as search where conversions occur. Therefore, search received credit for the conversion. Advertisers continue to increase their budgets spent on digital marketing channels. (IAB, 2017) However, most advertisers still formulate their digital marketing strategies based on the click-through rate and cost-per-conversion that each channel can generate. As shown in the figure above, adding display channels can not only increase website traffic and brand awareness, but also improve the performance of paid search and promote organic traffic. Research by Kireyev et al. proved that, in the long run, paying too much attention to performance-driven channels will cause advertisers to lose conversions, because auxiliary channels such as display will directly and over time improve search performance. 2.1.7. Methods of measuring display advertising campaigns and return on investment It is difficult, if not nearly impossible, to measure return on investment from display advertising campaigns. Read Less