Summary:
1. The click-through rate of display ads in the industry is notoriously insignificant, less than 0.1%.
2. The revenue increase of post-display optimization is ten times higher than that of post-click optimization.
3. In controlled tests, customers who saw IMVU (the name of a virtual world) ads were 10% more likely to become paying users, regardless of whether they clicked on the ads.
A century ago, John Wanamaker once said, "Half the money I spend on advertising is wasted; the trouble is, I don't know which half." Today, online marketers are still trying to overcome the same problems in measurement analysis. question.
The answer seems simple, because in the online world, you can track clicks. The problem is that clicks and click-based analytics don’t hold up. Not only does the number of clicks fail to tell the whole truth, it can even turn it upside down, especially when used alone.
Because of the availability of web analytics tools, many marketers attribute website activity (engagement, conversions – semwatch editor’s note) solely to click-based campaigns, such as clicks on display ads. However, this is a very limited approach.
Because the click-through rate (CTR) of display advertising is very low, less than 0.1%; most people who see online advertising will not click on it. In addition, the number of clicks is not proportional to the number of clicks. About 85% of clicks come from 8% of people. Many industry studies have been conducted on this issue.
However, a low click-through rate doesn’t mean the ad isn’t working—quite the opposite, in fact. Consumers often make a purchase shortly after seeing an ad without clicking on it.
In a recent test, a virtual social network called IMVU that allows you to purchase virtual items tried to find out what happens when free IMVU users (those who have received marketing emails and seen ads in the virtual world) are in the real world. Are you more likely to become a paying user when you see IMVU's online ads?
In controlled tests, customers who saw an IMVU ad were 10% more likely to become a paying customer, regardless of whether they clicked on the ad. This 10% increase is in addition to all existing marketing efforts compared to the control group. The control group had the same chance of seeing other marketing campaigns as the test group. The only difference between the two groups was whether they actually saw the ads. The test group saw IMVU ads, while the control group saw irrelevant ads.
IMVU used the same method to test whether paying users would be willing to spend more money if they saw advertisements that stimulate consumption in the real world. On average, IMVU members who see ads promoting virtual products spend more than twice as much as those who see irrelevant ads, regardless of whether they click on the ads. Again, this boost is in addition to promotions via email and virtual worlds. Companies like IMVU sell virtual items as if they were printing money.
Let us focus on an e-commerce company again. The company relies heavily on website analytics tools to analyze post-click user behavior data (to optimize marketing campaign effects - semwatch editor's note) (post-click date) (the website's traffic and revenue are brought by advertising). Advertisers hope to use only post-click user behavior data as the basis for optimization. Since the client does not track display-related revenue (post-view revenu), there is no way to optimize it.
Let’s review these two situations: optimization that allocates conversion contributions based on post-click data (post-clicks) vs. optimization that allocates conversion contributions based on post-impression data (post-view). The incremental revenue from post-impression optimization is ten times higher than post-click optimization. When analyzing revenue from a post-click perspective, we call the best ad Ad A and the worst ad C. But when analyzed from a post-show perspective, the results are exactly the opposite. C is the best and A is the worst. This leads to completely different optimization solutions.
One might argue the opposite and argue that post-impression analysis overstates the work of online advertising. Because a potential consumer is likely to purchase a product regardless of whether they have seen an online advertisement, and these advertisements probably did not influence their decision. However, after testing again and again, we found that the results are exactly the opposite. We analyzed the window of time between seeing an ad and purchasing a product. Data shows that the rapid increase in conversions occurs within a short period of time after consumers see the ad, which reflects the impact of post-impression attribution. In the example below, half of the conversions occurred within six hours of the ad being shown, and 70% of the conversions occurred within 24 hours of the ad being shown. If post-impression attribution had no such effect, we should see conversion rates randomly distributed over time to follow a linear pattern rather than a curvilinear pattern.
The bottom line is that every advertising campaign is different. They should all be optimized based on as much data as possible. Don’t rely solely on click-based analytics. It's better to make the most of your strengths and resources.
Commentary from Tianan:
Analysis of the effectiveness of display advertising should refer more to various factors, such as:
1. What is the purpose of display advertising? Is it for branding or sales promotion? At which stage of consumer decision-making does this advertisement hope to have an impact?
2. What are the characteristics of the industry you are in? How long is the consumer decision-making cycle?
In fact, display ads are paid media, and the website you enter after clicking is owned media. But in fact, both are media that advertisers can control. They can decide the content, method and content to be displayed. time. For advertisers, clicks are just a transfer from the display of one information module to the display of another information module.
From the consumer's point of view, both are conveying information to him. The difference is only the amount of information and his own reading focus. Clicks represent a relatively strong degree of participation, but display is also a transmission of information. The impact of this information transmission is the "top of mind" in decision-making. For example, if I saw an advertisement for fitness equipment, maybe I glanced at it but did not click on it; but if I have similar needs and search for fitness equipment through a search engine again, I see the same name because it is the second advertisement. impressions, so I might have a sense of familiarity/trust that leads to a click.
The current status of display advertising data analysis may be largely restricted by the difficulty of data collection. In owned media, with the popularization of website analysis technology, click data is easily collected and can be easily applied in practice. However, in paid media, especially in the media environment of the domestic display advertising market, A lot of data cannot be collected or shared with advertisers, so it is destined to be ignored in analysis and optimization. What's more, in order to reach the above conclusions, a large amount of data needs to be analyzed very carefully. This can be regarded as a pain in the development process of the industry.
Original text: http://www.imediaconnection.com/content/29020.asp
The author, Jarvis Mak, was born in biology and served Yahoo and Neilson respectively in customer analysis including the MegaPanel project. Now focusing on digital media and marketing in the retail industry. ,
Translation source: http://semwatch.org/