What is the paid search effect? Here are four revelations for corporate promotion

"I know that half of my advertising expenses are wasted, but unfortunately, I don't know which half is wasted." This wise saying can be called the " Goldbach Conjecture " in the advertising industry. The person who posed this difficult question was John Wanamaker, known as the father of the department store and a very successful retailer.

Advertising plays a role in promoting the survival and development of enterprises, and advertising expenditures also account for a large part of corporate expenses. But measuring the effectiveness of advertising spending has always been a challenge. With the development of information technology, advertising expenses invested in the Internet and mobile Internet have grown rapidly. Of this advertising cost, paid search (mostly SEM ) accounts for a large portion.

The IAB's recently released "Internet Advertising Revenue in the First Half of 2016" report shows that advertising revenue, including PC search and mobile search, accounts for 49% of US Internet advertising revenue. Search giant Google generates hundreds of billions of yuan in annual revenue from search advertising .

In the Internet age, users' various online behaviors can be better tracked and recorded, which provides a basis for us to evaluate which half of the advertising costs are wasted.

Let’s look at some empirical studies first

eBay Research Lab and several researchers from the University of Chicago and the University of California, Berkeley, jointly designed an experiment to study the impact of search placement on eBay. In the United States, eBay spends a large amount of money on paid search (mainly SEM) every year.

eBay’s paid search keywords mainly include two categories: brands and words containing brands, such as eBay, eBay mobile phone, eBay TV, etc.; words without brands, such as mobile phone, TV, memory, etc. This experiment verified the effectiveness of the two types of keyword placements. The effectiveness was measured by the increase in user visits brought to corporate websites by paid placements, or the increase in revenue brought to them .

When placing the first type of keywords, many companies, especially some well-known companies, also like to purchase brand areas in search engines , similar to the first display area shown when we search for JD.com in Baidu.

In eBay's experiment, several researchers used massive amounts of data and found that whether it was putting eBay keywords alone in the search engine, or eBay plus keywords of a certain category, such as eBay shoes, there was almost no effect. After these keywords stopped being used, most users entered eBay through non-paid links.

Let’s look at the situation of placing the second type of keywords.

eBay has a library of more than 100 million keywords, which are placed on the search engine platforms of Google, Microsoft and Yahoo according to specific algorithms. These keywords are not like brand keywords. They will be ranked at the top even without natural search results.

If non-brand keywords are not placed, there will be great uncertainty in the display position on the search engine. The experimental results of non-brand keywords show that SEM has a small and very insignificant contribution to sales revenue. Users will not buy more products after seeing eBay’s paid search promotion , which also makes the advertising revenue negative.

To further study the impact of different users, several researchers divided users into two groups: users who frequently shop on eBay and users who rarely or have not shopped on eBay in the past year. The research results show that SEM has significantly led to an increase in new eBay registrations and shopping behavior on eBay among the latter group of users. But for the former user group, SEM does not bring any positive impact. More specifically, users who have shopped on eBay more than three times in the past year already have a certain understanding of eBay and are less affected by SEM.

In contrast, new users who have never shopped on eBay are more likely to be influenced by advertisements and register on the site. Users who have shopped on eBay once or twice in the past year are more likely to be influenced by advertising and increase their shopping frequency. In a paper published in 2001, Ackerberg stated that users who have not used a certain product are more easily influenced by advertising than users who have already used the product. This also leads to a similar conclusion from another level.

The experiment on eBay produced a negative evaluation of paid search advertising, but Daisy from Lehigh University and Michael from Harvard Business School conducted an experiment on Yelp, the largest restaurant review website in the United States, and came to a completely opposite conclusion. Yelp can be seen as a vertical search engine focused on the catering industry, where restaurants can pay to have their search positions ranked at the top.

In the experiment, Daisy and Michael ran search ads for 7,000 restaurants over a three-month period that had never run such ads on Yelp before. Since then, Daisy and Michael have closely monitored the number of views on the Yelp pages of these restaurants, as well as other indicators that measure whether users are inclined to consume in-store, such as looking up addresses, calling restaurants, etc.

The results showed that the effect of the advertisement was quite significant. These restaurants saw a 25% increase in Yelp page views, an 18% increase in user address requests, a 9% increase in visits to the restaurant's official website, a 13% increase in phone calls to the restaurant, and a 5% increase in restaurant reviews on Yelp. Ultimately, all of these metrics combined resulted in an 8% increase in restaurant revenue.

Daisy and Michael demonstrated the positive effect of search advertising on restaurants on Yelp, but this does not contradict the results of the eBay experiment. eBay's paid search works better for new users or less active users. Most restaurants on Yelp are relatively small and unknown, so paid search ads are more likely to be presented to new users or less active users, so the effect will be better.

How is paid search wasted?

The empirical research above shows that paid search advertising is more effective for new users or users who are not very active, but most of the costs of search promotion are occupied by some active old users. Everyone knows that SEM is paid on a click basis. Old users are more likely to click on content they are familiar with, but these clicks do not contribute to incremental revenue. For old users, even if they do not see the promotional information of the product in the search engine, they will find and purchase the product by directly visiting the website or other forms. Regarding this part, let’s look at a case first.

Suppose you are the head of marketing for a women’s jewelry retailer and are listening to the digital marketing team report to you on the effectiveness of the company’s online marketing campaign . The team tracked user behavior during this marketing campaign and analyzed the impact of user behavior on sales. They found that the sales conversion rate for users who were not exposed to the ad was 0.02%, the sales conversion rate for users who were exposed but did not click on the ad was 0.9%, and the sales conversion rate for users who were exposed and clicked on the ad was 2.8%. From this, they came to two conclusions: advertising worked and increased sales conversion rates; and stimulating users to click could further and significantly increase sales conversion rates. Is this conclusion correct? Should the corresponding strategy be adopted?

Because you are a very wise leader, you asked your colleagues in the membership management department to carefully compare the characteristics of these three types of users in the CRM system. The result may surprise you. The three types of users mentioned above have very fundamental differences in attributes. The users who were not exposed were basically men, and the vast majority of those who were exposed were women. Men are not the target consumers of the product at all.

Going further, you find that 14.1% of users who click on the ads are the company's most loyal members, while this proportion is only 2.3% among users who do not click on the ads. This finding makes the above strategy recommendations completely untenable. Users are not exposed because they are male, and the click-through rate is high because they are already loyal.

In paid search promotion, the situation we described above often occurs. Companies doing promotion will first choose keywords that best reflect their advantages, and then screen the target user groups that best match them. For some well-known large companies, the user groups targeted by paid search are often already old users of the company.

Paid search is paid on a click basis, which means that most of the clicks are actually contributed by old users. This is why the overall return on investment in the eBay experiment above is negative. It should be noted that for the company's old users, even if they do not see the paid promotion , they will complete the purchase by directly entering the company website or other means. In the experiment, the researchers eliminated this part of the interference by setting up a control group and adopting a double difference model.

What insights can we draw?

The effectiveness of paid search promotion depends on whether it can bring fresh and useful knowledge to users, which is actually very consistent with the scenario in which users use search engines. When users turn to Google or Baidu for help, most of the time they have encountered something they don't understand and need to find an answer. At this point, if users are presented with conclusions that they already know or are obvious, it will be difficult to meet their needs. This is an important principle that needs to be grasped in paid search promotion.

Therefore, returning to the above experiment, we can find that when users have no intuitive concept of the products or services provided by the company, search ads will play a greater role. Some well-known brands, such as eBay, have very high visibility and a wide audience, but do not benefit much from paid search. Part of the positive benefits come from users who have never used eBay or have not been active on eBay in the past year.

Because most users already know their brand and have a general understanding of the products and services they provide. In contrast, most of the restaurants on Yelp are small, unknown brands, so the effectiveness of their advertising is much better. This also indirectly shows that paid search (SEM) plays more of a notification or reminder function and has no persuasive effect.

Combining all the above information, we can draw the following insights on the paid search promotion strategy of enterprises:

1. Some well-known large brands may consider reducing their search advertising budgets, but this does not mean that they do not need to advertise. Instead, when placing search ads, it may be more effective to target new users or old users who have been inactive in recent period. This is the key to the success of search advertising. However, this involves the strength of search engine data support.

2. When placing search ads, try to make consumers aware of things that are not within their knowledge system . For example, when placing an ad on JD.com, you can set more keywords such as clothing and cosmetics that perform better on Taobao , while the strategy should be the opposite when placing an ad on Taobao or Tmall . Well-known brands should use paid search to promote products or services that are not in the minds of consumers.

3. For some less well-known smaller brands, increasing search advertising budgets may be beneficial. As I write this, I think of Baidu’s medical search ads that have been criticized for a long time. It is precisely because users come here to seek knowledge, and do not understand most of the knowledge, that the effect of this type of advertising is so good. This also puts Baidu under great regulatory and public pressure.

4. There is still a long way to go to achieve refined delivery. If we can understand more about user data when we launch advertising, we can set different keywords for different regions, different user groups and different categories, so as to meet users' expectations for search ads to the greatest extent and maximize the benefits of advertising. But this obviously requires more data, especially the connection of data between enterprises, to form a portrait of users across the entire network. There is still a long way to go, but it does point out a direction for effort.

Mobile application product promotion service: APP promotion service Qinggua Media advertising

This article was compiled and published by @贞元(Qinggua Media). Please indicate the author information and source when reprinting! Site Map