Explain the basic logic of Internet advertising all at once!

Since the advent of the Internet, the age-old industry of advertising has been changed forever.

I always think that popularizing the basic logic of Internet advertising is something meaningful to the industry, so I made a 40-minute video to share with you.

I hope this popular science course video can achieve two goals - first, even novice users who have never worked in advertising can understand it and understand the logic behind it. Second, even industry veterans can be inspired and gain something ;

It is not easy to achieve this goal at the same time. Everyone is welcome to leave a message. If you don’t like watching the video version, you can directly watch the text version below!

The following is the full text version:

Today, let’s talk about the basic logic of Internet advertising.

“It’s everywhere, you can’t escape it.”

This was a slogan when WeChat Moments ads were launched. Indeed, ads have become a part of our lives. When it comes to ads, many people’s consistent response is: “Don’t talk to me about ads, I hate them!”

But I would say that advertising is extremely important.

You may say that you are in advertising, so of course advertising has meaning. The meaning I am talking about is not only for advertisers and advertising media, but also for each of us, advertising has important meaning.

So what exactly is the point of personal ads?

The first overlooked significance of advertising is that it can lower the price of goods.

You may wonder, where do Volkswagen spends $6.6 billion on advertising every year, Coca-Cola spends $5 billion, Procter & Gamble spends 8.2 billion, and Unilever spends 8 billion euros? Aren't we, the consumers, paying for it?

If he didn't advertise, I would definitely buy a car much cheaper, but is it really like that?

Let’s look at two cases:

As a first example, the price of a box of Coca-Cola, which spends $5 billion on advertising on JD.com every year, is $59.9.

The same specification is 500ml, 24 bottles, and it has never been advertised, so you may never have heard of Laoshan Cola. The price is 68 per box.

You see, spending money on advertising actually makes the product cheaper.

Let's look at a natural example:

In 1972, Professor Benham of the University of Chicago published a paper in which he studied the prices of glasses in different states in the United States in the 1960s.

The results showed that the average price of glasses in states that allowed advertising was $26, while the average price of glasses in states that did not allow advertising was $33.

If you don't advertise, it will be more expensive, but if you spend money on advertising, it will be cheaper. So why is this?

The answer is the scale effect brought by advertising. Through advertising, consumers can increase the scale of production, and the increase in scale will in turn reduce the average price.

Next, let’s look at the second overlooked meaning of advertising, which is that it can reduce our cost of selection.

If all shampoos now do not have advertisements, how will our supermarkets choose anti-dandruff shampoos?

We have to read the instructions at least many times and pick out the ones with anti-dandruff function. We may also compare their ingredients, origin, efficacy and price;

Compared to walking into the supermarket and taking a bottle of Head & Shoulders directly, the series of actions we just took is our choice cost.

The third overlooked meaning of advertising is that it allows us to use many products for free.

Google is free, WeChat is free, Weibo is also free. Some people say that these have always been free.

Let’s not forget that Windows and Photoshop without ad mode are so expensive:

The retail price of Windows is 1088, and the surprise price of Photoshop is 3485.

Some people say that they have never paid for Windows. That's because you are illegally using pirated versions, or it is already included in the price of the computer when you bought it.

Advertising, and even lowering the price of hardware. In the United States, the ad-supported version of Kindle is $25 cheaper than the ad-free version.

So advertising is of great significance.

The most important change that has occurred in the advertising industry in the past few decades is the shift from traditional advertising to Internet advertising.

But we may still have many doubts about Internet advertising——

• For example, the first three advertising groups launched on WeChat Moments were Coca-Cola, BMW, Zhongbo and vivo. Some people would ask, why did I see Coca-Cola, but my husband saw BMW? Is my husband an invisible rich man?
• Some people may wonder, isn’t it about precise delivery? Why do I, a boy, see Always ads on my phone?
• Also, why did I see a camera on JD.com during the day, and then see an advertisement for this camera when I was browsing Toutiao at night?

The above three questions reflect that we still don’t quite understand the operating model of Internet advertising.

Not only do we not know, but even US senators do not know:

During the congressional hearing of Facebook CEO Zuckerberg, Senator Hatch asked such a question: "Isn't Facebook free? How do you make money?"

This was an easy question, so Zuckerberg answered easily: "Congressman, we have ads."

Advertising is the largest business model of the Internet. How does Internet advertising work? What is its basic logic? What significance does understanding these basic logics have for our advertising?

This is the topic we are going to discuss today. I will analyze the most basic operating logic of Internet advertising from both macro and micro perspectives.

1. Macroscopic part

Let’s first look at the macro part and start with an extremely macro figure: the proportion of advertising in GDP.

First, let’s look at the data for the United States. From 1919 to 2017, the proportion of advertising in US GDP remained basically unchanged, hovering between 2% and 5% for a long time.

The proportion of advertising to GDP has remained basically unchanged, and this is not unique to the United States. The same is true in China, but China's proportion is lower than that of the United States, only 0.88%. In comparison, China's defense spending accounts for 1.3%.

You can see the growth trend chart of Google's revenue in the course. We can see that Google's annual growth rate is actually more than 20%. We just said that the proportion of advertising in GDP remains unchanged, but GDP growth is actually relatively slow. For example, China is about 6%, and the United States is about 1%.

So how do advertising-centric Internet companies like Google achieve rapid growth of up to 20% per year?

Some people say that it is because Google is a leading company and its performance is better, but in fact, the advertising revenue of mid-tier companies like Meitu is also growing rapidly, and its growth rate is even faster than that of Google.

Then someone may say, since the top companies are growing faster, the small media at the bottom must be suffering, right?

In fact, the advertising expenses of long-tail media such as my public account "Wei Xi Zhi Bei" this year are much higher than last year.

Then the question arises - the total advertising pie remains basically unchanged, or is growing extremely slowly, while both large Internet companies and conventional media advertising revenues are growing rapidly, so who is actually falling?

The answer is that traditional advertising is declining and has been replaced by Internet advertising.

Whether in China or the United States, while Internet advertising is booming, traditional advertising is inevitably declining.

You can see the trend chart of the global advertising market in the course. In 2000, the share of traditional advertising was almost 100%, but today, 20 years later, its share has dropped to 50%. That is to say, half of the global advertising share has been replaced by Internet advertising.

Then the next question will be - why did traditional advertising become Internet advertising?

The answer is actually very simple, because the Internet takes up everyone's time.

Let's look at an interesting data change - Internet analyst Mary Meeker publishes an annual report on the Internet every year. Since 2009, there is a page of PPT every year to disclose the proportion of advertising revenue and user time of various media.

We directly compared the data from 2009 to 2018. The share of print media dropped from 26% in 2009 to 7% in 2018, while the share of the Internet increased from 13% in 2009 to 51% in 2018. The Internet also happens to occupy 51% of people’s time.

The reason why we say that advertising is the largest business model on the Internet is that the most profitable Internet companies in the world, such as Google, Facebook, Alibaba, etc., almost all make a living from advertising.

Whether in China or the United States, the Internet advertising market is basically divided up by several giants.

The next question is - why do these giants have such a large market share?

Of course, the giants occupy most of the user time. From the pictures in the course, we can see that Tencent, ByteDance, Alibaba, and Baidu occupy more than 70% of the user time on China's Internet.

So what is the difference between Internet advertising and traditional advertising?

First, Internet advertising changes the structure of the advertising market;

Second, Internet advertising changes the logic of advertising implementation in terms of technology and products.

An important change that the Internet has brought to advertising is that advertisers have become more diverse.

If I ask – How many advertisers does CCTV have in a month? How many advertisers can Ogilvy serve in a month? How many commercials can be aired during the Super Bowl? The answer is: very little.

Then let’s take a look at how many advertisers Facebook has in a month? How many advertisers does Google have in a year? How many advertisers does Baidu have in a quarter?

The answer will be more than you can imagine -

According to its financial report, Facebook had more than five million monthly active advertisers in 2017;

Google has not disclosed how many it has, but in 2018 alone, Google blocked 723,000 violating accounts.

According to Baidu's financial report, there were as many as 529,000 advertisers in Q4 of 2018.

Then let’s see how long it takes from start to finish if we run an advertisement on CCTV. This is definitely a very long process, but how long does it take for us to run an advertisement on Kuaishou, Weibo and Douyin?

The answer is one minute. We select advertising content, ad targeting, and pay immediately on our mobile phones, completing an advertising campaign very easily.

So why can advertising in the Internet age do this?

2. Microscopic Part

Next, we will move on to the micro part of advertising logic.

Many people face a lot of doubts when placing Internet advertisements.

For example, placing advertisements on the Internet is really too complicated. Are there any principles that need to be followed?

From which aspects should I comprehensively improve the return on advertising?

Next, I will talk to you about the micro-logic of Internet advertising operations and how we should place Internet advertisements through six questions raised by a case.

You can see this picture in the course. Let’s start with an ordinary WeChat Moments advertisement. This is an advertisement made by Liu Tao for Yili that I received in Moments.

We see such advertisements every day, but advertisements are like an iceberg. The part we can see is only the part above the water surface. In fact, there are more parts hidden below the water surface. Behind it is actually a very complex system.

OK, let’s take a look at what’s underneath the iceberg. I have six questions:

• First, what information does WeChat collect to place advertisements?
• Second, why did you see this ad and not another one?
• Third, what is the difference between the placement of this advertisement and traditional advertising?
• Fourth, why is the avatar Liu Tao? What is the meaning behind this?
• Fifth, why can we only see two to three advertisements on WeChat every day, but we can see more on Weibo and Toutiao every day?
• Sixth, what do the likes and comments below the advertisement mean to the advertisement?

These six questions can actually be divided into two categories. The first three are related to technology, and the last three are related to products. Each question corresponds to a keyword. The six keywords are data, bidding, automation, native advertising, frequency control and interactivity.

Question 1: What information does WeChat collect about me to place advertisements?

What information does WeChat collect about me to serve advertisements? The corresponding keyword is data.

In the advertising industry, there is a conjecture called Goldbach's conjecture, which is:

I knew that half of my advertising dollars were wasted, but I just didn't know which half.

This was said by a retail tycoon in the 1970s, and it was essentially an indictment of the inaccuracy of advertising, but today the problem has been solved to some extent.

Internet advertising is like a matchmaker. The matchmaker connects men and women. The advertisement is based on the advertiser’s needs on one side and the user’s characteristics on the other side. The advertising system matches the two sides together.

Next, let’s talk about the evolution of advertising precision:

At the beginning, we targeted through media, such as CCTV-5 and Southern Metropolis Daily;

In the portal era of the Internet, we target through content, such as the various channels of Sina.com;

Then in the age of search, we target by intent;

Finally, in today’s era of short videos on social media, we target advertisements based on crowd attributes.

Just now we talked about what information WeChat collects to place advertisements, which involves the user portrait system of WeChat advertisements:

Because Tencent has a large product matrix, it can obtain cross-data from six platforms to build a very three-dimensional and accurate user portrait.

These six platforms are: social platforms, such as QQ; WeChat game platforms, such as the mobile games we play every day; advertising platforms, such as Tencent's Guangdiantong, brand advertising, video advertising, etc.; and media tools, such as Tencent.com, Tencent Video, etc.; as well as basic platforms and e-commerce life platforms.

Then let’s take a look at how Alibaba mines advertising data. Contrary to everyone’s impression, Alibaba actually has more than just e-commerce and payment data . Through the acquisition of Amap, UC Browser, Youku, etc., Alibaba is able to build a complete user portrait from all angles.

Then let’s take a look at ByteDance’s advertising targeting. Through apps such as Douyin, Toutiao, Xigua, Volcano, and Dochedi, Toutiao can also accurately know the user’s interests and his corresponding behaviors.

This picture in the course is Facebook's advertising targeting background. It collects relevant data through applications such as Facebook, Instagram, and Whatsapp. The targeting dimension of Facebook's advertising system is also very accurate.

Of course, the importance of data is different.

Let's compare two different types of data. In the early days of the Internet, products represented by Baidu and Taobao mainly solved useful problems . They were more about saving time for everyone, meeting existing needs, and being stock data. The data at this time was highly important, but the dimensions and volume were relatively small.

Today, products represented by WeChat Moments and TikTok mainly solve interesting problems . They are more about killing time, tapping into people's new needs, and increasing data. At this time, the importance of data is not that high, but the dimensions and total amount are much greater.

Data has revolutionized advertising.

On Google's advertising information page, it will directly tell you what information it collects about you to place ads.

You can see this picture in the course. This is the label given to me by the Google Ads Settings page——

I actually seldom use Google products. I only use search, translation and YouTube occasionally. But the amazing thing about Google is that the labels it gives me are extremely accurate. 90% of these labels are very accurate data, which makes a huge difference in the accuracy of advertising.

Xie Xiaohu, a former core member of Google's advertising system, revealed that the click-through rate of the first ad on Google's homepage is as high as 30% to 40% , which means that almost half of people will click on Google's first ad. In comparison, in the newspaper era, if 5% of people were interested in an ad, it was a remarkable achievement.

Next, let’s use two stories to intuitively understand the accuracy of the advertising system.

First, let’s talk about Zhang Yiming’s early entrepreneurial story. When Zhang Yiming, CEO of Toutiao, started his business in his early years, the company needed to use Python extensively. At that time, Python as a language was not very popular, so it was particularly difficult to recruit people.

Zhang Yiming did not advertise on vertical recruitment websites such as Lagou, so how did he do it?

He chose to go to Baidu to buy keyword advertisements related to Python , such as some functions in Python, because only Python engineers would search for these keywords. As a result, these keywords were unpopular and cheap, and the recruitment effect was also very effective.

Then let's look at the story of a Silicon Valley engineer. There was an engineer in Silicon Valley who wanted to work for a website called Reddit. He thought of a very interesting method. He first wrote a very high-level article on how to improve the recommendation algorithm of this website.

Next, he found some unique advertising targeting through the website's CEO Hoffman's public Facebook account. Then he used these targeting methods through Facebook's advertising system to deliver his article to 197 people. Among these 197 people, Hoffman was surprisingly the target. This promotion only cost him $10.6 . In the end, his article was recognized by Hoffman, and he was successfully accepted by the website.

Through these two stories, we can see that with accurate data, wide publicity will become narrow publicity.

So in this sense, Internet advertising should actually be called narrow advertising .

Data is the oil of the new era, so Facebook and Google are the Standard Oil of the digital age, and Tencent and Alibaba are the CNPC and Sinopec of the digital age.

OK, let’s summarize this section - how do we use data to deliver ads to our target users?

The answer is to fully tap into valuable data for advertising , including first-party data, second-party data, and third-party data.

The so-called first-party data is the advertiser’s own data, such as the customer’s email address, mobile phone number, etc.

Second-party data refers to the data that advertising platforms can provide, such as the data of various dimensions of Alibaba and Tencent mentioned above.

Third-party data is data provided by third-party data companies, such as data obtained through data markets.

Question 2: Why is this ad shown to a person instead of other ads?

OK, let’s look at the second question - why is this ad shown to a person instead of other ads? The keyword involved is called bidding.

Let's understand bidding through an extremely simple example -

There are three red apples and three green apples in a basket. There are two people. The first person only likes to eat red apples and not green apples, so he is willing to bid one dollar for a red apple and 0 for a green apple. The other person is just the opposite.

Now the problem arises - we have two ways of selling. The first way of selling is to sell the whole basket to one of them. At this time, our income is three dollars.

The second way is that we can sell the red apples and green apples to them separately, and our income will be three dollars plus three dollars, which equals six dollars.

If we regard advertising as the apple in the example just now, then through this example, we can draw the following conclusions:

• First, the premise of bidding is that people’s value judgments on different advertisements are actually different.
• Second, Internet technology can effectively differentiate different advertisements.
• Third, the purchase price can achieve the optimization of resource allocation.

OK, let’s go back to the question at the beginning, why do you see this ad instead of other ads?

You can see this diagram in the course. Suppose the advertising system recognizes through data that Xiao Ming is a 25-year-old man living in Beijing who is interested in fitness.

So how was the advertisement delivered to him?

Advertising system delivery is actually very complicated, but there are two key steps :

The first step is called Target, which means matching. In this step, ads that match Xiao Ming’s targeting will be found in the entire ad library, such as A, C, and E in the figure.

The second step is called Rank, which is bidding sorting. In this step, the advertisement at the front will be selected and sent to Xiao Ming, which is C in the figure.

At this time, Xiao Ming saw the advertisement C.

So the question is, how is the bidding ranking of advertisements done?

Suppose we now have two advertisers - Nike bids one dollar per click and Adidas bids two dollars per click. We have a total of 100 impressions. Who should we give it to?

Many people would intuitively say, of course it has to be given to Adidas, it offered a higher price!

But if we take the click-through rate into consideration, the result will be quite different. If Nike's click-through rate is 5%, 100 exposures will generate five clicks worth one dollar, and the revenue from each click is five dollars.

At this time, if Adidas' click-through rate is 1%, then 100 exposures will generate one click, and the income from one click is two yuan.

So for the two dollars versus five dollars, we should give it to Nike, not Adidas, because its overall profit is higher.

We should not think that this case is very simple, but it reveals several most important rules of pay-per-performance bidding advertising.

• First, advertising platforms care about the total revenue of their traffic;
• Second, advertisers usually bid based on performance, for example, this time they bid based on a single click;
• Third, the advertising platform needs to convert click bids into revenue for ranking;
• Fourth, CTR or click-through rate builds a bridge from clicks to revenue, without providing a foundation for sorting.

Therefore, in bidding advertising, click-through rate is an extremely critical indicator.

Then the question arises - the click-through rate is a result indicator, that is, after we have launched the campaign, we can count the click-through rate. Before launching the campaign, we actually don’t know it, and each campaign needs to be sorted.

So how should the click-through rate be calculated?

Many people think of a method - "Try to put it out"

So does the “give it a try” approach work?

The answer is no, why not? Let’s look down -

Let's still use a simple example to illustrate:

Suppose there are four advertisers now - one selling high heels, one selling suits, one selling dolls and one selling game consoles, and their bids are all one dollar per click.

If there are 4,000 users on the entire platform,

As I just said, we don’t know the click-through rate of each advertiser at the beginning, so we can try it out. For example, we bid 100 times for each advertiser. As a result, we find that after the delivery, the click-through rate of each of them is 25%, exactly the same. At this time, sorting becomes very difficult.

Therefore, we can only deliver the advertisement fairly and randomly. At this time, our number of clicks is 4,000, multiplied by the click-through rate of 25%, which equals 1,000. Our total revenue is 1,000, and the total revenue is 1,000 clicks multiplied by one dollar, which is 1,000 dollars.

However, if we knew the characteristics of these 4,000 users, we would not estimate the click-through rate in this way.

So what is the truth?

The truth is that these 4,000 users can actually be divided into four categories based on age and gender - mature men, mature women, little boys and little girls . Each category has 1,000 users. Each of them has a 100% click rate for items they like and a 0% click rate for items they don't like.

For example, a little boy will definitely order a game console and will never order anything else, while a mature woman will definitely order high heels and will never order any other products.

So our correct estimate at this time should be that when a little boy comes to visit, we can accurately estimate that his click-through rate for the game console ad is 100%, while the click-through rate for others is 0%.

With this estimate, everyone’s click rate will be 100%. The number of clicks will then become 4,000, and the total revenue will also become 4,000, four times the previous 1,000.

The above example tells us that we need to estimate the click-through rate in real time based on advertising characteristics and user characteristics.

So what features are needed to estimate the click-through rate, or what factors are related to the click-through rate?

There are three main aspects that affect click-through rate - advertising side, user side and platform side.

It is easy to understand the advertiser's industry advertisements on the advertising side. The copy, advertisements, pictures and the form of advertisements will affect the click-through rate.

On the user side, we just mentioned age, gender, region, mobile phone interests, and usage environment, etc.

Then on the platform side, the platform's frequency control time, bidding strategy and traffic distribution strategy will all affect the click-through rate.

Estimating click-through rate is actually a very complicated process, and it is one of the most core technologies in Internet advertising.

We can simplify click-through rate estimation into two steps. The first step is called feature engineering, and the second step is called model training.

Feature engineering is to find many features related to click-through rate, and through model training we can obtain a specific click-through rate.

In this sense, click-through rate prediction is actually a black box. By inputting many features, the estimated click-through rate can be output in real time.

The prediction of click-through rate is widely used. The core technology of the recommendation system of Toutiao and Douyin, which we are familiar with today, is also the prediction of click-through rate. What we usually observe is that they recommend what I have seen, which is very accurate.

What this actually does is to estimate the click-through rate of all the content that may be delivered to you, and then push the content with the highest click-through rate to you .

Advertising has one more step in sorting than recommendation. Recommendation only needs to know who is ranked at the top, while advertising needs to know the specific value because it needs to be ranked comprehensively together with the bid.

By understanding the significance of CTR estimation to ad ranking, we can dispel a misunderstanding about pay-per-performance:

Some people say that it’s great to pay based on clicks and transactions, and there’s no risk if there are no clicks and no money!

So can I pay one cent per click? I don’t even have time to make good advertising materials, but I still want a million clicks a day?

In fact, there is no such good thing. You pay one cent for one click. If the advertising platform generates a click, it will only deduct one cent from you. If there is no click, it will not deduct any money from you.

But the most likely result is that your bid is too low or your click-through rate is too low, so you are ranked at the back and cannot be delivered at all. So there is indeed no risk, but there is also no benefit.

OK, let's summarize this section: How to improve the efficiency of advertising through bidding. Paying by performance is indeed a billing method that is friendly to advertising groups, but it is also a fair billing method. There are only two ways to get more clicks:

• First, increase the bid. Of course, the bid should be based on your affordability.
• Second, we improve our click-through rate and conversion rate, and continuously optimize our advertising creatives, advertising targeting, and advertising delivery data.

Question 3: How was this advertisement placed?

OK, let’s look at the third question. How is this advertisement delivered ? The corresponding keyword is automation. After studying this section, we will know how to use automated tools to improve the efficiency of advertising.

Advertising, in fact, has placed great emphasis on experiments from the very beginning. In Ogilvy's time, advertising experiments were conducted like this: publish advertisements with different copywriting in the same newspaper, on the same page, and on the same date, and then observe the difference in the number of free sample requests brought about by the two copywritings.

So how are advertising experiments conducted today?

Facebook has developed a creative experiment system. By inputting different pictures, common titles and different descriptions, the system will automatically generate multiple advertising creatives for advertising experiments.

According to Google's chief economist, Varian, the Google advertising system ran more than 12,000 experiments in 2016 alone.

Toutiao has developed an advertising creation robot - Miaobi. By inputting industries and keywords, it can immediately create massive amounts of advertisements.

I tried it one day. I chose the gaming industry and typed in the keyword "two-dimensional". In an instant, a large number of texts were created.

Moreover, intelligent optimization bidding has made many advertising optimizers unemployed. As long as the advertising group determines its own conversion cost advertising, the platform will automatically bid intelligently and let the users who are most likely to convert see our ads. Today, Facebook, Google and Tencent are promoting this smart bidding method.

OK, let's summarize this section:

How to use automated tools to improve the efficiency of advertising. Archives: We should not fight against machines, we need to cooperate with machines. We need to make full use of existing tools to help us with advertising.

Question 4: Why does Liu Tao appear in the advertisements on WeChat Moments?

Let’s talk about the fourth question. Why did Liu Tao appear in the advertisements in the circle of friends ? The corresponding keyword is native advertising. By understanding the meaning of native advertising, we will know how to create popular advertising content in the Internet era.

You can see two pictures in the course - the left one is a cool picture, and the right one is a simple text link. Let’s guess which ad has a higher click-through rate?

The answer is the text link on the right.

So why does a simple text link have a higher click-through rate? It’s because that text link looks like a function rather than an advertisement, so its click-through rate is higher.

That’s the power of native advertising.

So what is native advertising? Let’s take a look at it through the following example:

Unlike traditional TV commercials, the fancy spoken ads in "The Debaters" are actually a kind of native advertising, such as KFC's "One person eats chicken, the whole family is honored"

So-called native advertising means making the advertisement and content look similar.

There are two types of native advertising: form-native and intent-native. The so-called form-native means that the advertisement looks the same as the content or function.

Let’s take a look at some different kinds of native ads - on Zhihu, an ad is a question; on TikTok, an ad is a vertical video; on Beauty Camera, an ad is an icon; on QQ Nearby, an ad is an avatar; on the mini-game Jump, an ad is a small box.

Next, let’s take a look at what intent-native is. This is an advertisement that Tmall launched on Douyin during the Double 11 Shopping Festival.

We can see that this advertisement does not have any celebrities, and the whole scene is very lifelike, just like the TikTok videos we watch every day. This is actually a native intent advertisement.

The so-called native intent means that the tone and intention of the advertisement are the same as those of ordinary content.

We see this picture in the course. This is a skin color detection advertisement customized by Meitu for UD. It can recommend foundation suitable for your skin color by taking a photo. It deeply combines advertisement and function, allowing users to experience the connotation of the brand in the function. This is also an intention-native advertisement.

Then you might ask, are there some advertisements that are both form-native and intent-native?

The answer is yes. Search ads are a typical example. First of all, the ads look very similar to search results, and then the ads are closely related to the search keywords.

OK, let's summarize this section:

How to create popular advertising content in the Internet era? The answer is one center and two basic points:

The one core is to make the advertisement look like something other than an advertisement. The two basic points are to make the form of the advertisement look consistent with the content, and to provide creative content that is as consistent with the tone and intention as possible.

Question 5: Why do I only see two or three ads on WeChat every day?

OK, let’s look at the fifth question. Why do we only see two advertisements on WeChat every day , but we can see many headlines on Weibo every day? The corresponding keyword is advertising frequency.

By understanding the impact of frequency on advertising, we will know how to reasonably control the frequency of advertising.

Let's start with a complaint. Momo's president Wang Li, whose nickname is Boss Wang, complained publicly on Weibo that he wanted to permanently block a direct-sending ad that was pushed to him every day.

So why is this ad exposed to the same person so frequently?

This involves the issue of the frequency of advertising, which profoundly affects the three-party balance of advertising , because advertising is a system that needs to consider the interests of three parties: advertising platforms, advertisers, and users.

• For the platform, frequency means that if the inventory frequency is strict, the inventory is small, and if the frequency is loose, the inventory is large;
• For advertisers, frequency will affect conversion rate and click-through rate. The click-through rate and conversion rate will definitely be different if the content is viewed once a day and five times a day.
• For users, frequency means the core user experience. If there are too many ads, users will leave.

Let's take a look at what factors affect the frequency of ads from the platform's perspective.

The first is the scenario. Different scenarios will affect the frequency of advertising. For example, as we just said, WeChat Moments has two posts a day, while Toutiao can have five pieces of content, just one advertisement.

Some people make a comparison and conclude that WeChat advertising inventory still has a lot of room for growth. This statement is actually inaccurate because Moments is a private space and the content we see is posted by our friends.

Toutiao is a public platform, and the content we see is unfamiliar, so it doesn’t matter if there are more advertisements. In this sense, it can never be as many advertisements as Toutiao.

The second factor that affects frequency is the attraction of the content. This is easy to understand. Youku’s popular TV series can post 120 seconds of advertisements, but we can’t post money and patches in front of Meipai.

The development stage of the product is also affected by the frequency of advertising. In the early stages of product development, fewer advertising may be experienced for users, and when commercialization is needed later, more advertising will be available.

User behavioral characteristics will also affect the frequency of advertisements. For example, there are some apps, Toutiao and Weibo, which will dynamically adjust the frequency of advertisements based on the user's activity.

Let’s look at a frequency case, which is probably a frequency strategy of Toutiao two to three years ago.

It includes three levels, the ad main level, the advertising creative level and the negative feedback level.

An advertiser refers to the same ad group, and the same ad plan only appears once within 12 hours.

The advertising creativity part is the same advertising creativity, the same advertisement landing page, the same application, and the same article only appear once within 12 hours.

Negative feedback means that after the user clicks negative feedback, the same account, the same application and the same landing page will no longer be displayed to the user within two weeks.

OK, let’s talk about how often we can talk from the perspective of advertisers that frequency means impression bonus for advertisers.

A question for many people is why household brands like Coca-Cola continue to advertise . Can we not drink Coca-Cola if we don’t advertise?

In fact, it is really possible. Let’s look at a natural case. The story of Jianlibao In the 1990s, Jianlibao was a household name for Chinese people. In 1997, it sold more than 5 billion yuan, which was more than the total of Coca-Cola and Pepsi.

Later, Jianlibao declined. Of course, there were many reasons for its decline. We cannot simply attribute it to not advertising, but it provided us with a natural case, that is, if we do not advertise, consumers will forget it.

Today, Jianlibao has not gone bankrupt. It is still sold in supermarkets, but because it has not advertised for many years, our consumers today have almost forgotten it. Behind this is the forgetting curve at work.

The forgetting curve of German psychologist Ebbinghaus tells us that any memory needs to be strengthened repeatedly before it is forgotten.

One indicator that brands are very concerned about when placing advertisements is coverage frequency.

Usually an advertisement will be played on TV for four weeks and then stopped for two weeks. When it is served on the Internet, it will cover the same user at least three times. The purpose is to make you repeatedly strengthen your impression before forgetting.

OK, let’s summarize this section, how to reasonably control the frequency of advertising.

For performance advertising, we need to constantly experiment and control the frequency of advertising to optimize the overall conversion rate.

For brand advertising, we need to follow the forgetting curve and reach the users' minds many times in a short period of time.

Question 6: What does the likes and comments below the ad mean?

Let’s look at the last question, what exactly does the likes and comments below the ad mean? Its corresponding keyword is called real-time interaction.

By explaining this issue, we know how to better interact with advertising audiences effectively in the Internet era.

Real-time interaction is an important difference between Internet advertising and traditional advertising—

Traditional newspapers and TV advertisements are actually unable to interact with each other. For consumers, they can express their attitude towards advertisements very clearly.

For merchants, interaction can collect specific feedback from audiences on advertisements in multiple dimensions.

Let’s take a look at a deep interactive case, negative feedback, this is the interface of Toutiao’s negative feedback. Clicking on a credit card advertisement for Everbright Bank will appear next to it.

First of all, it provides many negative feedback options, such as spam and blocking options.

Secondly, it provides accurate blocking options in multiple dimensions, including first-level classification and second-level classification in the advertising department. If we look closely, we will find that the three options of Everbright Bank, Credit Card, and Banking Finance are actually of hierarchy. In this way, it can know the audience's real advertising needs and advertising preferences in real time.

Let’s look at another deep interactive case, YouTube’s skip ad button, which is a very classic case.

YouTube's patch ads can be skipped after five seconds. This skip button allows advertisers to add a dimension to measure the quality of their ads. For example, whether they skip in the seventh second or the 23rd second, it will accurately reflect the ad's attraction.

This small skip button can be either negative feedback, such as skipping immediately, or positive feedback, such as insisting on reading. These metrics will optimize advertising and provide a very broad space.

Regarding this case, you can listen to the 626th issue of Logical Thinking. Teacher Luo Zhenyu made a program of Wei Xi’s article.

OK, let’s summarize this section together, how to better interact with advertising audiences effectively in the Internet era, the answer is two points:

First, analyze interaction, second, feedback interaction.

The so-called analytical interaction means making full use of various interactive indicators, adjusting advertisements in time, and increasing the conversion rate of the entire advertisement; the so-called feedback interaction means that if the audience interacts with you, you must immediately apply your advertisement to become lively.

OK, let’s review, we need to sort out the micro logic of Internet advertising through the six keywords represented by six questions. These six keywords are data, bidding, automation, native advertising, frequency control and interactivity.

There are many definitions of advertising. I will give the advertisement my own definition here based on my understanding, that is, to deliver the right advertising content to the right people efficiently in the right way at the right time.

The keywords we just talked about are exactly the same as this definition—

Frequency control allows us to seize the right time; through native advertising, we can create the right content; bidding is a suitable way; and automation can achieve high efficiency; data can help us find the right people;

We will find that there is actually one keyword missing, so I added a little to the definition, which is to repeat continuously based on feedback, corresponding to the keyword interactive.

Internet advertising is a comprehensive process. If we want to get high returns, we need to do every link above to the extreme.

Conclusion

At the macro level, Internet advertising has replaced traditional advertising and became the mainstream, and the basic logic behind this is the transfer of users' time.

At the micro level, Internet advertising has completely reshape the logic of the entire advertising operation from a technical and productive perspective.

Author: Wei Xi

Source: Weixizhibei