For the sake of Jeff, Google’s star engineer, Xi Xiaohu is willing to come in person to meet the star newcomer of Google Brain.
But this does not mean that he will directly hand over the core data of the advertising system to Meng Fanqi.
Xi Xiaohu believed that he had the power and the need to check whether the young man in front of him was qualified and qualified to do this.
After all, if the algorithm is really changed, then the platform that Xi Xiaohu has been responsible for will eventually be launched online, AdWords.
If something goes wrong, he will be held responsible.
Google's AdWords is an online advertising platform. To put it bluntly, it is Google's version of PPC.
Advertisers who want to place ads can bid on this platform and display their services or products.
Whether it is the display of Google searches or the display of advertisements in related apps and videos, orders can be placed on this platform.
The service fee is paid according to the number of times the user clicks on the advertisement. If the user does not click on the advertisement, there is no charge. After the mobile phone version became available to almost everyone, it became basically the standard for advertising.
But at the beginning of the 21st century, it was still a relatively advanced design.
Don’t think that this is a conscientious move by service providers like Google. This is actually forcing advertisers to improve their advertising quality.
Some ads are of low quality, and not even one out of a hundred people will click on them; some ads are of higher quality, and will be clicked on ten times if they are shown a hundred times.
Under such circumstances, advertisers with poor ad quality must bid more than ten times the price per click to obtain the right to place ads.
From this perspective, advertisers themselves will also largely influence the price of services.
The pay-per-click method is generally more popular with advertisers than hanging banners directly.
Because the number of clicks is a kind of interaction between users and advertisers, it can be used to judge the user's attention and interest and make certain targeted adjustments.
Banner ads are placed directly in every corner of the website. Not only do they have no interactive effect, they are also easily annoying. It is a very poor way of advertising.
In terms of user experience, Google's strategy is very cautious.
The strategy they adopt is to strictly control the overall delivery quantity, use bidding methods to force advertisers to improve the quality of their ads, and then use their own big data to push customized advertising content to each user.
Under the Google advertising system, the average click-through rate of users is almost 2-3 times that of the second advertiser, which is quite an amazing result.
According to Meng Fanqi's understanding, about 60% of users simply did not notice that many of the links and interfaces they saw were advertising content.
Among the remaining 40%, nearly 10% of people knew it was an advertisement and were willing to click on it.
Xi Xiaohu, who led the team to achieve such results, is indeed qualified to be confident.
"Meng Meng is a young and promising person, and his breakthroughs in computer vision are obvious to all. However, there is a big gap between the advertising recommendation algorithm and the image algorithm. In particular, Google has very little external information about this part of the content. I wonder how much Meng knows about it?"
“As far as I know, there are roughly two types: search and non-search. The search algorithm will trigger ads to be placed on relevant pages and apps based on specific keywords, while the non-search algorithm will trigger ads when the user has not made a search. When taking action, we will specify specific and personalized recommendation content based on the information we have already mastered.”
Meng Fanqi gets angry when he talks about non-search ads. He will push whatever he wants after buying something. He doesn’t know who made this crappy algorithm. Who wants more after already buying it?
"Search recommendations are more important here because they involve bidding for related keywords. As far as I know, Google's ad placement does not just depend on bidding. It also evaluates the quality of the ad itself, the information attached to the ad, and the user Make a comprehensive judgment based on your own search environment.”
Meng Fanqi also couldn't understand why Baidu had used such a primitive method of bidding ranking for so many years and still sold Tieba, making itself notorious.
How much does it cost to sell Tieba?
Google combines the content quality of the advertisement itself with the subject and nature of the searcher's search, the geographical location and time of the search, the device used, the placement of other peer advertisements, and many other factors.
Each search will yield customized results. If there are 10 users with different search habits from different countries searching for the same keyword or product at the same time, then the advertising rankings given by Google will be different for these ten search users. identical.
For advertisers, even if your competitors bid higher than you, you can still achieve lower prices by using highly relevant keywords and high-quality advertising content, or by targeting regions and habits. Price wins higher rankings.
Xi Xiaohu asked a few more general questions, such as the panda algorithm and penguin algorithm currently used by Google, and Meng Fanqi answered them fluently.
"The Panda algorithm mainly punishes content farms and keyword stuffing very severely, and improves search quality by excluding web pages with low-quality content." The two are somewhat similar, and can basically be understood as keyword stitching monsters.
This kind of web page basically does not have any meaningful content, and it does not exist for users to see.
Its only purpose is to fill up all kinds of hot words on the Internet to trick you into clicking in. It is a pure garbage dump.
Whether search engines will find such inferior content will have a great impact on the user experience.
This thing is actually more similar to the product names of later small commodities, each with a dozen keywords strung together.
"The Penguin algorithm mainly punishes the link profile of the web page. If there are suspicious links from unknown sources on the website, it will greatly affect the website score."
Xi Xiaohu got a little carried away and asked a few more methods that had just been put into use in 2013, such as the Hummingbird Algorithm and the Pigeon Algorithm. Meng Fanqi saw the tricks. Except for some small details that he couldn't remember clearly, overall he gave very good answers. correct.
“The Hummingbird algorithm enhances Google’s understanding of user search text, while the Pigeon algorithm is designed to improve the ranking parameters of physical distance.”
The ranking parameter of physical distance, simply put, means that stores and search results that are close to you in reality will be more likely to appear on your search interface.
A very reasonable operation, it is always possible for people to go to a place closer to the actual location.
"Emmm." Xi Xiaohu failed to get what he wanted. After going back and forth several times, he found that Meng Fanqi could answer these questions very naturally. He was a little suspicious in his heart. This kid seemed to be prepared.
Could he really make any decent breakthrough? It shouldn't be.
Xi Xiaohu also spent a few minutes checking Meng Fanqi's resume.
I studied a completely unrelated major and researched a completely unrelated field.
Xi Xiaohu has seen many geniuses, but there are not many who can suddenly switch to other fields and shine like this. Most of them continue to make breakthroughs in a certain field.
Is it true that people are born with knowledge?
