The biggest technical problem facing the advertisement (I.)

In the interactive advertising market, there are many technical problems. Some problems, people can change the industry habits, advertisers cooperation and other ways to solve. If we cannot solve the problems we are facing, we will not know how the advertising market evolves and how advertisers apply interactive advertising patterns.

One of the challenges: stock forecasting

At present, the biggest and most difficult technical problem facing advertisers ' advertising servers is inventory forecast. Each advertiser must make its own marketing campaign with visual impressions of the ads, and can book advertising sales. But because of the complexity of the algorithm, it is very difficult for advertisers to predict the future inventory.

Even predicting a user's access to a Web page is a big technical challenge. If we don't take into account the effects of disasters, storms, wars and scandals, we can get seasonal variable predictions. In addition, most advertisers sell stock through methods such as CPC and CPM.

Imagine the number of impressions on MSN and Yahoo sites on December 19, 2007. If one advertiser wants to buy a website to buy 10 impressions, if the only variable is the number of impressions, then the portal has a great opportunity to sell the ads.

But if the main purchase target impressions, how can the portal to convey its impression number? It is difficult to predict not only how many users have visited the site on the same day, but also how well they are located or the demographic structure they represent, or the Internet habits they have used over the last 60 days. If there are more changes, we can hardly solve these problems.

This is a big technical problem, if someone can solve the problem, then he will certainly get the math prize. In the industry, with strong technical strength of the company is not uncommon, this is plagued the industry for many years, but also customers often encounter problems. There are great technical limitations in predicting the evolution of the industry. The forecast stock is even more difficult for small businesses. This is a technical issue that has developed for many years and cannot be solved by an emerging company. Of course, startups can find and solve new problems, using an ad distribution system that solves the problem.

Problem Two: Target customers

People have been discussing whether the target technology can solve all the problems faced by the industry. Because advertising has nothing to do with itself, all people generally do not like advertising activities. Therefore, we should know more about the user's internet situation, send the target and related advertisement to them. In this way, everyone will like advertising.

At the same time, the target ads will also be cited in the stock forecasting problem, which is what I am trying to discuss in this column. This issue is of great value, so we need to continue to study it further.

First, let's look at behavioral targeting ads. You can create an anonymous identity for each visitor by technical means and remove all identifiable information from that user. You will then be able to track the activity of the anonymous user visiting various websites. Based on the content of the site, you can create an anonymous user profile, and the ads offered to them will be more valuable.

Because you have a large number of unique files, you can classify each file as an action segment. You can also assume that this behavior segment helps sell ads. The biggest problem with behavioral target advertising is that you need to know a lot of users ' information to realize their worth. There are many ways to ensure the privacy of a user, and there is no secret that large companies are doing such activities today. But if we want to really benefit from behavioral advertising, we have to learn more about anonymous users. The information collected on a website will not have much value and we need to understand cross site behavior activities.

In the past, we can only do this through client-side technology such as Claria (Gator) and WhenU. But the user and the industry objected to the system, so it eventually evolved into a new business model. Enterprises such as Tacoda have established a large number of anonymous user information databases through various large websites.

Other types of targeted advertising are different, including demographic targeting and geo-structured targeting advertising. Geo-structured target ads are implemented by the opposite IP address lookup. In general, advertisers have all IP address data for geographical location. If the place name is famous, its accuracy is higher. But some IP addresses are still only geographical.

The character structure target advertisement needs some means, for example, can realize the advertisement purpose by applying the well-known data. If we have geo-targeted data, we can apply for the demographic data of the famous area. Some advertisers also get demographic data through user registration data.

The biggest challenge facing the target ad is actually not implemented. The fact is, we don't have enough advertisers to advertise and make the vast majority of ads for individual users. In fact, only 20% of the users agree with the parameters that are widely told. Now, we can assemble advertisers through multiple websites and advertise to target customers to realize the value intention of advertisers.

(Original starting on December 18, 2006, compiled: Song)