Brief discussion of user recommendation

Attention to the ' user recommended ' opportunity is now a shopping site's ' Guess you like ', although the recommendation is to change the skin, but really a day began to accurately guessed my heart good. I guess I am a lot of operation, adhere to disdain to use the product, leaving enough clues to the database, which was hit by an arrow, let the program see through the preferences.

One day meeting, business and design in dispute ' recommend ' module content, in the end by the business push? Or do you want to push it according to your preferences? In other words, are you arranging a blind date or letting free free?

What do you mean by "user recommended"? Or am I looking at mine? Everyone purpose is different, as a user, rely on their own wisdom to make choices, far more than those colorful recommended ads to reliable, but the interesting result may be the same.

Today small Discussion user recommend

The recommended pattern that emerges from your mind:

Recommended in the online site:

A. Oversized advertising lightbox +flash scrolling

B. Sub-page two-sided ' recommended ' module

C. Page end ' recommend ' module

D. Recommendations for preset options when searching

E. Recommended search results

F. Nooks and corners Dogskin ' recommended ' Entrance

Outside site recommendation mode:

(Similar advertisement)

A. Email, regular mail

B. Instant Messaging

C. Dependent on other application software for ' recommended ' position

(It should be said that the advertising position is more accurate)

Recommended objects:

A. Current product users

B. Users who are not the user of this product, but may have an intersection with the software/information related to this product

C. No screening for all

Recommended frequency:

A. Recommended content for each page refresh update

B. Fixed limited area timing scrolling

C. (for advertising, targeted outside/offline recommended) According to product activities, special user events (such as birthdays, use cycle ...) Recommended at specific points in time

Recommended effect:

Full-spread advertising recommendations-specific users will be aware that the information is not the point of interest users will often ignore or hold suspicious attitude

Recommended for personal preference--the rate of acceptance of information increases in order of recommendation accuracy

(But we still will delete the so-called product new information mail, also never look in the mailbox of the member recommended ads, see unreliable book music recommended, all ignore Delete, finally manually find their own)

A lot of hot push products accounted for more than a lot of space, with a variety of PS crystal Bright pictures of the bombing eyeball, but the purpose of a clear and skilled users, incredibly little through the home page to buy products, a lot of pages to look at the turn off. In fact, the recommendation of this thing stand in the market point of view, nothing more than to sell products, especially unsalable possible products, so a variety of recommendations, advertising offensive, hope that users know, but the user ungrateful! Business needs chatter to be more violent, but with little or no progress. This turned out to be a double win transaction, the user is not an idiot, the occasional mentally retarded, but also from the emotional exclusion step by step follow, the sale of the fierce cover leg sales model. What users need to know is the information that really relates to me. Don't waste my time! Not to mention the money!

From this perspective, individuals are more likely to encourage targeted user referrals, and this user recommendation has long been done everywhere, but not so much. Some of the best name of the user recommendation is actually not accurate. This is a bit embarrassing, such a user recommendation is chicken ribs, neither open and hard to recommend the distance of advertising, and did not do considerate private customization services. That's doing the white ~!.

Find user interest points and behavior habits:

The reference is as follows:

Http://wenku.baidu.com/view/7711db64783e0912a2162a3d.html

1. User browsing document behavior habits

2. Users focus on the page information

3. The user has performed the save, the printing and so on the movement related information

4. User-flagged types of information

5. The user performs clicks jumps to a link the behavior

6. User access to a high degree of duplication of the corresponding functions or information

7. The function or information that the user operates many times

8. User browsing Time information

9. User-entered information manually (including search keywords)

Make recommendation Guide:

A. Best price recommendation for similar products

B. Recommendations of the latest products of the same kind

C. With the series of products recommended (such as the user bought shampoo, that is not the way to recommend a hair conditioner? ）

D. The same keyword recommendations for fuzzy categories (such as the user purchased a large number of red clothing, whether it can be ' red ' as a recommended keyword, corresponding to other non-clothing products)

E. Extension recommendations (such as user preferences suspense books, whether you can recommend similar movies or plays?) Even activities with logical analysis factors, related to the theme of the site of the travel of virtual people? Go to England to find Sherlock Holmes to Japan to find Conan and so on? Pull away ... ）

The benefits of a highly targeted user's personal recommendation:

--Increased site adhesion, users can quickly access to the use of high-value data, more intimate

--Improve effective feedback and product trading chances

Recommended mode in life case:

"Bookstore" will be the latest listing and the topic of the book in the first place, corresponding to different kinds of books, in the eyes and layer placed the most popular books, unpopular books are often at the highest level, entertainment fast sales books tiled on the flat table for easy browsing.

"Supermarket" famous beer and diaper case ~

There is often the most close to the shelf life of goods in the most important position, price promotion. On the contrary, there is very little advertising, and the location is not obvious. To the members of the implementation of the advertising brochure issued.

"Café" some coffee shop to provide a claim cup, establish products and customer contact, record customer purchase history, customers into the shop to do the corresponding product recommendations.

"Market" often go to the market aunt will know the neighborhood, know each other buying habits, the occurrence of time, taste preferences, budget range, will recommend the latest fresh vegetables, may not be the cheapest, but suitable for the neighborhood home cooking materials.

"China Pharmacy" This is probably the best recommended by the user, different with the supermarket shop aimlessly selling things, doctors record patient cases, the right remedy.

When it comes to the end, it's the right remedy ~!

So is there a shift in the long-term maintenance model when referrals are made and accepted for the user's perspective?

Will ~! When users are connected to a business platform, the user mindset is open, and users may try to accept new recommendations in addition to the inherent recommended product categories. Is this a good opportunity to insert a business purpose recommendation? possible! For example, recommended users 10 personal products are the best price, and long-term have good feedback, in the successful trading of multiple transactions, the system inserted 1 Although the price is not optimal, but the experience of the best of the same products, users will try to understand. Will you try to buy it many times later?

User behavior itself is also changing, the system data will be updated in real time, recommended mode needs to be adjusted, at the same time, a small number of mixed guidance for other purposes, non-user general mode, but with the user preferences have the intersection of products, whether it will succeed to achieve the goal?

Under the premise of building trust, we can discuss the other products effectively.

Or the example of a vegetable market aunt A, for example, you go to buy vegetables every day, aunt a know you love to buy vegetables, in short, your family and the youngest are like to eat vegetable, (this is what the assumption ...) Khan ... Then one day aunt a wholesale a lot of fresh eggs, aunt recommend to you to ensure that she sold eggs and her 10 years to sell you the same other vegetables, is the most fresh soil eggs, you do not have to go to another booth to see, by the way and vegetables together to buy, looking at aunt a sincere eyes, you are willing to do? Egg is not Ganoderma lucidum, not a few hairs, maybe you and vegetables together to buy a few noodles to eat ...

But if you go to a new booth and come up with an aunt B, you never bought any vegetables from her, she strongly recommended her soil eggs, really, the market selling eggs so much you can not see what is different, if the price and the same as the aunt a booth, you will return to aunt a there to buy, despite the warm recommendation of Aunt B.

As you often go to the old hotel, the boss will ask: "We have a new dish called #￥%......※, you want to try?" If it's not ridiculously expensive, usually you'll say yes. But to the new restaurant, this answer to be more cautious, ask the price, ask the material, ask later may also refuse.

Why are we always recommending? Is the final goal achieved?

Usually:

A. Not competitive, but the business returns a good product, will recommend the list

b． Competitiveness is not high, even under competitive products, need ammunition, find an excuse to recommend

In short, businesses do not lose money, to sell goods ~!

In fact, even if the competition is not high products, there may be some of the user needs of a fitting, if the beginning can find a relative to the user base to recommend, success rate on the big increase.

The same is competitive under the product, even if the advertising, recommended to do a wonderful, but wide net, there may be thin income, because did not find to the user launched the offensive. (Users are not stupid, know the recommendation must have a merchant price appeal)

Instead, the mind spent on each product corresponding to different users, mining products and user intersection point, such a recommendation is effective, targeted, even if the product price is expensive, but the experience is good, still have customer base. Finally, it is likely to correspond to the actual needs of the user, or it may be the same as the user's search for the product. So everyone's happy? You sell what you want to sell, and the user buys what he wants to buy.

Source Address: http://ued.ctrip.com/blog/?p=2717