Recommended Products I know

Product recommendations are divided into 2 kinds, one is to marketing for the purpose, one is really feel good. Today we only talk about the first kind.

Before we have done a, since think it is good, customer and customer service feedback or, unfortunately not http://www.aliyun.com/zixun/aggregation/18783.html "> official online, so far hold." The idea is simple, "the customer tells us how much he is willing to spend [X], what effect [Y] we want to get, and we help them choose the right path [P]." "Where x is relatively easy to obtain, Y has to be subdivided and abstracted into quantifiable features because of its diversity," he said. Then for each feature, select the path p1.p2 ... Pn, do combinatorial analysis, and finally, all the features of Y are put together for global optimization. Of course, the investment may be more than money, may be time or something, but then, it is too complicated.

We finally finished the first step of such a grand ideal, that is, when the customer says I want to take the first road, we tell him what shoes you should wear ... Say good details ... At that time we can see the path on the one or two, still full of mud, for fear of the customers are not willing to lead, or walk on the wrestling, the first road to the people do not give force, has been saying "or wait for us to fix the point, you bring people to?" "Eh, it's very tangled anyway."

The final conclusion is that the time has not come, and still need to wait.

Sad reminders!

These days again think of this matter, found that the idea of product recommendation is not always reliable, only in certain circumstances, can do meaningful.

First, the customer's appeal is too difficult to abstract and quantify. They used to say, "This is not a good product," it's not what I want. are very difficult to deal with the feedback, when you ask further or express the hope that they can be more detailed, they will think, answer two words "feeling" ah ... In short, there are not many customers who can clearly describe their needs. If you ask him again, do you pay attention to a or B? The answer all want the customer basically accounted for 90%.

Second, the client is not as stupid as you think, but most of them are lazy. A new product means a new set of content that requires new operations, and for most people there is no strong impulse to try if the present is enough, especially when it comes to paying. Therefore, in the new product before the launch of a small flow test, in order to recruit more voluntary customers, often take a free trial or at least half-price promotion. From the practical experience, it is good to recruit 8 or so big customers.

Third, do customers really need product recommendations? This week I also devoted to this topic to find a few familiar customer service, they also helped me to ask the customer. The answer is similar, customers do not need, customer service can see. Customers tend to be skeptical about what they recommend, and they know that your goal is marketing, and the subconscious mind is asking if you want me to pay. Think about the reaction you received when you got the phone call from the insurance company. Customer service, after all, have work pressure, their previous recommendation experience is mostly from personal accumulation, colleagues sharing and some help documents, few real-time reference data and information, and no similar tools to assist, so very welcome. Also said that customer service does not want such products open to customers, especially with detailed data and trend map, the main reason is afraid to explain the cost will increase.

There are good news, the good news is that most customers love to ask others how to do, competitors ah, the same industry are doing, see a good product will also be active inquiries. For example, the recent very hot Baidu open platform, the consultation is quite a lot. Unfortunately, when customer service asked me, I did not understand ...

This is the one below.

Speaking of which, I thought, the product recommend the different form will be better?

Last night to go to Thailand good taste dinner, each table on the menu on the list of last month's sales, the first is the winter soup, sold more than 500, the second is like a pineapple rice, a total of more than 10 dishes, have written a clear June sales figures. Before going to eat to check the public comments, to see what is the signature dishes, the restaurant of this act poured out of the first diners do not know what the trouble, looks so intimate, than the waiter said "We here XXX sell very well", a look at the price, heart filled with "expensive dead" aversion to the mood much better.

Can the product recommend this?

Now that the customer wants to know, we are properly informing others of what they are using, or what they have been doing recently to make the most of the herd mentality. If you can guarantee the authenticity of data and information, can establish a smooth feedback mechanism, so that users can clearly express the need for what not, with a certain amount of accumulation, and then start to recommend.

The usual scenario for recommending now is to face a client who has no avatar preferences and may even make a mistake about his gender. Generally, the basic information from his user ID and registration, and at most the performance of other product lines, can be inkling. We used the same type of customer data to make estimates, and we tested the estimates with real data, and the results were inaccurate. Think is also inevitable, with the past experience to predict the future is not necessarily accurate, with other people's past situation to see how possible.

I know the product recommended so much, summed up.

1. It would be much simpler if the customer clearly stated the input and the desired output and did not have any preference for the path, that is, to follow the results of your recommendation.

2. The product recommendation is based on a full understanding of the customer and the product. Both are changing, it will be difficult.

3. In fact, is the problem of cost-effective and optimal allocation, but a lot of data unknown, difficult to find the best solution. The solution has already been, this is not convenient to disclose.

4. Recommendation is a trend, and the time has not yet come. In the present, perhaps does the chart class the effect to be better. For the same data, customers will have different interpretations and choices. They will be responsible for their own decisions, but not necessarily willing to pay for your recommendations.

Source Address: http://www.wangxuntian.com/?p=1230