See how the through train operates the precise crowd
We all know the distribution mechanism of Taobao’s personalized traffic. Mobile Taobao search + mobile Taobao homepage is a typical example. We often operate through trains to drive natural traffic, especially the operating crowd. Now, if you don’t understand the operating population of the through train, it is basically impossible to do well. Crowds are directly affected by keywords, but the impact is different in standard and non-standard products. Today, let’s talk about the operation method of accurate crowd.
The function of being a crowd is tabs and pull tabs. When the crowd in your store is chaotic, you can use crowd optimization to correctly pull the baby’s tag. Especially when the baby is new, you can make good use of it and draw accurate labels to drive traffic. Let the baby get more accurate flow.
When we combine people, what should we do?
- Industry market
Business consultants can see data on the entire industry market.
- Visitor analysis
See storewide data. If the main traffic of the store is concentrated in one type, or the population of the store’s products is basically the same, you can refer to it at this time.
Baby customer group
Baby attributes
Combined with the baby’s actual situation, generally speaking, the baby’s audience can be analyzed, the scope can be appropriately enlarged, and the covered crowd can be appropriately released. If the coverage area is too small, it will affect the display and be restricted. Can be added to the population for testing.
The above four methods can be combined, generally in the following order:
Two combinations: age + gender
Three combinations: age+gender+single/age+gender+monthly average
(Depending on the category, some categories may have more attributes, such as clothing style, mother and child age, etc.)
Handling of Standards and Non-standards
High flow standard
method one:
New plan - add keyword - exact match card position - do not add crowd (you can also add, but the premium should not be too high)
To accumulate data volume, start optimizing keywords, delete and add. (The new product has a long operating cycle) - Add a group of people, and add a second- or third-level group as appropriate
Delete the observation crowd data, start the inquiry, lower the quotation, and increase the premium
When pricing, reduce the keyword data every day, but the amount should not be too large, generally 5%-10% rate, until the plan cannot run the daily limit or the click rate has dropped significantly. At this time, it is necessary to increase the crowd premium that has been raising, in order to quickly run past the daily limit. In this way, we will scale down the bid and repeat the operation so that the planned traffic reaches the crowd.
Method Two:
New Scheme - Add Keywords - Exact Match Card Position - Do Not Add Crowd
Accumulate data and start deleting keywords - add first-level population directly
The first-level population data is large enough, leaving good people, many people will be deleted, and then the bad data people will be subdivided into secondary groupings.
And the population remains the same, the secondary population remains the same, and if not good, the tertiary population will be subdivided and merged.
Generally speaking, the categories of standard products are sufficient for the second-level population, and the third-level population is too little covered.
method one:
New Solution - Average Bid - Exact Match - Crowd II/III Combination Test
Synchronized optimization of keywords and Crowd - the amount of accumulated data to delete and optimize because Crowd data is based on keys because the crowd premium of words also affects keywords.
Method Two:
New plan - average price bid - as accurate as possible - plus first-level population test
It is not easy to refine the test data of the population, delete the good population, subdivide the good population, and conduct the secondary population test
After the second-level population was tested, the adverse population was deleted, and the third-level population was further organized and merged. This is the limit, otherwise the population covered will be relatively small, we will not get so much data, and we will not achieve the desired effect.
GN finally said: the operation point of the crowd is to cover the crowd and cover the crowd repeatedly. If these two points are well controlled, it is not a big problem. The premise is that there is no problem with the keyword. Standard products leave good products and then combine with bad products. Non-standard and good merge deletes. The reason is the number of people covered. Standard product categories are generally smaller and non-standard products larger.