Amazon’s internal logic for capturing reviews

The full text has 1959 words and takes 10 minutes to read.

Making reviews on Amazon is one of the most effective promotion methods, but reviews also carry great risks. But where do the risks of reviews come from? What is the logic behind Amazon's review collection? What kind of review method is safe?

01 The logic of Amazon’s review collection

Amazon relies on big data + seller’s fake order records + buyer’s purchase history to catch fake orders.

First, Amazon can use the basic conversion rate and review rate (for example, the conversion rate is obviously abnormal) and other big data to catch some sellers who are obviously manipulating the rankings. For example, if SellerA is caught, Amazon will block SellerA's account (of course, it will not be blocked immediately if it is suspicious, and it will first be subject to advanced review), and then ask SellerA to write a complaint report, requiring a detailed list of each abnormal VP and review. This means that you have to confess and strive for leniency. (If SellerA does not confess, the store is basically finished. If he tells everything, the store will restore the sales authority, and of course all the problematic reviews and even products will be removed)

At this time, Amazon will get a new batch of data, which is BuyerA, BuyerB, and BuyerC provided by SellerA. ... Then, based on a comprehensive analysis of these Buyers’ VP and review history data, Amazon will follow the clues and catch SellerB and SellerC. ... Then SellerB and SellerC are banned from selling, and the above steps are repeated. ... ... This is a bit like interrogating a prisoner.

02 Where do the risks of assessment come from?

The risk of evaluation comes from these Buyers. For example, if the Buyer you find helps ten Sellers to do evaluations, and there is a problem with the data of one of them, you may be implicated.

03 What kind of evaluation method is the safest

There are two main evaluation methods now. One is to find foreigners or service providers on FB, and the other is to do it yourself.

After more than two years of Facebook experience, I have transformed from a newbie to a so-called "experienced user". My personal summary is as follows:

1. Many foreigners use their own small accounts to look for "free lunches", so when looking for reviews, you must ask the foreigners to send a profile, and then match it with the account that left the review. Many foreigners deliberately use their old accounts to seduce you, and then use their small accounts to purchase reviews to reduce the risk of their main accounts.

2. Many fake order intermediaries pretend to be foreigners to take orders without reviews and maintain their own accounts. If you see that the reviewer's Facebook page is relatively blank, most of these people are intermediaries or have small accounts pretending to be foreigners and asking you to fake orders.

3. Many of those who claim to be foreign students are also impersonating agents. Generally, they ask you to transfer money via WeChat or something. Most of these are agents because they may not have a foreign PayPal account.

4. Generally, the studios that help others do reviews make a living from this. They must have many buyer accounts and use VPS or multiple virtual machines. This is not fundamentally different from our own agency reviews.

Therefore, it is really rare to find high-quality reviewers on Facebook. I mainly work in the UK and US markets. Reviews from the UK are relatively easy to find, but reviews from the US are really scarce!

To sum it up in one sentence, it is difficult to find good and real buyers, and most of them take orders in large quantities. Referring to the logic of Amazon's interrogation of prisoners mentioned above, the risk is very high.

The second type is self-brush orders. Let me analyze several common evaluation environments for you:

1. VM/VBox +911

How to do it: Install VMware/Virtualbox virtual machine software, install Win10/Win7 image in the virtual machine, register a 911 account, install and run 911 in the virtual machine

Advantages: easy to operate, low cost. Disadvantages: When Amazon obtains virtual machine hardware information, there will be some special marks (for example, the hardware name of the VM will start with vm), which is easy to identify. 911 shared data center IP, if the same IP is evaluated by multiple sellers at the same time, it may trigger Amazon's advanced review. The result is self-evident.

2. Alibaba Cloud/Google Cloud

How to do it: Register an Alibaba Cloud/Google Cloud account, create a pay-as-you-go US cloud server, and use remote desktop for operations.

Advantages: Pay-as-you-go, low cost, and real data centers abroad Disadvantages: Pay-as-you-go IPs need to be changed frequently, which is not suitable for number maintenance. When searching for IPs, the IP number segment will be displayed as Alibaba Cloud or Google Cloud.

3. Foreign VPS partition operation

How to do it: Buy a large VPS server from some service providers, and then divide it into multiple small VPSs, and use virtual network cards to virtualize multiple IPs.

Advantages: Real computer room, fixed IP, can keep the account logged in for a long time, suitable for account maintenance Disadvantages: Use virtual network card for multiple IPs when partitioning, and use intranet penetration tools will be useless. The cost is high, and you need to buy a larger VPS server first.

4. iPhone+Luminati

How to operate: Use software to partition the iPhone, modify different local parameters, register Luminati to set a fixed exclusive residential IP, and use a disguised iPhone to connect to the residential IP for operation.

Advantages: Compared with PC and Android, iPhone is the best environment. Uses static fixed IP, suitable for account maintenance. Using camouflage software, a 64G iPhone can open 1000+ environments. Disadvantages: The operation is relatively complicated and not easy to use. Luminati IP is also relatively expensive $12.5/G

Without going into the conclusion, let's first talk about the conditions required for security assessment:

I captured the Amazon app (I had 5 years of IT experience before doing cross-border business), and found that the main information Amazon obtained from us included: device model, system version, IP, DNS, location, network operator, system cache, device name, UDID, iMEI, MAC address, IDFA, etc.

If we completely simulate these data, can we virtualize countless evaluators?

Combining all the above conditions, a perfect evaluation method is available:

Compared with PC and Android, iPhone is the best environment, no doubt about it (because Apple prohibits apps from obtaining basic hardware parameters, we can use tools to disguise all parameters obtained by Amazon). Using Luminati, we can connect to the home IP of the country that needs to be evaluated (IP address, DNS are all country-specific)

Fast speed, only need to log in for the first time, then you can directly enter without logging in. With the fixed IP of luminati, it is invincible to maintain the account. The total cost is low, a 64G iPhone can install 1000+ environments.

Statement: The article comes from the Internet and the copyright belongs to the original author. We respect the originality. Due to various reasons such as numerous reprints, we are unable to confirm the true original author of the article. If there are any copyright issues, please contact us for immediate processing.