How to calculate Amazon's "average category click-through rate and conversion rate"? A brief analysis of the meaning behind the data of "Amazon Product Selection Compass"

Question: Observe that the ratios of Click Counts/ Glance Views before October were close to 1, and only after October did the ratios become smaller. Was it search traffic before October (unlikely)? Or did Amazon adjust the data statistics in October? Or did I misunderstand these two data?

3. Search to purchase ratio (search to purchase ratio) as shown below

Search-to-purchase ratio, do you not understand this data? It is mentioned in "Misunderstanding of Conversion Rate". Here is another meaning of this data: it can estimate the search conversion rate.

The search-to-purchase ratio is the number of ASINs purchased/the number of ASINs displayed. The search-to-purchase ratio is 2.35‰, which means that if you buy 2.35 ASINs, you need to display 1,000 ASINs.

If you combine the data from Search Query Performance under Brand Analysis, you will find that in fact, about 24 ASINs are displayed for each search (this is my experience, your observation may be different) .

Assuming that 24 ASINs are displayed in one search, then to buy 2.35 ASINs, you need to search 1000/24=41.67 times, that is, to buy 1 ASIN, you need 41.67/2.35=17.73 searches (number of searches). It is estimated that the product search conversion rate (number of search orders/number of clicks) is greater than 1/17.73, or greater than 5.6%.

Why is it greater? Because by observing the search query volume and total click count under Search Query Performance, you will find that the number of searches is greater than the number of clicks, which means that some searches may not have clicks (why not? Maybe the search results are not what the customer wants). [If there is any error in the explanation here, please correct me]

The accuracy of this estimated search conversion rate (Search Performance has already mentioned one conversion rate calculation, this is another) depends on the accuracy of the rule of thumb (displaying about 24 ASINs). There is a difference between these two conversion rates.

Amazon did not provide such important data as Conversion Rate per Price Range. Return Ratio and Reasons for Returns. The former can be used in profit calculation, and the latter can help improve products.


4: Offer & Selection Metrics (offer and selection indicators) are as follows:

12 Months:
90 days:

The two charts above are data charts for December and 90 days. Number of sellers The number of sellers is the number of seller accounts. Amazon does not give a specific number, and it is not very meaningful to give a specific number.

Number of new brands The number of new brands can reflect the number of sellers entering the market during the selected period. There is at least one seller account for each brand.

Number of ASINs: The number of sub-listings. I personally think that this value includes ASINs that are no longer on sale but still exist on the detail page without changing to dogs. Because I feel that this value is too large, those that have changed to dogs should not be included.

You can choose 7 days, 30 days, and 90 days as references, and calculate the number of ASINs deleted within 1 month and 3 months (in the screenshot above, the number of ASINs in 7 days is 556.9K, 30 days is 593.1K, and 90 days is 652.8K. The difference is 36.2K and 59.7K, which means that the number of ASINs reduced increased by 36.2K/1 which is greater than 59.7/2, or compared with the number of newly added ASINs in 3 months, 32.4K). Is it getting bigger or smaller?

If it gets bigger, it may be that the number of people who choose to leave is greater than the number of people who enter, profits are decreasing, and competition is intensifying. Most nodes should be in this situation where the number of ASINs decreases and increases (Amazon is eliminating sellers), not to mention that there is more than one product in a node, so calculating this is actually not very meaningful for selecting new products. [The above analysis may be wrong, welcome to communicate]

Number of new ASINs The number of new ASINs is the number of new sub-listings in the selected time period. (The value of 7 days, 30 days, and 90 days has not changed. So the time of this new ASIN may be 90 days.) [The above analysis may be wrong, welcome to communicate]

Offers per ASIN The number of sellers for each ASIN is not very meaningful for reference. This is the average value. There are no sellers for discontinued ASINs, and some ASINs have multiple follow-up offers. The normal value is 1. If it is 2, it means that the category has a serious follow-up sale.

Star Ratings The distribution of ASIN quantity corresponding to the star rating. Normally, 4 stars and above account for the largest proportion. Generally speaking, in categories with high return rates, the proportion of 4 stars becomes smaller, but it is still higher than other star ratings.

To make this ratio more specific, you can press F12 in the browser to view the source code, find the pixel length of the horizontal bar, and do the calculation (in the screenshot above, 4+ pixels is 276, 3 is 55.1, 2 is 8.1, 1 is 4.1, and the calculated ratios are: 80.4%, 16.1%, 2.4%, 1.2%)
Advertisement Spend. The average daily spend is given below as 0.1 and the amount most people spend is 0.3.

The average of 0.1 may seem small, but there are 527.2K ASINs in 12 months, and the average daily expenditure of each ASIN is 0.1, which means the average daily advertising revenue of Amazon is 52720. Assuming that your main competitor spends 100 per day, that is 527.2 potential competitors (of course, the number of potential competitors cannot be calculated in this way).
 

 
02. Comparison of product selection compass data and third-party tool data

Although the data provided by Amazon has limitations, I trust Amazon's official data more than the data provided by third-party tools (this does not mean not to use third-party tools), because I think the data of third-party tools either comes from Amazon or is estimated through its own data model. Below, I will compare the data of Product Selection Compass with that of third-party tools.

The following two pictures show the sales data of the product selection compass and third-party tools at the same time point in January on the UK site:

The sales data are still relatively close (the total number of products has a big difference), after all, one is the sales of the top 100, and the other is the sales of the entire category.

As for the top sales, concentration index and other data provided by third parties, if you have the patience to use the monthly sales given by Amazon (as shown below), you can also calculate them, but it takes more time (that’s why third-party tools are charged, after all, time is money)

As follows: In January, the sales volume of the category was 183,130, and the total sales volume of BSR1 sellers in the last 30 days was 5,500+, and this BSR1 sales volume accounted for 3%+. At the same time, let's compare the sales volume of **Jingling, which accounted for 7.78% (as shown below)



Regarding data accuracy, here is a simple example: for example, when calculating sales, the data in **Elf, especially the multi-variant ones, are estimated using the information of a single variant, while Amazon's data are definitely statistics of the real data of each variant, so there is still a big error here.

Although the data of product selection tools may not be that accurate, they still play a big role in product selection and they cannot replace each other.

03. Last

When I started this article, I wanted to write everything in detail, as shown below (backstage screenshot, English and Chinese): The following picture shows the advantages of the mall product guide given on the Amazon help page. I think that except for the official data being accurate, there are no other advantages~~~///(^v^)\\\~~~

For example, the path to enter the tool is as follows:
To enter this tool, you can go through the "New Product Recommendations" card on the backend homepage, or you can go through the navigation menu-Growth-Product Selection Compass (under this growth menu, there is also a product selection tool-Business Opportunity Detector. The difference between these two tools can be considered that one provides data for categories, and the other provides data for products, brands and keywords).

After entering, there are 3 tabs, one is the function introduction page of Marketplace Product Guidance, one is Global demand for your products, and the other is Category Insights.

Later I thought about it, this is not my summary, but more like a tutorial for newbies. It is long-winded and useless. It omits a lot of content and goes directly to what I think may be more important.

"Wonderful Reply"

Hey, no big deal - Amazon Operations

Agree with: wkming, Ponyo is here

Thanks for sharing, it is very useful. I would like to ask a few questions:

1. I searched for the keywords in the query column of the table with the corresponding classification node ID, but I couldn't find the specific and accurate sub-category, or I couldn't find the corresponding one. I feel that Amazon's search accuracy is not very good. Or is there something wrong with my search? Please help me, experts

2. For purchase searches, about 24 ASINs are displayed for each search. Does this mean the number of ASINs displayed when searching for a certain keyword on the front end?

To summarize:

1. Search Performance = Clicks Count / Search Volume (the difference from advertising: advertising is clicks / impressions), which can be used as a reference for the click-through rate of keyword advertising search positions;

2. Estimated conversion rate: =Units Sold/ Click Counts and =Units Sold/ Glance Views. The difference between Click Counts and Glance Views: 1 is the average conversion rate of all search terms at this node, and 1 is the average conversion rate of the entire category of products. Click Counts is the number of clicks (one person clicks multiple times and counts once), Glance Views is the number of page views (one person clicks multiple times and counts multiple times)

3. Search to purchase ratio = number of ASINs purchased / number of ASINs displayed. The search to purchase ratio is 2.35%. That is, to purchase 2.35 ASINs, 1,000 ASINs need to be displayed. It can be used to estimate conversion rate.

4. By observing the changes in the number of ASINs and brands over 7 days/30 days/90 days, you can infer whether there are more entrants or exits in this category, whether profits are declining, and whether competition is intensifying.

Forget • Shenzhen • 2024-03-04 14:55

1. Not all nodes have it. You will not find the corresponding one by searching with keywords. Search layer by layer according to Category>Product Type>Item Type Keyword 2. Yes, the number of ASINs displayed by searching a certain keyword on the front end

kakitang

The data of the product selection compass has a certain lag, but the data is aggregated by Amazon through its own platform big data. It is a periodic data, but it is very real. For example, it is of great reference value when estimating product costs. For plug-in selection, the data is relatively real-time, but the plug-in selection is based on data captured by crawlers. If the data is not captured, there may be some impact. Secondly, it may be estimated based on algorithms. Generally speaking, it is not as accurate as the data from Amazon's backend.

As long as the category you want to sell is not monopolized by the top sellers, you can enter the market. However, I don’t know if the data in the product selection compass includes the data of Amazon’s own sales. After all, self-operation also accounts for a large proportion.

Anonymous user

Agree with: komoribe, Ponyo is here

If you combine the data from Search Query Performance under Brand Analysis, you will find that about 24 ASINs are displayed for each search (this is my experience, your observation may be different)-------about 24 ASINs are displayed for each search, how is this data obtained?

Weird Uncle Born After 1985 • United States • 2024-03-08 14:40

Search term performance report, find the relationship between search volume and display volume to know, 1 search volume is generally 24 display volumes, 4% is the common probability between the two, that is, 1/24, I wrote a related article about the calculation of this relationship 2 weeks ago: https://feishu.cn/docx/H5OHdxLEmorRTaxhXQWcCn7mnBg , you can take a look if you are interested, welcome to discuss. But to be honest, the category of the product selection compass is BSR, and many of the products here are not purified, so the above indicators can only be used as a reference. For products in the niche market, they are unpurified products and are not very accurate.

Anonymous user

The search click rate section says: This value is used as a reference for the click rate of keywords in the ad search position. I feel that the value calculated in this way is too high. After all, it is calculated at the category level, but we are making a certain specification product in this category.

Forget • Shenzhen • 2024-03-05 17:42

When you segment a product, the difference from the node average value may be large. This value is just a reference. From this value, you can get an expected fluctuation range. There is actually no accurate value for product conversion rate. There are different conversion rates in different time periods, different keywords correspond to different conversion rates, and different listing conditions may also have different conversion rates. The conversion rates we usually use are all average values.