Seller Central
Seller Forums

Amazon Seller Forums » Selling on Amazon » General Selling Questions

Thread: Slow Sales, Search Words don't Matter, Suppression of listings by Algorithm
This question is answered.


Permlink Replies: 173 - Pages: 12 [ Previous | 1 2 3 4 5 6 7 | Next ] - Last Post: 08 Nov, 2017 9:04 AM by: Foot Supports I...
SusanH@SellerSu...

Posts: 4,580
Registered: 14 Mar, 14 8:50 PM
Posted on: 20 Jul, 2017 9:28 AM   in response to: Vase Candle in response to: Vase Candle  
Click to report abuse...   Click to reply to this thread Reply
Hello Sellers,

We understand your concern with the input of search keywords.

Amazon launched a feature that limits the length of the generic keywords attribute to less than 200 bytes in the IN marketplace, 500 bytes in the JP marketplace and 250 bytes in every other marketplace except the CN marketplace. The limits have shown to improve the quality of Search results. To date, the Search team has been suppressing generic keyword attributes that exceed these limits. This will apply to newly registered ASINs as well as existing ASINs.

We recommend following the guidelines below to optimize your use of the generic keywords attribute.

Key Guidelines for the Generic Keywords Attribute:

Keep the content within the prescribed length limit (less than 250, 200 for IN, 500 for JP):
  • The length limit applies to the total content in all generic keyword fields (A max of 5 attributes).
  • Number of bytes equals number of characters for alphanumeric characters (i.e. a–z, A–Z, 0–9) while other characters, such as umlauts (e.g. ä), each correspond to 2 bytes or more. Examples of other character byte sizes include £ (2 bytes), € (3 bytes) or ❤ (3 bytes).
  • Spaces and punctuation (“;” “,” “.”) do not contribute to the length limit, but words should still be separated by spaces. There is no need for punctuation, such as commas, between words.
Guidelines for optimizing keyword content for Search discoverability:
  • Do not include keywords that are not descriptive of the product.
  • Do not include brand names (even your own) or other product identifiers.
  • Do not duplicate content that is present in other attributes, such as title and bullet points.
  • There is no need to repeat keywords; once is enough.
  • We recommend using keywords that are synonyms, hypernyms, or spelling variations of content in visible attributes (e.g. if the product title contains ‘whiskey’, the generic keyword attribute may contain spelling variations (‘whisky’); if the product is a ‘guitar’, the generic keywords may contain a hypernym (‘musical instrument’).

For further information, please refer to the optimize listings for search and browse help page .

-Susan

I had a question sent to me in a Private Message, and I wanted to make a clarification here.

The prescribed length limit is less than 250 in ALL markets, except IN, JP and CN. Therefore, the prescribed length limit is less than 250 for the NA marketplaces, .ca, .com, .mx

Edited by: SusanH@SellerSupport on Jul 20, 2017 10:48 AM
Was this reply helpful? Yes 22   No 4
LiliC

Posts: 657
Registered: 17 Sep, 15 9:56 PM
Posted on: 23 Jul, 2017 9:15 PM   in response to: SusanH@SellerSu... in response to: SusanH@SellerSu...  
Click to report abuse...   Click to reply to this thread Reply
This would have been helpful had Amazon let us know the changes. In Handmade we were given 5 search term boxes each can have about 1K characters or "bytes" Not once in the handmade guide talked about limits..just relevance and no repeating.

Then we started reading about the 250 characters and a lot of us in the Handmade community reached out to Amazon and were given different responses. Some have no idea what we're talking about, others said yes, it changed to 250 characters, others told me to flat out that no, that these were all rumors and not do anything about it.

Summers are always slow but when you are 41% down from last year, that's really slow!

as of today 7/24, the Handmade community has yet to get an official email from Amazon about the whole 250 characters issues.

I know that some will say well "handmade" is not special, I don't think we are special but we do have differences from the Marketplace. We are all on the same team and we all want the same thing to be successful. If these 250 characters are the issue why our sales have taken a nose dive, it would have been great to hear about it from Amazon so we can change them.

Unlike the Marketplace, the handmade community does not have Bulk editing..those of us that have 1500+ listings have to edit our listings one by one and that takes a heck of a lot of time...it is a very painful task that requires us to get some sort of heads up so that we can start changing things...right now we are all just changing things based or rumors and hearsay and in business that's never a good thing.
Was this reply helpful? Yes 18   No 1
Moonwishes Sewi...

Posts: 1,399
Registered: 01 Aug, 12 7:10 PM
Posted on: 23 Jul, 2017 9:56 PM   in response to: Vase Candle in response to: Vase Candle  
Click to report abuse...   Click to reply to this thread Reply
Vase Candle - I don't know what the rules are for Handmade, but in 'regular' Amazon one the rules is that the primary, default photo be of the item that is for sale. I easily found your listings, but what I did notice is that many of your listings had photos that might represent the fragrance, but not a picture of the item itself. Such as the lovely roses, lilacs, even a sheet. None of these are vases or candles and are thus not permitted, unless as I said that rule doesn't apply to the Handmade category.
Was this reply helpful? Yes 9   No 4
Vase Candle

Posts: 33
Registered: 19 Sep, 15 1:04 PM
Posted on: 28 Jul, 2017 3:38 PM   in response to: SusanH@SellerSu... in response to: SusanH@SellerSu...  
Click to report abuse...   Click to reply to this thread Reply
Attachment Thumper.JPG (87.4 KB)
Attachment VaseCandle2.PNG (55.0 KB)
Attachment VaseCandle1.PNG (344.8 KB)
Thank you for your response Susan.

But we both know that the key words don't really matter in the search. It is the algorithm that makes the decision as to whether a customer will see an item or not.

Christmas 2016 - I have the screen shot of Vase Candle being listed as a business on the left. You click on it and it didn't show any listings even though I had over 200 at the time. That didn't have anything to do with key words. I reported it on a ticket and for the whole month of December and into January it stayed open because Amazon couldn't (wouldn't) stop the algorithm from intentionally blocking my product.

I work for Amazon too and I tried to find a solution from my position here. My husband and I thought it might be roaming brown-outs on the servers. We even did tests to try and prove it. But it wasn't an unintentional server problem - it was the algorithm doing what it was supposed to do - suppress my listings so that I would purchase more advertising.

You see what I believe is that (whether intentional or not) the algorithm isn't there to provide good service to the customer, its function is to work the Seller.

I believe that if you are caught in this algorithm, the basic pattern is that there are good sales for a while - say the first 8 to 12 months. Then, when you become comfortable and build your business out to meet the demand, your sales will drop 30% to 70% for no real reason.

So, to save your business, you pump more money into advertising to try to keep the sales coming in (with Seller Prime and FBA being pushed later). They just take you up the steps.

Christmas 2015
200 Vase Candle listings
All Handmade section
I used 1/2 the title characters and just the normal description
I had three key words - literally - Candle, scented candle, jar candle.
$30,000 gross
Fees around $4,500 with Advertising at around $1,000

Christmas 2016
400 to 600 listings
Both regular Amazon and Handmade
Optimized everything - title, description and key words
$18,000 gross
Fees around $2,700 with Advertising at around $2,500

If you consider Customer Sales as being a finite resource, even though Amazon lost about $300 in fees from us – this opened up the customer space to allow another seller to make up the $12,000 difference. The 15% of the other seller's sales generated an additional $1,800 in fees for Amazon. Amazon came out $1,500 ahead on that same $30,000 in sales by restricting our sales. By manipulating our sales downward, Amazon increased its profits.

I’m not a lawyer, but what this looks like is a Racket – “A racket is a service that is fraudulently offered to solve a problem, such as for a problem that does not exist, that will not be put into effect, or that would not otherwise exist if the racket did not exist. Conducting a racket is racketeering. Particularly, the potential problem may be caused by the same party that offers to solve it, although that fact may be concealed, with the specific intent to engender continual patronage for this party.” Wikipedia

I will include a picture of the International Head of Handmade's response confirming that it isn't a server problem, but the algorithm is working just fine. I will also include the screen shots where my product didn't show up.

So what a Seller can do is to check their account if their sales have gone down like this and compare what they paid in fees and advertising to last year. What you should see is that the fees remain close to the same while your sales have gone down drastically (40% to 50%). That's what the footprint looks like. It looks like Amazon created the problem and then sold the solution.
Was this reply helpful? Yes 22   No 4
MG85

Posts: 65
Registered: 15 Jan, 15 12:04 AM
Posted on: 28 Jul, 2017 3:49 PM   in response to: Vase Candle in response to: Vase Candle  
Click to report abuse...   Click to reply to this thread Reply
I'm currently experiencing something like this and believe that Amazon is suppressing at times, starting with this Tuesday I believe. My sales LITERALLY dropped about 80% overnight. 1 day it was fine and the next its like everything stopped...I dont know what to do really. I contacted seller support and they have a "specialist" working on it...

I kind of agree with what you described...thats easy money for Amazon and they would do that...
Was this reply helpful? Yes 19   No 1
CabinFeverBooks

Posts: 5,033
Registered: 23 Mar, 12 7:09 PM
Posted on: 28 Jul, 2017 3:56 PM   in response to: Vase Candle in response to: Vase Candle  
Click to report abuse...   Click to reply to this thread Reply
Oh, Now I see how "The Sprinkler" Works...

Thanks for the Heads Up....

Whats kind of amazing is all the effort being done to micro manage sales...

Why not just let buyers see everything and choose what they want and from whom...?

There must be a serious need to "spread the wealth"....to go though all this trouble.
Was this reply helpful? Yes 10   No 10
Skeeter

Posts: 929
Registered: 04 Mar, 12 3:20 AM
Posted on: 28 Jul, 2017 4:05 PM   in response to: CabinFeverBooks in response to: CabinFeverBooks  
Click to report abuse...   Click to reply to this thread Reply
CabinFeverBooks wrote:
Oh, Now I see how "The Sprinkler" Works...

Thanks for the Heads Up....

More than a sprinkler is going on, it does appear like you have pretty good assessment on that.

We started doing some research again and one thing discovered by accident is what appears to be a bubble. At this point just don't consider it for fact because more work needs to be done to confirm that.

Pricing seems to be another issue that is removing buy box eligibility.
Was this reply helpful? Yes 6   No 4
RKF Tools

Posts: 1
Registered: 04 Sep, 16 7:58 PM
Posted on: 31 Jul, 2017 4:14 PM   in response to: Vase Candle in response to: Vase Candle  
Click to report abuse...   Click to reply to this thread Reply
It seems that i have fallen into the same problem as the rest of you. I just started selling at the beginning of this year.

I had 1 product that was selling 100 units a month for 3 months, then in June it stopped overnight. june i sold 30 units. i was in the middle of a deal to buy out a brand and inventory, total of 3 products and about 1000 total units.

So in july, with these 3 new products now available, i have a total of 6 active products and a total of 40 sales for the month!

in july i really focused on optimizing my listing, i also tried optimizing my PPC, and nothing seemed to work.

I also learned with the help of other friends that are sellers, if keyword fields have more than 50 characters the entire list will not index! So its a total of 250 characters combined between all 5 lines. and it was updating in 15 mins or less. as soon as i had 51 characters non of the words indexed. i made it 50 or less and with in 15 i was indexing again.

i am so disappointed, i was enjoying learning how to sell on amz. then they made these changes, or put me in the "game" and now i just want out because of how much money it will cost me. I wanted this to be a big part of my income and allow me to quit my job and work for myself. unless things get better soon i will have to find another source to replace my job, and just sell out my stock and be done with amz.
Was this reply helpful? Yes 16   No 1
JustHaningOn

Posts: 46
Registered: 02 Mar, 12 5:12 AM
Posted on: 31 Jul, 2017 10:02 PM   in response to: Vase Candle in response to: Vase Candle  
Click to report abuse...   Click to reply to this thread Reply
This is the exact situation we are in currently. Sales are down 60% YoY and Amazon's portion of the pie is bigger YoY.

My question is, when you are caught in this downward spiral, how do you get out of it? Is it even possible to get out of it? At this rate, we are likely to do 60% of this year's lousy sales next year and that will be the end of our business.

So far, doesn't seem like anything we do(adding new SKU, lowering price, advertising etc) matters. No matter what we do we're stuck in 60% of last year's sales, and can't seem to get out of that pattern.
Was this reply helpful? Yes 11   No 0
AwesomeSeller

Posts: 650
Registered: 02 Jan, 15 6:48 AM
Posted on: 31 Jul, 2017 11:32 PM   in response to: Vase Candle in response to: Vase Candle  
Click to report abuse...   Click to reply to this thread Reply
Attachment Capture.JPG (55.6 KB)
Attachment Capture.JPG (71.2 KB)
Attachment Capture.JPG (49.3 KB)
Vase Candle wrote:
Thank you for your response Susan.

But we both know that the key words don't really matter in the search. It is the algorithm that makes the decision as to whether a customer will see an item or not.

Christmas 2016 - I have the screen shot of Vase Candle being listed as a business on the left. You click on it and it didn't show any listings even though I had over 200 at the time. That didn't have anything to do with key words. I reported it on a ticket and for the whole month of December and into January it stayed open because Amazon couldn't (wouldn't) stop the algorithm from intentionally blocking my product.

I work for Amazon too and I tried to find a solution from my position here. My husband and I thought it might be roaming brown-outs on the servers. We even did tests to try and prove it. But it wasn't an unintentional server problem - it was the algorithm doing what it was supposed to do - suppress my listings so that I would purchase more advertising.

You see what I believe is that (whether intentional or not) the algorithm isn't there to provide good service to the customer, its function is to work the Seller.

I believe that if you are caught in this algorithm, the basic pattern is that there are good sales for a while - say the first 8 to 12 months. Then, when you become comfortable and build your business out to meet the demand, your sales will drop 30% to 70% for no real reason.

So, to save your business, you pump more money into advertising to try to keep the sales coming in (with Seller Prime and FBA being pushed later). They just take you up the steps.

Christmas 2015
200 Vase Candle listings
All Handmade section
I used 1/2 the title characters and just the normal description
I had three key words - literally - Candle, scented candle, jar candle.
$30,000 gross
Fees around $4,500 with Advertising at around $1,000

Christmas 2016
400 to 600 listings
Both regular Amazon and Handmade
Optimized everything - title, description and key words
$18,000 gross
Fees around $2,700 with Advertising at around $2,500

If you consider Customer Sales as being a finite resource, even though Amazon lost about $300 in fees from us – this opened up the customer space to allow another seller to make up the $12,000 difference. The 15% of the other seller's sales generated an additional $1,800 in fees for Amazon. Amazon came out $1,500 ahead on that same $30,000 in sales by restricting our sales. By manipulating our sales downward, Amazon increased its profits.

I’m not a lawyer, but what this looks like is a Racket – “A racket is a service that is fraudulently offered to solve a problem, such as for a problem that does not exist, that will not be put into effect, or that would not otherwise exist if the racket did not exist. Conducting a racket is racketeering. Particularly, the potential problem may be caused by the same party that offers to solve it, although that fact may be concealed, with the specific intent to engender continual patronage for this party.” Wikipedia

I will include a picture of the International Head of Handmade's response confirming that it isn't a server problem, but the algorithm is working just fine. I will also include the screen shots where my product didn't show up.

So what a Seller can do is to check their account if their sales have gone down like this and compare what they paid in fees and advertising to last year. What you should see is that the fees remain close to the same while your sales have gone down drastically (40% to 50%). That's what the footprint looks like. It looks like Amazon created the problem and then sold the solution.

You seem to be suggesting that Amazon would rather collect the money from Ad sales versus fees from selling. That doesn't make sense. In a perfect world, for Amazon, Ad sales would drive sales to increase their take for each one of your ASINs. So while I understand your frustration. I'm not sure that I'm sold on the idea of Amazon "flipping" a switch to stop your sales.

What you'd need to prove, statistically, is that your sales did not follow the trend of total candle sales. If the total went down, it's very likely that your share took a proportional hit to sales. I can't say either way, but at this point, I'm not really sure you've got enough data to make a case either way. Although, I really do feel like Amazon should do more to explain how products are weighted.

My guess is that there may have been a change that impacted your ability to show up in the search results based on your search terms. Uniqueness in branding is why apples would be difficult to find if you were looking for apples. The ambiguity means that the algorithm has to get more information from either the user, their search history, purchase history, recent page history, product sales history etc.

When Apple Computer's brand was created, it was created prior to widespread consumer use of "indexed" semi-structured search. Although some users had spent time in library databases, many people were unaware how information was retrieved from SQL databases. SQL was created in the 1970s, but only SQL database folks were very familiar with schemas and how E.F. Codd and other created the framework. Now, fast forward to the internet and search engines. SQL was great for structured data that was designed for a specific use case. The internet created a new opportunity for unstructured data to be "indexed."

See, most people don't want to spend all day going through the "search results" looking for your particular brand. What the customer would like to do is find the product that they would like to buy. This gets a little trickier when anyone desires to get to the head of the pack. Say for instance, I wasn't a candle gal/guy and was the Headphone Jack brand of awesome headphones, or Black Backpack (as someone else mentioned). I could easily get to the top of search results beating well known headphone makers and headphone jack makers because both headphone and jack would give me great search results. In fact, I might be able to game the search algorithm and not really realize I had been given an unfair advantage. I've used "google" to illustrate the point. Of the 4 results in ads, 2 of them are not for headphone jacks at all. That means that someone has tilted the playing field because someone decided they'd pay Google ad dollars. Now imagine I was a headphone maker whose brand was Headphone Jack. I've just gamed the search results.

To be sure, I know that your $30,000 in candle sales were a lot of money. But, we're talking about a $3.2 Billion Dollar market. If I were a competitor, I'd be fuming that Amazon displayed your candles for as long as they did.

Just for fun. Here's Apple. Imagine the apple growers' dismay over the strangeness of what google expects you want information on. You think I could change it? I could probably train my history? But, so few people may be looking for Apples on google that even my search history, pages I'd visited today, my geoloc, the time of day, month of the year, phase of the moon, my average time on the site... won't change the results! lol

The reason I mentioned SQL (an RDBMS) databases was that a good number of people have a general understanding of how SELECT commands work, maybe not the actual command itself, but how the data is pulled from a database using venn diagrams. It becomes more complex when you're dealing with semantic queries.

So why are semantic queries needed? To solve the problem of people not wanting to look through the 100,000 different candle ASINs. Amazon has to create a method to achieve the impossible. And then explain it to people who don't trust them.

But, I guess if they moved you to the top of the search results, it may be the time to go nuts and brand two word combinations of similar products and add them to the brand registry. Ha!

Was this reply helpful? Yes 6   No 8
Vase Candle

Posts: 33
Registered: 19 Sep, 15 1:04 PM
Posted on: 03 Aug, 2017 3:42 AM   in response to: AwesomeSeller in response to: AwesomeSeller  
Click to report abuse...   Click to reply to this thread Reply
I'm going to try to put this simply for you.

Sales are stagnate at Amazon. According to the Wall Street Journal, sales did not increase for Amazon during the 4th quarter 2016, but the profits did increase. Have you noticed that Amazon is buying up brick and mortar stores?

I actually work at Amazon. I thought that the slower sales (not just for me, because we could see the slower sales in the facility) could have been a sign of roaming brown outs on the servers. I pressed really, really hard to try to have it looked at all the way up the executive ladder (and I do mean all the way up).

The answer that I got from the executives was that it wasn't the servers, what I was seeing was the algorithm.

The algorithm's purpose is to make the most money for Amazon.
Was this reply helpful? Yes 25   No 0
CabinFeverBooks

Posts: 5,033
Registered: 23 Mar, 12 7:09 PM
Posted on: 03 Aug, 2017 3:50 AM   in response to: Vase Candle in response to: Vase Candle  
Click to report abuse...   Click to reply to this thread Reply
Last 7 daze have been Slow to point of Stopping...

Prior couple weeks Great...even with my now limited inventory.

Its like "The Sprinkler" moved. Going to Suddenly Nothing makes it hard to plan or even to want to plan... I have other projects to work on if AZ is going to go dormant.

Whats the point if its all determine by some algo?

We all sort of know "Something" is going on besides "bad inventory" or "too high prices"...

Vase Candle now exposes what most of us already suspected...Its the Algos

Why is AZ "buying up" B&Ms? Why are sales declining "across the board" How does that boost profits...?

Sears went from a Catalog company to Stores and they're barely staying out of Bankruptcy...

They were the 1917 version of AZ?

Perhaps Customers just aren't Spending or On-Line Shopping has PEAKED?
Was this reply helpful? Yes 7   No 8
Skeeter

Posts: 929
Registered: 04 Mar, 12 3:20 AM
Posted on: 03 Aug, 2017 5:21 AM   in response to: CabinFeverBooks in response to: CabinFeverBooks  
Click to report abuse...   Click to reply to this thread Reply
CabinFeverBooks wrote:
Last 7 daze have been Slow to point of Stopping...

Its like "The Sprinkler" moved. Going to Suddenly Nothing makes it hard to plan or even to want to plan... I have other projects to work on if AZ is going to go dormant.


We all sort of know "Something" is going on besides "bad inventory" or "too high prices"...


Why is AZ "buying up" B&Ms? Why are sales declining "across the board" How does that boost profits...?


Perhaps Customers just aren't Spending or On-Line Shopping has PEAKED?

Since we are multi channel and have an un-manipulated organic for base line reference I have a pretty good source of data. Deriving market trends from the seller side data of Amazon is a useless because your sales are heavily manipulated. Historically, this site does not have such a steep drop during the summer slow season due in part to the dynamics of Amazon's scale. This season is way different, the Amazon channel has the largest percentage decline, about 32% greater decline than our other channels.

At this point I have pretty good suspicion what is going on, I will let you fill in the rest.
Was this reply helpful? Yes 4   No 3
TFC-Josh

Posts: 2,578
Registered: 05 Mar, 12 10:00 PM
Posted on: 03 Aug, 2017 5:55 AM   in response to: Skeeter in response to: Skeeter  
Click to report abuse...   Click to reply to this thread Reply
CFB is a conspiracy theorist who has long held the belief that his poor choices in inventory, refusal to reprice, and general unwillingness for many years to cull inventory was all Amazon's fault. The blame of an unknown and unspecified algorithm he calls the "sprinkler", which directs sales to specific merchants in specific bursts at specific times for non-specific reasons, is just the most recent iteration of his desire to not assume responsibility for bad business decisions.

Don't be suckered in by him.
Was this reply helpful? Yes 10   No 19
Skeeter

Posts: 929
Registered: 04 Mar, 12 3:20 AM
Posted on: 03 Aug, 2017 6:29 AM   in response to: TFC-Josh in response to: TFC-Josh  
Click to report abuse...   Click to reply to this thread Reply
TFC-Josh wrote:
CFB is a conspiracy theorist who has long held the belief that his poor choices in inventory, refusal to reprice, and general unwillingness for many years to cull inventory was all Amazon's fault. The blame of an unknown and unspecified algorithm he calls the "sprinkler", which directs sales to specific merchants in specific bursts at specific times for non-specific reasons, is just the most recent iteration of his desire to not assume responsibility for bad business decisions.

Don't be suckered in by him.

That might very well be true however CabinFeverBooks "conspiracy theory" shares parallelisms to some of our observed Amazon marketplace sales data. Attempting to discredit their comments by claiming "conspiracy theorist" at this point is not good, you personally do not now for fact how the platform operates so why would you attempt to discredit CabinFeverBooks observations?
Was this reply helpful? Yes 14   No 5