Hello ToyDemon Team,
I wanted to share some feedback regarding product visibility on your site. While browsing under the “Masturbators” category, I’ve noticed that not all in-stock items appear in the listings. Some products only show up if you already know they exist and search for them directly by name.
I’ve also noticed that discontinued or out-of-stock products can’t be found through browsing either, even though their listings still exist on your site. This makes it difficult to get a complete view of your catalog, both past and present, and limits how much information I can gather as a customer.
At the moment, the search toolbar and category dropdowns feel restrictive, and it takes extra effort to track down products that I know are available but not appearing in the normal menus. If this could be improved, it would greatly enhance the shopping experience and help customers make more informed purchasing decisions.
It would be especially helpful if the catalog view displayed all relevant products — whether in stock, out of stock, or discontinued — as long as the listing still exists. That way, customers can easily explore your full lineup without needing to already know what to search for.
Thank you for considering this improvement, as it would make browsing, shopping, and planning future purchases on your site much easier.
I wanted to share some feedback regarding product visibility on your site. While browsing under the “Masturbators” category, I’ve noticed that not all in-stock items appear in the listings. Some products only show up if you already know they exist and search for them directly by name.
I’ve also noticed that discontinued or out-of-stock products can’t be found through browsing either, even though their listings still exist on your site. This makes it difficult to get a complete view of your catalog, both past and present, and limits how much information I can gather as a customer.
At the moment, the search toolbar and category dropdowns feel restrictive, and it takes extra effort to track down products that I know are available but not appearing in the normal menus. If this could be improved, it would greatly enhance the shopping experience and help customers make more informed purchasing decisions.
It would be especially helpful if the catalog view displayed all relevant products — whether in stock, out of stock, or discontinued — as long as the listing still exists. That way, customers can easily explore your full lineup without needing to already know what to search for.
Thank you for considering this improvement, as it would make browsing, shopping, and planning future purchases on your site much easier.
**ToyDemon** ( 11202 )
@GeassComplex Which in-stock item are you not seeing in the masturbator listing? Please give us an example of an item you know is available but not appearing in the normal menus so that we can check into the matter.
Out of stock items do show up on the listing. We double checked and they are showing up on the listing and are filterable.
As for discontinued items, they do not show up on the listing by design. We have been around for a long time and the amount of discontinued items number in the many thousands. If we were to list them all, at least 5 out of every 6 item on the listing page would be a discontinued item.
We used to list them a long time ago but the amount got to a point where it got in the way of the browsing experience. However we've made many of the discontinued item pages still available if needed as you've mentioned. In general discontinued items' page are limited in their helpfulness as they aren't available to purchase anymore.
Out of stock items do show up on the listing. We double checked and they are showing up on the listing and are filterable.
As for discontinued items, they do not show up on the listing by design. We have been around for a long time and the amount of discontinued items number in the many thousands. If we were to list them all, at least 5 out of every 6 item on the listing page would be a discontinued item.
We used to list them a long time ago but the amount got to a point where it got in the way of the browsing experience. However we've made many of the discontinued item pages still available if needed as you've mentioned. In general discontinued items' page are limited in their helpfulness as they aren't available to purchase anymore.
ChonkBoyeee ( 20 )
@**ToyDemon** I think the site works pretty good. Not to hijack this post or dismiss the OPs concerns. But I’m gonna slightly hijack the post with a related but different comment on the search and discovery functionality.
Im not a “gotta collect em all” person but I am a platinum member. Honestly I’ve pretty much bought them randomly but there’s only so far I’m willing to take that. I’ve got some turds in the batch, or at least they just weren’t for me. Probably 1/2 of them I’m like “eh”. If there was a way to increase my confidence I’d enjoy the ones I got, I would spend more money, period.
Here is where you could improve: make it easier to know what an onahole is going to be before I buy it, and improved recommendation engine. Both of these can be solved trivially with AI, I believe. I’m a software engineer and this stuff is cheap and easy.
The first is the simplest. There’s a ton of information on the boxes. I’ve taken screenshots or pictures with my phone and had ChatGPT translate them. There is so much info that is just too inconvenient to do with a bunch of these for each user. Then I have to worry it’s going to show in my ChatGPT account I sometimes share, or I’ll save an image somewhere, it’s just a hassle. It would be trivial to do it all at once via the Open AI API and store that data along with the rest of the description and images, and once you do it for a product it’s done. The api is cheap and probably on the order of cents per product. For a massive value for users that ultimately gives them more confidence to purchase, and makes it more likely they’ll love it and buy again.
The 2nd part is more open ended. It’s more powerful and is even more value. This would be a recommendation engine that goes beyond “people who looked at this bought that”. This can be easy and cheap too, or it can be as complicated as you want to make it. I prefer simple. This example is conceptually simple but would likely provide shockingly good recommendations. It requires pre-work (translating the boxes) but that work is straightforward and a known amount of effort that is not particularly difficult for any one item.
First, assume you’ve got the boxes translated into text. You’ve used an LLM to summarize the attributes of each item into 50 words or less, just stuff like size, hardness, brand, size, whatever is on the box. To make a recommendation, you ask the user for 2-3 of their favorites. Pull the summary for those, and also the summaries for the top 500 onaholes. Write a prompt that says “a user prefers X, Y, and Z. Given this, pick the top 10 onaholes that this user is likely to enjoy. Here are the choices: <massive list of item name + summary>”.
There are infinite ways to improve this. Generate embeddings (so you can add your entire catalog rather than a subset). Add in reviews (although one benefit here is getting recommendations for unreviewed items). There are ways to make this cheaper than straight up text prompts. You can tailor this in many ways to prioritize certain products you want to sell while still making users happy. This is just an example of how simple it can be.
You probably don’t even need to put in the other product data to get halfway decent results but it’s going to work much better.
This is not something that would be easy to do for any particular user, nor is it really something that makes a ton of sense as a separate site. I signed up for some recommendation site but I’m just not going to spend my time entering purchase information into a separate page, then cross referencing recommendations with your site to see what works and is within budget. All the pieces are here and really, only here.
Me personally, I’m willing to try a bunch to find what I like. But if someone is buying just one for the first time, you want that one to be something they love. Then they’ll be back. If that second one is great too, and the third and forth, youll need to put up a hotline number to stop them from throwing cash at you. Thats a great place for any business to be.
I’m a software engineering contractor and can implement this. I’ve said what I want from the site so I can find more onaholes that I like personally. But if you’re interested I can totally make this happen. I’m not the cheapest by any means, but I am very capable and reliable, and have experience leveraging AI for this sort of thing. And I’d take payment in store credit which is probably a bit less common. :) reach out if interested. If not, find someone else to do it so I can use it! Thanks for your time.
Im not a “gotta collect em all” person but I am a platinum member. Honestly I’ve pretty much bought them randomly but there’s only so far I’m willing to take that. I’ve got some turds in the batch, or at least they just weren’t for me. Probably 1/2 of them I’m like “eh”. If there was a way to increase my confidence I’d enjoy the ones I got, I would spend more money, period.
Here is where you could improve: make it easier to know what an onahole is going to be before I buy it, and improved recommendation engine. Both of these can be solved trivially with AI, I believe. I’m a software engineer and this stuff is cheap and easy.
The first is the simplest. There’s a ton of information on the boxes. I’ve taken screenshots or pictures with my phone and had ChatGPT translate them. There is so much info that is just too inconvenient to do with a bunch of these for each user. Then I have to worry it’s going to show in my ChatGPT account I sometimes share, or I’ll save an image somewhere, it’s just a hassle. It would be trivial to do it all at once via the Open AI API and store that data along with the rest of the description and images, and once you do it for a product it’s done. The api is cheap and probably on the order of cents per product. For a massive value for users that ultimately gives them more confidence to purchase, and makes it more likely they’ll love it and buy again.
The 2nd part is more open ended. It’s more powerful and is even more value. This would be a recommendation engine that goes beyond “people who looked at this bought that”. This can be easy and cheap too, or it can be as complicated as you want to make it. I prefer simple. This example is conceptually simple but would likely provide shockingly good recommendations. It requires pre-work (translating the boxes) but that work is straightforward and a known amount of effort that is not particularly difficult for any one item.
First, assume you’ve got the boxes translated into text. You’ve used an LLM to summarize the attributes of each item into 50 words or less, just stuff like size, hardness, brand, size, whatever is on the box. To make a recommendation, you ask the user for 2-3 of their favorites. Pull the summary for those, and also the summaries for the top 500 onaholes. Write a prompt that says “a user prefers X, Y, and Z. Given this, pick the top 10 onaholes that this user is likely to enjoy. Here are the choices: <massive list of item name + summary>”.
There are infinite ways to improve this. Generate embeddings (so you can add your entire catalog rather than a subset). Add in reviews (although one benefit here is getting recommendations for unreviewed items). There are ways to make this cheaper than straight up text prompts. You can tailor this in many ways to prioritize certain products you want to sell while still making users happy. This is just an example of how simple it can be.
You probably don’t even need to put in the other product data to get halfway decent results but it’s going to work much better.
This is not something that would be easy to do for any particular user, nor is it really something that makes a ton of sense as a separate site. I signed up for some recommendation site but I’m just not going to spend my time entering purchase information into a separate page, then cross referencing recommendations with your site to see what works and is within budget. All the pieces are here and really, only here.
Me personally, I’m willing to try a bunch to find what I like. But if someone is buying just one for the first time, you want that one to be something they love. Then they’ll be back. If that second one is great too, and the third and forth, youll need to put up a hotline number to stop them from throwing cash at you. Thats a great place for any business to be.
I’m a software engineering contractor and can implement this. I’ve said what I want from the site so I can find more onaholes that I like personally. But if you’re interested I can totally make this happen. I’m not the cheapest by any means, but I am very capable and reliable, and have experience leveraging AI for this sort of thing. And I’d take payment in store credit which is probably a bit less common. :) reach out if interested. If not, find someone else to do it so I can use it! Thanks for your time.
**ToyDemon** ( 11202 )
@ChonkBoyeee Thank you for your suggestion.
Long before ChatGPT became a thing, we had for a short time (well, we've been around for a long time so comparatively short time) a chat bot on our website and a recommendation engine which were a lot of time and work back then to develop. The use data coming in during the time it was available was mostly people trying to test how far they could carry a conversation with it and vast majority of interaction wasn't anything that had to do with the website or items. Not many took or used the recommendation made.
Since then we've focus our tech development to the back end where improvements help in ways that isn't visible to visitors. LLM has exploded in recent years and we've thought about how some of our old system would work with recent developments in tech. However we are quite protective of our customer's data privacy and giving current third party access to those data even if done anonymously for recommendation isn't something we want to do.
As for building off the box information, we've touched on this a bit in other posts in the past asking about posting translation of the boxes. Japanese items aren't immune from having marketing fluff written on the box. Many users currently use their phone and something like Google Lens to translate the Japanese information written on them. Often people see tons of text and feel like they must be missing out on a massive amount of useful and technical information but once you translate a lot of it, it isn't as useful as they thought. Many technical data are already available in the items' Details tab.
The other thing is that there isn't a standard for attributes for onaholes so LLM reading the box information will likely lead to inaccurate data pool. What does a ToyHeart Baby Skin feel in softness compared to Ride Japan's Baby Touch, or Magic Eyes Skin, or Tomax's Very Soft? Every brand has their own custom blended material, name, and softness scale. Since everything is handmade, softness for the same item can vary even from production batch to production batch. An item's softness can even change as it sits on a shelf over time. We imagine a LLM would struggle with accuracy on common questions people always have like I want an item with the softest material or the most real feeling material.
This isn't to say that the current LLM couldn't make some of the recommendation on the website better but we feel that with the way things are, we aren't at that point yet. (and there's still the data privacy issue) Every so often, we've "kick the tires" on using updated statistical modeling in-house to create a new recommendation engine. However it's just not a project we want to take on at this time.
Thanks again for making your detailed suggestion. (We tried to keep our reply shortish.) We truly appreciate bringing the topic up as it's probably on many other's mind also with AI being a part of everything now.
Long before ChatGPT became a thing, we had for a short time (well, we've been around for a long time so comparatively short time) a chat bot on our website and a recommendation engine which were a lot of time and work back then to develop. The use data coming in during the time it was available was mostly people trying to test how far they could carry a conversation with it and vast majority of interaction wasn't anything that had to do with the website or items. Not many took or used the recommendation made.
Since then we've focus our tech development to the back end where improvements help in ways that isn't visible to visitors. LLM has exploded in recent years and we've thought about how some of our old system would work with recent developments in tech. However we are quite protective of our customer's data privacy and giving current third party access to those data even if done anonymously for recommendation isn't something we want to do.
As for building off the box information, we've touched on this a bit in other posts in the past asking about posting translation of the boxes. Japanese items aren't immune from having marketing fluff written on the box. Many users currently use their phone and something like Google Lens to translate the Japanese information written on them. Often people see tons of text and feel like they must be missing out on a massive amount of useful and technical information but once you translate a lot of it, it isn't as useful as they thought. Many technical data are already available in the items' Details tab.
The other thing is that there isn't a standard for attributes for onaholes so LLM reading the box information will likely lead to inaccurate data pool. What does a ToyHeart Baby Skin feel in softness compared to Ride Japan's Baby Touch, or Magic Eyes Skin, or Tomax's Very Soft? Every brand has their own custom blended material, name, and softness scale. Since everything is handmade, softness for the same item can vary even from production batch to production batch. An item's softness can even change as it sits on a shelf over time. We imagine a LLM would struggle with accuracy on common questions people always have like I want an item with the softest material or the most real feeling material.
This isn't to say that the current LLM couldn't make some of the recommendation on the website better but we feel that with the way things are, we aren't at that point yet. (and there's still the data privacy issue) Every so often, we've "kick the tires" on using updated statistical modeling in-house to create a new recommendation engine. However it's just not a project we want to take on at this time.
Thanks again for making your detailed suggestion. (We tried to keep our reply shortish.) We truly appreciate bringing the topic up as it's probably on many other's mind also with AI being a part of everything now.
GeassComplex ( 11 )
@**ToyDemon** I’ve been busy lately as to the late reply, but I did notice your website underwent some maintenance just recently. As to your request of me using examples, it was a bunch of items in some of my wishlists that I was going to use as your examples (as no matter how many times I used the varying categories under the masturbators drop down, I’d completely exhaust the listings and would notice a large assortment of items missing that I knew for a fact were in stock). But as of right now post maintenance, a bunch of the items I am making the claim to are once more viewable again which I’d like to clear up.
There is something else that came to my mind after making this post, but it’d be nice if you can develop a quick view of your products while scrolling through your listings (being able to scroll through the pictures without having to open the link would be a huge navigational help and make shopping much easier imo).
Another thing I noticed is some of the brands for products you have don’t have workable links to their (a) main brand page, so fixing that issue would help as well.
As to the *discontinued* as well as *out of stock* part of this topic, thank you for clarifying and clearing up the matter for me. If I notice anything else in the future to give helpful advice on in terms of improving the website experience, I’d gladly help again in the future.
There is something else that came to my mind after making this post, but it’d be nice if you can develop a quick view of your products while scrolling through your listings (being able to scroll through the pictures without having to open the link would be a huge navigational help and make shopping much easier imo).
Another thing I noticed is some of the brands for products you have don’t have workable links to their (a) main brand page, so fixing that issue would help as well.
As to the *discontinued* as well as *out of stock* part of this topic, thank you for clarifying and clearing up the matter for me. If I notice anything else in the future to give helpful advice on in terms of improving the website experience, I’d gladly help again in the future.