InvisibleHand shows a discreet notification when the product you're browsing can be bought for a lower price elsewhere. It gives you a link directly to the product page at the competing retailer.
InvisibleHand
InvisibleHand
InvisibleHand
- Thread starter Osiris
- Start date
- Status
kmote
Seg Fault'n,
- Messages
- 5,801
How is that even possible?
invisiblehand
Beta member
- Messages
- 4
Hi All,
I'm one of the people behind InvisibleHand, just noticed that you're discussing this product
@kmote: It's possible, in fact. We use our in-house technology to match products from different websites together and though in some cases it makes mistakes, overall it's capable of finding the same product across different web-sites.
@Mak213: No, we don't use PriceChecker or PriceGrabber or any other database. InvisibleHand was developed from scratch without reliance on external databases. In fact, it would be a bad idea to use other databases as the prices there are very often outdated (more here: http://getinvisiblehand.com/wp-content/themes/invisible/images/pricegrabber-invisiblehand.jpg), while IH updates its prices every single time you use it. In other words, before it shows you a price, it actually goes to another site and scrapes it.
I'm one of the people behind InvisibleHand, just noticed that you're discussing this product
@kmote: It's possible, in fact. We use our in-house technology to match products from different websites together and though in some cases it makes mistakes, overall it's capable of finding the same product across different web-sites.
@Mak213: No, we don't use PriceChecker or PriceGrabber or any other database. InvisibleHand was developed from scratch without reliance on external databases. In fact, it would be a bad idea to use other databases as the prices there are very often outdated (more here: http://getinvisiblehand.com/wp-content/themes/invisible/images/pricegrabber-invisiblehand.jpg), while IH updates its prices every single time you use it. In other words, before it shows you a price, it actually goes to another site and scrapes it.
Ah, since you posted here; InvisibleHand looks great, does it work in the UK too? i.e. will it search on UK sites. Since your screenshots are in USD.
Edit: Ah, after a quick visit to the site I see that the answer was in plain view, thank you. I'll be donating to you, depending on if it saves me money in future
Also, I gave you rep for you obvious dedication to InvisibleHand, coming to our lowly forum to correct a few facts
Edit: Ah, after a quick visit to the site I see that the answer was in plain view, thank you. I'll be donating to you, depending on if it saves me money in future
Also, I gave you rep for you obvious dedication to InvisibleHand, coming to our lowly forum to correct a few facts
invisiblehand
Beta member
- Messages
- 4
@Yamikotai: I hope it works for you and saves you money IH database is currently biased towards books and electronics since they are much easier to match to each other than, say, clothes, but we're aiming to cover all product categories and all major retailers. Let me know if you've got any questions about IH, I'll be happy to help.
IH database is currently biased towards books and electronics since they are much easier to match to each other than, say, clothes, but we're aiming to cover all product categories and all major retailers. Let me know if you've got any questions about IH, I'll be happy to help.kmote
Seg Fault'n,
- Messages
- 5,801
Hi invisiblehand, a couple of questions if I may. When you say it goes to another site and scrapes it do you mean that the application running on my PC goes to all the sites on an internal list of some kind and searches for cheaper matching products or does it look the product up on your database and return cheaper matches, allowing your servers to do the scraping?
invisiblehand
Beta member
- Messages
- 4
@kmote. IH operates as follows:
1. When you start Firefox, IH downloads the list of retailers it supports: amazon, currys, bestbuy etc. The reason for this is that you don't have to update the extension when we add a new retailer to our database and it happens fairly often.
2. When you open a new tab in firefox, it checks whether it's a retailer. So if the address begins with, say, 'http://www.amazon.co.uk/...', it proceeds to the next step, otherwise (for example f you open 'http://www.google.com/) it just stays inactive.
3. When it determines that you're at one of the supported retailer, it sends a request to the server, asking for alternative urls. Say, IH sends an amazon url to the server and the server replies that this product is available at currys and johnlewis as well.
4. Then IH opens these urls (currys and johnlewis) in the background (you don't see it) and scrapes the price for every page.
5. Finally, IH compares all prices it's got and shown a notification (for example, 'This product is 10% cheaper at curry's than at amazon) and shows you a link to currys.
6. You save money!
So, the answer to your question is: no, IH doesn't check all retailers, it asks the server to do it for it but it scrapes the price locally, when you need it to ensure it's accurate. The reason we decided to scrape prices on the fly is that they really change often, sometimes several times per hours. So the list of alternative urls is provided by the central database but the prices are fetched in real-time a fraction of a second before you see them.
1. When you start Firefox, IH downloads the list of retailers it supports: amazon, currys, bestbuy etc. The reason for this is that you don't have to update the extension when we add a new retailer to our database and it happens fairly often.
2. When you open a new tab in firefox, it checks whether it's a retailer. So if the address begins with, say, 'http://www.amazon.co.uk/...', it proceeds to the next step, otherwise (for example f you open 'http://www.google.com/) it just stays inactive.
3. When it determines that you're at one of the supported retailer, it sends a request to the server, asking for alternative urls. Say, IH sends an amazon url to the server and the server replies that this product is available at currys and johnlewis as well.
4. Then IH opens these urls (currys and johnlewis) in the background (you don't see it) and scrapes the price for every page.
5. Finally, IH compares all prices it's got and shown a notification (for example, 'This product is 10% cheaper at curry's than at amazon) and shows you a link to currys.
6. You save money!
So, the answer to your question is: no, IH doesn't check all retailers, it asks the server to do it for it but it scrapes the price locally, when you need it to ensure it's accurate. The reason we decided to scrape prices on the fly is that they really change often, sometimes several times per hours. So the list of alternative urls is provided by the central database but the prices are fetched in real-time a fraction of a second before you see them.
kmote
Seg Fault'n,
- Messages
- 5,801
That's a very good way of doing it. Well thought through.
Sorry but i have to laugh still. Even according to your own picture that you posted on why databases are bad to use.
http://getinvisiblehand.com/wp-content/themes/invisible/images/pricegrabber-invisiblehand.jpg
If you actually read it, it states only 16% of all searches. Out of a low field of 143 searches that seems like a lot. That is just about 23 searches. What is the % of searches per day that are wrong of the thousands or millions? I bet that their numbers would be very similar to yours. As you stated yourself:
Yeah i will say that the databases are a bad way to go cause they are not updated. But to provide data and consider it fact on such a small number of searches, not even including what was searched, means absolutely nothing. For all we know you could have went out specifically to find these errors just to report them. You could have went to the main sites to find the daily deals just to show that the database was wrong.
Sorry i know a thing or 2 about putting a spin on words. Linking me to a picture that tries to show me proof, but doesnt back it up with any real facts just a couple of pictures and a blurb about how many searches were performed is not good enough. I could do the same thing with your software. Would that mean that i am proving my point and it is fact cause i can do 143 searches and find 10 errors? No.
So show me real facts. Get me at least 1,000 searches. Get me the items that were searched. Then i will give the product a bit more credit.
Gotta love those advertising twists to make 1 product look bad and another look good. IH sounds decent enough. The implementation sounds alright. I may try it.
http://getinvisiblehand.com/wp-content/themes/invisible/images/pricegrabber-invisiblehand.jpg
If you actually read it, it states only 16% of all searches. Out of a low field of 143 searches that seems like a lot. That is just about 23 searches. What is the % of searches per day that are wrong of the thousands or millions? I bet that their numbers would be very similar to yours. As you stated yourself:
Yeah i will say that the databases are a bad way to go cause they are not updated. But to provide data and consider it fact on such a small number of searches, not even including what was searched, means absolutely nothing. For all we know you could have went out specifically to find these errors just to report them. You could have went to the main sites to find the daily deals just to show that the database was wrong.
Sorry i know a thing or 2 about putting a spin on words. Linking me to a picture that tries to show me proof, but doesnt back it up with any real facts just a couple of pictures and a blurb about how many searches were performed is not good enough. I could do the same thing with your software. Would that mean that i am proving my point and it is fact cause i can do 143 searches and find 10 errors? No.
So show me real facts. Get me at least 1,000 searches. Get me the items that were searched. Then i will give the product a bit more credit.
Gotta love those advertising twists to make 1 product look bad and another look good. IH sounds decent enough. The implementation sounds alright. I may try it.
- Status
Similar threads
