Ask HN: Best Price for any product
2 comments
What are your biggest type of clients? I see you used a phone as an example, is that one of the most common products that people ask the price for? Asking because I'm trying to get the mindset of the proper client.
Also, for whatever product your most common client is, what are the reasons that they use your service as opposed to comparing the prices on amazon/ebay/Google?
Also, for whatever product your most common client is, what are the reasons that they use your service as opposed to comparing the prices on amazon/ebay/Google?
Yes, phone makes about 20% of the requests for us. However, we have received requests all over the board from cement, steel, home appliances, industrial machinery, computer/laptop etc.
From our experience, the prices published online are not necessarily negotiated price. They come to us because we do online research as well as in person negotiation with bunch of retailers/distributor to get them the price. For example, we sold a TV worth USD 2300 for 1911, a price not available anywhere online.
In a way, this crowd sourcing platform will help discover negotiated prices that people are buying the products for.
NOTE: Negotiation is a common habit in India.
From our experience, the prices published online are not necessarily negotiated price. They come to us because we do online research as well as in person negotiation with bunch of retailers/distributor to get them the price. For example, we sold a TV worth USD 2300 for 1911, a price not available anywhere online.
In a way, this crowd sourcing platform will help discover negotiated prices that people are buying the products for.
NOTE: Negotiation is a common habit in India.
Hmmm, I see. That is very difficult. I would definitely make sure that point is made somewhere on your website or app so people unfamiliar with it are aware of your value proposition since I'm sure there will be others like me wondering why Google wouldn't suffice. That will probably make it worth the hassle of downloading another app/taking a picture of a receipt worth it for about 10% of your customers I would guess. As for the rest, I'm not sure how you could incentivize them better or what alternative methods of gathering this data could be.
Another problem you'll encounter are outlying data sets and different terms. For instance, if somebody buys a phone without a contract it could be about $600, but if you get the same phone with a contract it could be free. While that's just one example, it points out a clear problem with the taking pictures of receipts because you don't know the terms at which the product was acquired. Also, demand and intelligence may contaminate your numbers. Let's say somebody is selling a used TV which they bought for $2400 and selling it for $2300. Let's also say it's super bowl weekend and somebody just broke their tv and they're huge football fans and want to go buy a new TV really quick. They buy this used TV for such a high price because they're crunched for time and are desperate. I know it's not the best example but it gets the point across. Somebody else will rarely be able to recuperate the same value for their tv as that person just did. So then you think you'll just remove the outliers. However, if you do that, unless there's a ton of people getting great deals, you're going to be left with what the majority of people are paying for products, which will most likely be market rate, or the going rate without any negotiations.
Idk if you've encountered those problems, have solved them,or are currently working them out. If that is a real problem for you then it seems tricky to automatically knit pick on a case to case basis.
Keep fighting the good fight, jjalan.
Another problem you'll encounter are outlying data sets and different terms. For instance, if somebody buys a phone without a contract it could be about $600, but if you get the same phone with a contract it could be free. While that's just one example, it points out a clear problem with the taking pictures of receipts because you don't know the terms at which the product was acquired. Also, demand and intelligence may contaminate your numbers. Let's say somebody is selling a used TV which they bought for $2400 and selling it for $2300. Let's also say it's super bowl weekend and somebody just broke their tv and they're huge football fans and want to go buy a new TV really quick. They buy this used TV for such a high price because they're crunched for time and are desperate. I know it's not the best example but it gets the point across. Somebody else will rarely be able to recuperate the same value for their tv as that person just did. So then you think you'll just remove the outliers. However, if you do that, unless there's a ton of people getting great deals, you're going to be left with what the majority of people are paying for products, which will most likely be market rate, or the going rate without any negotiations.
Idk if you've encountered those problems, have solved them,or are currently working them out. If that is a real problem for you then it seems tricky to automatically knit pick on a case to case basis.
Keep fighting the good fight, jjalan.
> What are some of the leanest way you can think to test this experiment?
What is your most dangerous assumption that you need to be true?
What is your most dangerous assumption that you need to be true?
People are willing to share their purchasing info or receipt in exchange of
(1) Ability to request a negotiated price (which we will do on their behalf) AND/OR
(2) Ability to see at what price other people are buying the product nearby.
(1) Ability to request a negotiated price (which we will do on their behalf) AND/OR
(2) Ability to see at what price other people are buying the product nearby.
Currently, we physically go and negotiate with local sellers and research online to get the best price. We are now considering a pivot to scale it. At the core of this pivot, we realized that people are already negotiating before they buy today. If we can incentivize them to share this knowledge (at what price and from where they are buying), probabilistically we will get the best price in our system in long run.
Here is how we are planning to execute it: we are now introducing an app which will be top up with virtual currency (say 3 coins). User can still ask for best price for any product (it will cost them - say 1 coin). Once they run out of coins, we will ask them to upload a receipt of any recent purchase they made - say iPhone 6. We will extract the price and seller information from the receipts. Later on, if other user is looking for best price for iPhone 6, we can surface all the information from the crowd who has submitted receipts for iPhone 6, ofcourse anonymously. In long run, system should grow by itself and we should get out of business (or maybe doing it for free) of getting the best price.
(1) What are your thoughts on this model? (2) What are some of the leanest way you can think to test this experiment?
Looking forward to some constructive thoughts.