We found that lots of universities were still using Facebook groups to sell. It's inefficient, hard to find what you want, can't filter other than the chronological feed, and occasional searching of keywords if you know exactly what you want to buy. Thus, we sought to build a better solution.
What it does
Here at SlickMarketplace, we're proud to stand behind Slick Resale, an online platform for student to buy and sell items, particularly textbooks. Thus, we wanted to build a true searchable repository where people could easily post and research items with the kinds of filters you'd expect of a modern shopping interface.
Streamlining the buying process was something we wanted to devote a lot of time to, so our team built an image recognition system such that when students take a picture of their textbooks, we can recognize what they are, what the edition is, and estimate the resale value by comparing to existing/past listings on websites like Amazon and eBay.
A lot of the friction in selling is that students don't know how to price their item, and may severely underprice (and sell immediately, but lose out on money) or overprice (and never sell). Thus, we see our tool as a helpful third party appraiser that assists in the buying process - eventually, we see the ML/Image recognition becoming a virtual valet, if you will, to help you sell smarter.
How we built it
We utilized Twilio to verify users (so we avoid the scamminess of something like Craigslist). We utilized Coinbase as the current pay function such that individuals could pay/receive bitcoin for their goods - an easy, painless way to get more bitcoin into the hands of students (more effective than airdrops because students get a chance to earn some that otherwise may not have been interested/aware prior to using the platform).
Accomplishments that we're proud of
Working backend using Google Cloud Platform, and a lot of work in Google Sheets! Was super fun utilizing sheets as a backend to work with more visual data.
Our front-end marketplace leaves some styling to be desired, but we're super glad that all of the filtering and sorting works!
The GCP and AWS allowed us to do predictions on pricing and analyze/recognize books --> a fairly uniform, standard market. We'd love to see how flexible it would be to tackle markets where there's lots of models, lines with no system like ISBN to help us, such as blenders, where there could be thousands of undocumented models.
Challenges we ran into
We tried to make an elegant react-native app such that people could seamlessly upload pictures and quickly browse through our web catalog, but it was to no avail, so we still have a sad google form as the listing page (though at least the attach file works ok on mobile!)
Customizing Google forms is a lot more time consuming than we thought.
What we learned
We got Twilio and Coinbase set up fairly quickly, such that we could verify users/sellers and get our system ready for the cryptocurrency revolution, and allow for payments to be transacted through bitcoin :) *one comment for coinbase, we'd love to see an option to sent in any coin all at once, not 3 separate links to 3 separate widgets if we want to utilize any of the big 3 coins.
Also, GCP is super powerful, and we learned a lot by utilizing everything within the GSuite to hook things up :)
What's next for SlickMarketplace
We envision having enough data on second-hand resale value of items to actually understand brand sentiment (nobody wants Blackberry phones anymore, value is only 5% of MSRP face value) which we could then use to create actionable insights to perhaps trade (Blackrock api!) and pick some winners/ avoid some losers that do particularly well in resale value of products. This data we think would also help companies that really only see first-hand sales (ex: Apple with their certified retailers) and thus by providing good data vis on pricing trends over time of depreciating goods and bid/ask spreads between what sellers and buyers want to offer, we can help visualize demand/supply curves to better help companies price their products in the first place.
Built by: Jonathan, Peter, Calvin, Alex, and Sudipta