Ask HN: My project is getting traffic but the product fails. What do I do?

13 points by meric ↗ HN
My website is http://textbookcentral.com.au/. I launched it during Australia universities' semester break at the beginning of last month.

My plan was whenever a student types "<course code> textbook", they'll get my website near the top. Try google "COFA0217 textbook" and see for yourself.

It worked and I'm getting on average 50+ visits a day, 75% from google, 15% bounce rate, 5 pages per visit. It has declined gradually by around 10-20% since the beginning of the semester, because, I'm guessing, students are no longer looking for textbooks now that we're approaching mid-semester.

How it works: Visitors find their university course and click "Buy textbook" or "Sell textbook". When there are buyers and sellers for the same course, they are both notified.

I've had around 400 offers+requests made in the past month. Considering there were only 1700 visits last month, an average of 0.23 offer/request made for each visit sounds great.

The problem: I've matched less than 10 courses so far. All the students were making offers/requests for textbooks for different courses, rarely matching up. I thought e.g. popular subjects would get more offers/requests but... not really. With 400 offers and requests and a 2.5% success match rate? I think that blows. :(

15 comments

[ 4.0 ms ] story [ 47.6 ms ] thread
Two sided marketplaces are hard. Is there any way you can fake the existence of one side of the marketplace? For example, there has to currently exist some way of buying used textbooks in Australia. It presumably is crazily inefficient, but it exists. Either a) roll up your sleeves and do it or b) pay some VA in the Philippines to do it for you.

Customer A: "I want a textbook for COFA0217 textbook." Site: "We've got it for $30." Customer: "Yay!" Site: "We'll send you an email when we ship it." Site in email to freelancer or yourself: "Acquire a copy of COFA0217 textbook for less than $30 and ship to $CUSTOMER." Freelancer, 4 hours later: "Did it!" Site: "Congratulations! We shipped your order. Tell your friends how much money they could make by listing their textbooks on our site!"

If you can't find a copy, you mail them personally, say that the book they wanted was out of stock, and refund their purchase (and/or you authorize but do not charge until shipment, which is similar to e.g. Amazon, but your credit card merchant account may or may not be happy with that).

I think Eric Ries called this the Wizard of Oz model: you can duplicate any arbitrarily complex system with near-zero engineering costs if you're willing to conceal a man behind a curtain. When you don't have traction, it's great, since it's just part-time work for yourself. Once you have traction, the "problem" of scaling solves itself.

There's two issues with that:

1. My website only exchanges buyers and sellers' contact details, I don't know when a sale is made. The idea is that since buyer and seller are in the same university, they'll just transact face-to-face.

2. I don't actually know the name of the textbooks for e.g COFA0217 unless the user enters it into the form. e.g. http://textbookcentral.com.au/26/university-of-new-south-wal...

If they do enter it into the form and I become the other side of the transaction, it'll be very expensive. Textbooks here cost $100 each. I'll have to make at least a $20 loss for each transaction.

What I need is more students looking for textbooks online, and I guess doing this will help. But matching a request/offer costs $20 in loss for the book (If you're buying a used textbook you're looking for at least a 20% discount), plus labor required to get the textbook.

Is a freelancer really going to be able to find a used textbook worth $80, send it to me, for the price $80? Also, I still have to send the book to the buyer/seller. This might sound suspicious; if I did the course why do I have to send it and not just meet at university? There will also be some students who receive the book by mail and not pay me, and I'll have to take the loss for that. (i.e I can do this for my university only.)

So the only way out is to change my website so I can use the tactic you described. That would be like changing it to the model of textbook.com, where they have an existing inventory and make dough from the margin between buyers and sellers.

I don't want to handle such a big business yet. (Higher stakes, requires lots of time) I want to run a website that helps people but requires minimal maintenance.

You should listen to patio11. He's basically saying 'fake it until you make it' and this is the key to success. You need to fake the transaction until you're big enough that you no longer need to fake the transaction. Reddit did this to get started and so have millions of other projects.

You need to build the system so that they pay you before they take possession of the book.

Also build or scrape an already existing book search website. Basically, take the cheapest one there and buy it and then try to sell it for cheaper than Amazon. Even put that so they can see it. Amazon, with shipping, is $98.54 but we are $94.94. Try to break even on each of these, and then scale it up.

As for why you can't meet them? You already graduated, you're selling it for your dead cousin, you found a bunch of your brothers old text books, you transfered universities, you got a job, etc.

A cousin poster said "I don't think it's gonna work because this isn't a site that will garner much loyalty. Students will eventually graduate, plus buying textbooks is seasonal. He's just gonna do a lot of manual work for one-time sales. It's not gonna improve traction in the long-term.".

Students can spread my website to their friends via words of mouth and improve traffic, but usually majority of a student's friends are in the same year...

This directly improves match rate though, because I'm trying to match as many as possible by myself. I'll think about it some more.

I think you just described why your business isn't working. Find the names of the textbooks for each course. If you can't match them up, send them to an affiliate site. Bottom line, though, is you didn't pick an easy thing to get off the ground. But you have a good starting base of traffic. Just keep working on the model until you can make it work by yourself.

In other words, the two problems you listed are fixable - fix them.

I don't think it's gonna work because this isn't a site that will garner much loyalty. Students will eventually graduate, plus buying textbooks is seasonal. He's just gonna do a lot of manual work for one-time sales. It's not gonna improve traction in the long-term.
2. I don't actually know the name of the textbooks for e.g COFA0217

Spend some manual labour (or pay someone else) to either (a) look at the course website (b) contact the local campus bookshop and ask them what the course is.

find a used textbook worth $80, send it to me, for the price $80? Also, I still have to send the book to the buyer/seller.

No point adding layers for no point. Get the freelancer to send it directly to the buyer/seller.

I think the biggest brazilian online bookstore does this with (imported?) books: if I search for, say, "Types and programming languages", they say its imported and will take 6 weeks plus shipping time to get to you. They never mention they are going to buy a single copy from Amazon for you and put some % on top of it. I don't know how they implement it to integrate with Amazon book list. I could do it for less % though...

On the other hand, they are providing some value: isolating customers from changes in USD rate and providing customer service.

Interesting. It seems like, if you had more traffic, it would work better. You have the typical chicken/egg problem of a marketplace: just not enough matching. The same problem dating sites have.

One possible approach: limit your site, to only 1 university, or only 1 topic (law or something). That might increase that match rate.

The other problem is that one university might call it PS101 and the other one PS1001 and they might use the same book, but he would never know.
Allow listing by text book, as an alternative to course number.
Good point; I shouldn't done some maths before I started. I thought by hiding how many users this website has, new users won't be daunted by the lack of other users.

But no.

Limiting my site will only serve to decrease traffic, I think this might be counter productive.

I'll think of ways I can increase traffic further cheaply.

A side-note: From where I am, this thread on Hacker News is currently the top result for "COFA0217 textbook" on Google.
Sorry, it can't compete with hacker news. :(