How can I recreate this website?

March 8, 2012 8:52 AM

How do I go about getting a website designed without getting taken to the cleaners, as I know next to nothing about the process?

I am a bookseller that currently sales on Amazon, Half, Valore, etc and after years of steady increase in sales and profit I have kind of reached a point where in order to grow my business more I will have to find new sources for getting inventory. One thing that I want to start is buying books from the general public through a website similar to this one. Basically, the seller would list their ISBN's and the site returns a quote for the books.

Here is a list of 5 ISBN's to try out on the site:

0803955022
0415232104
0521791243
9812615997
0521894662

As you can see, it rejects some but accepts others. I assume it uses an algorithm where it checks resale price and sales rank and (possibly) compares with the companies current inventory to come up with the price. Then it has a kind of inverted checkout process where they send a mailing label and the seller sends the books in and money is dispersed back out. The business part of it I can do, it's the tech part that I need help with as I know nothing about any of how this works and I don;t want to go into a meeting with a company/person who does this kind of thing and get fleeced because of my ignorance.

So, how much would something like this take? How much maintenance is required? What am I looking at cost wise? What are the questions I need to be asking, etc?

Anything you can tell me is greatly appreciated.

*Special thanks to Jessamyn for helping me frame this.

posted by holdkris99 to computers & internet (10 answers total)

To keep your costs low, you might look at freelance web developers. Ask them to write up a detailed proposal for you if they're interested in the work.

In my area (California) I would be a bit wary if I heard of a site like this being built for less than $10K. They've obviously got a marketing & development budget too, since they've had a TV commercial made for them.
posted by circular at 9:14 AM on March 8, 2012

I'm less familiar with the cost side (which varies widely), but you are looking at 2 basic needs:

1) A website with shopping cart software
2) An application that looks up an ISBN in a database, compares it to your inventory, and uses a formula to accept or reject it.

Number 1 is incredibly standard and should be easy to source inexpensively (<>
You basically have two options with number 2, since it's custom software (unless you can find pre-built software to do this). You could contract it out to a known company that has good reviews. It will probably be expensive, but it will probably also be done on time and easier to maintain. Second, you could go to a rent-a-coder type site and see what you get. You might get some very cheap bids, but the delivery from those companies is pretty spotty, so caveat emptor. Depends on how much you have to spend and how quickly you need it.
posted by zug at 9:16 AM on March 8, 2012

Good advice from zug. I'd advise against saying you're looking to get a website "designed" - as that implies the visual elements only. You're looking to get a website developed (and designed, but that might be a smaller part).

You're looking for a relatively complicated website that forms the foundation for a real business. This might be an unconventional approach, but I'd recommend beginning NOT by going to developers, but by working or partnering with someone who can take on a "product manager" type role. This person will work with you to create things like specifications and wireframes, with clear explanation of the functionalities and flows you are looking to have created.

Developers are generally NOT good at figuring out these kinds of things, but they're really important in making sure a) you get what you want; b) you get an accurate price quote; and c) your website is user-friendly and not just using the shortcuts a developer has chosen for a number of reasons.

Once you figure these things out, it's time to get quotes from 1) a developer; and 2) a designer. Perhaps there is a team you can find that works together. Ideally, the "product manager" you've partnered with has a sense of the different technologies that might be used and their relative pros and cons, and can help in selecting these people (and locating them). I would say $10k for this type of a project would be on the low side, but it's definitely possible.
posted by iamscott at 9:35 AM on March 8, 2012 [1 favorite]

I actually work on some book-related software, so here's my perspective.

There's a fair amount you can easily do with an ISBN and some web services: See if it's valid, see if Amazon has it, and get some pricing info from them. I don't know offhand if they give you sales rank info. Note that Amazon's usage policy pretty clearly states that you're only supposed to be using the API to drive customer traffic to Amazon, which you don't plan on doing. No one's gonna come arrest you, but be aware that you operate at their mercy and they may decide to cut you off one day without warning.

Beyond that, to be honest, there's a lot of magic that needs to happen. What you're dealing with is a dynamic price quote system, which presents 2 challenges:

- I don't think standard shopping cart software can deal with it anywhere near "out of the box", since you don't have a fixed catalog of merch. You're generating offer quotes that need to be generated, "purchased", and fulfilled within some window of time.

- Your thinking about a pricing algorithm is a good start, but consider some consequences. If it's checking existing inventory, that assumes a back-end system for managing that inventory. From a development standpoint, that's another piece of software- probably a more complicated one than your front end. Also, is there a threshold below which you won't give out an offer (though I assume you've got the arbitrage end of this figured out)?
posted by mkultra at 9:45 AM on March 8, 2012 [1 favorite]

Having managed getting a website built from scratch recently with some fairly complicated development, all I can say is keep 20% of your total budget in reserve - just keep the scope 20% lower in terms of funding than what you have on hand, or demand in the contract that the web developer hold 20% in reserve for change requests.

Also, prioritize your requirements ahead of time, and get into as much detail as possible. Get into the detail, but never lose sight of what your top 3 or top 5 must-haves are.

Do mockups yourself before you engage a developer. The mockups should give a general indication of the look and feel of the website, the functionality, and user behaviour.

Having a very good idea of what you want beforehand will save time, and time equals money.

Hire professionals, but aim for a smaller shop. At your size, be wary of places employing project managers or information architects.

But don't cheap out.
posted by KokuRyu at 10:02 AM on March 8, 2012

Oh, forgot one very VERY important thing.

Insist that the contract withholds at least 25%, and preferably 40-50% of the payment until contract completion. I've heard of a lot of cases where people pay all of the money up front and then don't have a working product at the end of it, especially with the rent-a-coders.
posted by zug at 10:10 AM on March 8, 2012

There is some great advice in this thread already. But I'm going to disagree with KokuRyu slightly because I feel that for a project like this you want to make sure that you either have someone with software project management/business analyst experience, or hire a shop that offers those services. When you describe what you want it seems simple enough, but if someone is going to be developing it from scratch you need to be able to explain the requirements in a way that a developer will really understand exactly what you want. This often means wireframes, functional diagrams/flowcharts, etc.

I get where KokuRyu is going though, some firms pad their pricing with project managers who don't do a whole lot, so maybe you could find a place that offers what I've described but has the developer working on those tasks instead of an actual project manager. You also want to build in some a prototype deliverable to the project so you can make sure the site/software is doing what you expect it to be doing.

This really doesn't have to be a huge expense for you if you're just looking for some basic functionality, but you don't want to just throw money away on a system that doesn't give you what you want.
posted by Kimberly at 10:22 AM on March 8, 2012

I assume it uses an algorithm where it checks resale price and sales rank and (possibly) compares with the companies current inventory to come up with the price. ... The business part of it I can do, it's the tech part that I need help with as I know nothing about any of how this works and I don;t want to go into a meeting with a company/person who does this kind of thing and get fleeced because of my ignorance.

Kimberly comments about the need to have someone who is a quality designer/pm but I'll suggest one way you can help yourself up-front. You say you can do the business part, so if you were the magic elf in the web server doing this by hand, how would you do it?

This is probably the most important part you can suss out here and you'll save yourself misery and time by determining exactly how you'd handle this yourself. Are you going to look at what the current sales price and inventory are on Amazon? Half.com? Is there some other database you'll be looking at? Will you then offer 50% of that sales price? 50% minus shipping?

There's nothing magic about an algorithm - it's just a distilled series of steps covering every contingency. You may need to hire a programmer to implement this but you will help yourself get a good product and an accurate cost estimate by finding the resources for those decisions and having them up-front. If you came to me and asked for a quote with a statement as generic as "decide what to pay for them" I'd assume I'm going to be spending a lot of time with you developing your requirements.

You'll also hire better if you have a good functional requirement document up-front. When you present it to someone you'll be able to get an idea based on what questions they ask of you whether they understand your problem. And when you have an idea of how hard it is to do by hand you'll be better able to determine when someone is pulling your leg about how hard it is to automate.
posted by phearlez at 11:24 AM on March 8, 2012 [1 favorite]

I get where KokuRyu is going though, some firms pad their pricing with project managers who don't do a whole lot,

I would add that it's that the shops *call* these folks project managers or information architects or whatever, but in reality they are little more than account managers or sales reps (but priced at $120/hour).

However, if they're the real deal, and can help narrow the scope of the project and keep it on budget while attaining your core features, go for it!

That said, bake deliverables into your contract.
posted by KokuRyu at 12:19 PM on March 8, 2012

it's the tech part that I need help with as I know nothing about any of how this works

I think the technical aspect here is really a business problem first. If I understand correctly, you want a system that accepts an ISBN (or several) from a book seller and quotes them a price you are willing to pay. I can think of three ways this might be done, all of them probably wrong:

1) You create a short list of books you really want and what you will pay for them. The odds of matching up seem slim here.

2) The system ties into your inventory system and observes that some titles (or possibly genres) sell like hot cakes. Using this, the software divines that the submitted ISBN should be acquired.

3) The system queries some other BigDatabase, such as Amazon, and determines a fair market value and your system returns a quote at some lesser amount.

From my outside perspective as a programmer this appears to be the missing piece of the puzzle. Don't ask the programmer to figure this out because the programmer will attract toward the solution that yields the more elegant internal architecture, not the solution that yields the widest delta between bid and ask prices.
posted by dgran at 2:06 PM on March 8, 2012