Auction Sites-Build Or Buy The Technology?

This not so, Anonymous Coward asks: "I am involved in a new Internet startup: a B2B Exchange. This will be an auction site where my commodity industry (based in Asia and Eastern Europe) gets together to buy and sell. The business premise is valid. My dilemma: "build or buy" the technology?" A not too uncommon question in the IT world, and one that greatly depends on your market and the resources you have available. Read on for more information...

"The technology is available to buy (everyone knows names like Ariba, Commerce One, Moai, etc., etc.). All great apps I am sure, but my impression (just an impression, so far I have not talked to the vendors!) is that these are just toolkits - that I would be paying half a million dollars to get some consultants for four months actually writing the app. And then I'd probably pay royalty fees, maintenance fees, etc.

The other option: build it ourselves. I could specify our requirements, then hire a team of programmers and tell them 'build it.' Advantages: quicker; more control; I can use open source technology; etc. Drawbacks: maybe we are re-inventing a wheel; and we do not benefit from other users' experience.

What do you all think? Is an auction mechanism doable by ourselves? Or are we, by doing that, just reinventing the wheel? Is auction technology mature yet to call it a "product"? Has anyone worked with these?

I'd highly value your opinion. Elsewhere I would just hear business hype."

ACS has one

[djweis]

Check photo.net [photo.net]. It's probably not as sophisticated as some you've seen but it's free software and represents a good start over starting from scratch.

Get the security right.

[bluGill]

I can write you an auction cgi. Not really hard. version 1.0 could (if you paid enough) work in a month. Not a pretty version, but it would work just fine.

I cannot make it secure though. Many people claim to be security experts. Most are lieing. To get all the security right is hard. ssl, and the like are simple technologies in theory, but there are too many subtile ways to screw them up.

The bad news, I don't think most consultants are better then me.

Check out Rocketry Online's auctions

[RocketJeff]

Don't know if it would meet your needs, but I really like the auction service on Rocketry Online. Rocketry Online is (IMHO) the premier Model/High Power Rocketry website.

Darrell Mobley, the owner/operator/programmer of Rocketry Online, does a great job in programming the auction area. I'd contact him directly to see if he'd be able to/want to do something similar for you.

Rocketry Online! [rocketryonline.com]
Rocketry Online Auction [rocketryonline.com]

Note: I'm just a happy user of Rocketry Online(but I wish I had more to do with it).

Why not buy ebay?

[lw54]

Why not just buy ebay? They've already got a working product. :-)

I all seriousness, buying one will always be the faster route. Even if you have to modify someone else's product, buying theirs will reduce the time it takes for your programmers to create yours.

Work that bottom line...

[Bald Wookie]

Instead of half-asking a bunch of half- knowledgable slashdotters, get out there and figure it out yourself. You mentioned several companies in your original post. Have you contacted the sales reps and gotten quotes from them yet? Nail each of them down on development and maintenance costs. Get a really solid picture of how much this software is going to cost over its lifespan. Keep factors like scalability in mind. Does their solution scale by throwing more hardware at the problem, or will there need to be significant software changes? Make sure that they understand the technology that you are planning on using (and vice-versa), and have a good grasp of the problem at hand. Find out how much of the project can be done with "shrink-wrapped" parts, and how much customization will need to be done. Once you have a number from each company, your work isn't finished.

Find out who their clients are, and go talk to them. Focus on the ones that are using technologies similar to your application. Ask them about issues like reliability, ability to meet deadlines, and hidden costs. Try to get the impressions of the tech workers who have worked with the products and consultants. If the sysadmin thinks that the consultants are a bunch of half-wits, that is a bad sign. If the client's technical documentation bloodstained, beware of serious problems. Some companies may be reluctant to talk about their experiences, but you can usually get creative about getting them to open up. Try to deal with individuals rather than the 'organization'. Doing your research is painful, but vital...

Once you have a pretty good picture of each company, pick the two or three that you are most comfortable with. Compare issues like fixed and recurring costs, maintenance and support contracts, quality of work as evidenced by their portfolio, and the comments of their clients. Read up on the base products, but ignore anything from a marketing department. Make a nice big file on each, and hang on to them.

Next, look at your options for doing this in-house. Set your budget equal to the cheapest pre-made solution that you would consider using. Then sit down and grind out how much it will cost you to do it yourself. When budgeting for personnel costs, dont forget to pick high-range salary estimates, factor in insurance and payroll costs, and do it right. Compare the costs of salaried personnel with hired guns. Can you even find qualified developers for the project? Does the infrastructure for development already exist, or do you need to buy a bunch of hardware? How many (and what types) of developers do you need? Is your office big enough to house a new development team? How long will it take to do it in house? What about maintenance? How many team members will you need to keep things running once development is complete? All of these are tough questions, and need to be asked. Keep in mind the goals of the project, the timeframe you have to work with, and budget for disaster. If in-house development turns out to be way cheaper, double check your numbers. You might have missed something significant...

Next, compare the costs of doing it in house with the costs of outsourcing. If the costs (in both time and money) turn out to be equal, you are pretty much at square one. How much is having total control worth to you? Will the level of customization made possible by in-house development pay of in tangible (monetary) ways? Are you better off trusting your project management skills, or letting the pre-packaged solutions provider deal with things? Remember, if you blow it in-house, you (and your investors) will eat every dime. If outsourcing to a reliable solutions provider gives you a higher chance of getting a working product in a reasonable time-frame, it is probably worth writing a bigger check. Plus, if they don't deliver, then you have someone to sic the legal department on.

Run the numbers compared to projected earnings and cash reserves. Can you afford this? When will you run out of money? If it turns out that you can do things better, faster, and cheaper in house then that is probably the way to go. Otherwise it is probably worth paying the piper and getting the best pre-packaged solution for the money. All of these questions are things that slashdotters can't answer. My gut says to outsource it if your plans are complicated or you are unsure of your ability to build a good team. If the solution is simple and can be churned out by a small development team, then that may be the way to go. The final call depends on information that only you have access to and the weight of your decision rests on your shoulders alone.


-BW