I've been thinking lately about opening a website similiar to Yahoo Answers / Answers bag, ofcourse with a lot of tweaks and new ideas added.
I've gone to a couple of web development companies and I got two estimates: 1. about 20k $ 2. about 200k $
The thing is that the company that gave me the 200K $ said that this job is impossible to get done at 20k $, there is no way that the first company can develop a site that will compete with the big players in the market (Yahoo Answers mainly) at 20k $. They both offer the same (with both I just talked and did some very basic drawing).
What the company with the 200k $ said is very logic from one hand but on the other hand... its not that comlicated that it should get to 200k $.
What would you say? did he just played around cause he doesnt want the job and gave me a high estimate? or is he right and a website with high standards, which would be able to compete with the big ones, should really cost 200k $?
I am lost. Please give me some of your knowledge or any hint which will help me decide.
Another question I was wondering about: If I put Google Adsense on this kind of website (QnA) do you think it should get a higher CTR than the useful sites? I know this question is a little vague, but I was wondering if this kind of website should be using a Google Adsense / Yahoo Marketing Ads or should it be based on independent adversiting?
Welcome to WebmasterWorld, talco.
You didn't state in so many words, but I gather that you went to the outfits asking them to quote the application development and that hosting etc. wasn't part of the service that they'd provide you.
If that's true, think about this: you've written up an application specification, whether it's as informal as a drawing on a napkin or as formal and complete as a big 20 page outline with supporting documents, photos with circles and arrows and a paragraph on the back of each one. To quote the job, a developer is going to estimate the number of hours to do the programming [and data entry if part of the spec is their typing in all the answers], then multiply that by his/her 'shop rate' of dollars per hour. The developer hands you back that dollar amount, whether it's sliding the napkin back to your side of the table or returning a big 20 page document with breakdown by feature and alternate options and demographics and all that marketing mumbo jumbo.
What you're going to do with the application isn't really part of the equation.
To a certain extent the programmer needs to know how many people might be accessing the application at the same time, but that information shouldn't, in most cases, wind up multiplying the end result by 10.
|What the company with the 200k $ said is very logic from one hand |
Not really. In the physical world a design is very dependent on its application - you wouldn't expect a subcompact runabout to climb around rocks and hills the same way a vehicle designed for off-roading does. Both have the same features: 4 wheels, stereo, engine, suspension. But the off-roader's suspension & structure are designed for its application, and it makes sense that it's more expensive - heavier chassis, tougher tires, suspension has much better range of motion. Software, on the other hand, while to some extent needing to be designed to the scale of its application, largely either works or it doesn't.
If part of what you want them to do is host the application, then the operational scale would determine how much hardware and bandwidth needs to be thrown at it and would impact the end number by a significant amount. Although I'm having a hard time wrapping my head around a factor of 10, I couldn't really blindly disagree with it on this side of my monitor looking at an internet forum.
Now it's entirely possible that the 20k people have underestimated how many hours' development your project will take, it even could be that the 20k people employ inexperienced undereducated hacks who were flipping burgers (nothing wrong with that ;)) last month while the 200k are staffed with top-notch cream-of-the-crop experts. However, with the information given, I suspect that 20k is closer to the mark than 200k.
Before spending either of those amounts, I'd move out of the napkin realm. I highly suggest that you locate a software engineer to draw up a comprehensive spec for you. It's kind of like having a lawyer write up a contract. A software engineer knows the lingo and the things that you don't even know you need. It will remove any ambiguity for the programmers and the people quoting what the programmers are expected to produce. Yes it's an extra expense up front, but could save you an awful lot of money and aggravation in the long run.
Where do I get this software engineer? or in this case... I guess it should be a website developer that will do only this diagrams for me...
Its seems like 20k $ is more down to the earth, but since I am a newbie, with a itch to invest, I should be really careful and check out that I am not jumping into an amazing car with no engine installed :)
What is your experience with developing websites abroad? some suggested me that if I am talking about this serious ammounts (20k $ +) I should be checking the costs and quality of developing such a website in India. any experience?
For the sake of comparison, digg.com was started for $200.00. Since then they've spent a lot more than $200k in hardware and software to scale it, but that was only after they started getting the heavy traffic to justify it.
Very good point.
Try to get something cheap-and-cheerful started, not the finest most-robust implementation, but not one that will embarrass you or leak personal or financial info for example. You could probably get something going for no more than $5k outside, if you explain that this is what you're doing.
Call this a feasibility study if you like. Be prepared to THROW AWAY the initial implementation, but make sure that you can preserve the data and move it to any "upgrade" site.
*IF* that starts to bring in the users, THEN pay the money to have the bells-and-whistles done, and you and your implementors will have a much closer idea of what needs doing anyway.
Incremental development (and the willingness to throw away earlier versions of code but keep the data, which is the valuable bit in most cases) are the core to working systems, in my personal and consulting experience.
talco, if you were building a 10 story office building, would you rather have a structural engineer design it or the guy that does the framing? The framer is likely an expert in his/her field and in some cases knows things that the 'college boy' engineer doesn't, but the engineer is [supposedly] qualified to look at the big picture and know how to integrate the foundation, framing, electrical, etc.
Same with a software engineer. There's the commercial exchange right here on Webmaster World. Try typing software engineer into a search engine. Find groups of software freelancers (search engines, programming magazines); there's bound to be SW engineers hanging in close proximity. Do a want ad, post on employment sites, etc. Then have them show you their qualifications. You want one that has experience in what you want to achieve: site that does this for this many people. You might even want the qualified candidate to supervise the whole project, who knows. The first step, though, is to very clearly define the scope and specifications for the project.
I don't think I have anything specifically against India, but personally I would prefer to keep it in my own country. Besides the obvious patriotism, I'd want someone closer to home for accountability - it's a lot easier to thwap someone upside the head when they're a few hundred miles away than in different legal jurisdiction oceans and thousands of miles away. I don't know where you live, but in the USA, I have to think that I can find some pretty outstanding talent in Iowa that can work at Iowa prices - while perhaps not as cheap as India, also not as expensive as the cats that have to pay for housing in Southern Cal. I can get face time for a plane ticket at a couple hundred bucks and no passport or shots required.
DamonHD, I understand what you're saying and tend to agree especially with the 'know when to throw it away' part. Incremental development is definitely the only avenue when you're charting unknown territory because you're learning from mistakes and unanticipated directions. On the other hand, talco's project isn't groundbreaking heretofore unconceived theory. talco knows "I want this, only I want that to be blue, I want this over here, and I want it to do that as well". With a software engineer experienced in large scale data systems, wouldn't it be much more efficient to design the application floor up to service its audience? Basically you're using his/her experience to start with version 5 - versions 1 through 4 were essentially paid for by previous employers/clients. What can't be programmed in 6 months? 40 * 4 * 6 * 75/hr = 72,000. More than 20 but less than 200. 6 months has to be worst case, I'd think 2-3 would be more reasonable. Then grow the hardware as necessary - so it starts at a server farm (or wherever) and works its way up to being a multi-processor internet backbone connected monstrosity.
Definitely mep00 and DamonHD have very valid points in starting small and working upward. It really depends on your budget and how you want to hit it - if you have or can procur the funds and want to hit hard and be ready for tons of traffic (which may admittedly never come), go for that. If 20K means belt-tightening, 21K is completely out of the question, or you just don't want to hit it hard, then small+growth is the way to go.
It seems to me that neither the 20 or the 200 number mean much at this point.
If they mean anything, it might be that they weren't given enough, or equal, information to work with.
A third number from a third provider might have let you know which end, if either, of the scale to begin to think of as a serious range.
I think the major unknown for the OP seems to be the project spec/management part, even if the OP believes that they know exactly what the service should look like (maybe they do, maybe they don't).
Thus my cheap-and-cheerful suggestion is a way of exploring the truth of this, with a low risk if the OP is wrong in their assumptions of spec/traffic/etc.
It's basically a risk-minimisation strategy unless you are SURE (as you suggest) and RIGHT to be sure. We may all believe the latter, but are we always right. Ahem... B^>
I curious, how different is yahoo answers from any of the 1,000s of forums on the net right now, webmasters world for instance?
I think the name yahoo is the diff
So , whether you spend $20k or $200, I guess you must have factored in all the outlay you'll require to prise users away from your competion