Build a Bebo!

Adam

Hi guys,

Following a discussion with a few different Irish developers, I've come to boards for a more inclusive discussion regarding build costs.

Let's go back to Bebo a year ago, when it was still a growing trend but without all these fancy "apps" and "widgets", and just a site of good solid coding and fewer excessive features. Social interactivity, video/music features and photo sharing.

Now with quotes from Ireland of €50k for example, and then businesses on elance.com and rentacoder.com quoting roughly $3k for bebo/myspace/facebook clones, where is the huge gap in pricing coming in?

yeraulone

$3k - really? I wonder if it would work!

A couple of friends and I have been putting together something similar and we are about 80% finished after 3 months. Our spec is pretty high though, with a lot of user management tools for our administrator, a lot of web services, along with all the other user functionality that normally comes with social network sites. If done properly it’s a big job. I can see how it would cost a lot to get it done with a third party company.

AARRRGH

Mirror wrote: »

Now with quotes from Ireland of €50k for example, and then businesses on elance.com and rentacoder.com quoting roughly $3k for bebo/myspace/facebook clones, where is the huge gap in pricing coming in?

Two things -

1. The Irish guys will provide a lot more documentation and spend more time working on the design spec etc. The end result will be better quality code (IMO outsourced code can be hard to maintain.)
2. Wage differences.

I outsource a lot of work (also, part of my day job involves managing outsourced projects) so I'd be quite familiar with the process.

The key is being really really specific with what you want.

EL_Loco

I suppose you'd have to look at where the coder lives, and do a comparison on living costs. That's the first major gap. Of course the whole rip off Ireland part may make up some of the gap too.

so take your pick of factors. Partly the same reason manufactoring is moving east, people don't get paid as much.

Adam

And what did you find to be the primary time consumer on the build? As with the outsourcing, even the time frame is a lot smaller.

And let's consider the design is done and it's purely for development.

tomED

Hi Guys,

I've already spoke with Mirror on this topic, since we were one of the developers he contacted.

As per my emails with you, there are a few key factors here:

1. Being specific
As DublinDude rightly says, the key is being specific. When someone approaches a web development firm and says I want you to build me a bebo, what do you expect them to do? Tell you they could build bebo in a week? If they do, don't use them, they are being unrealistic.

2. Comparing Irish Prices with Foreign Countries
You could get a decent COMPANY on Elance, the best, for about $20 per hour for one developer. A project to rebuild a site similar to bebo, would take at least 3 months. That works out about $10k. A similar price for 3 months work in Ireland would be €30k. That's just one developer, no account has been taken for a designer, tester, project management etc. Oh and that price is very competitive for a COMPANY in Ireland.

3. Underestimating the work
One of the biggest things people that approach us have in common is their complete under estimation of whats involved in developing a website. It's nothing new, everyone does it. Just look at how long it has taken yeraulone, who has a group of friends working on his site.

Speaking for 2bscene, if someone approaches us to "build a bebo", we wouldn't touch it for less than €50k.

Tom

CoNfOuNd

What's going to be so different about your bebo? Are you looking to do it better or just create another similar community and make money from advertisements etc.?

$3k is obviously completely unrealistic. €50k is a big investment but it does sound more realistic - but it all comes down to a very big contract with a long list of everything the company will code, when they'll code it by and when they'll get paid. Lots of planning in other words.

jmcc

Mirror wrote: »

Hi guys,

Following a discussion with a few different Irish developers, I've come to boards for a more inclusive discussion regarding build costs.

Just to put this in perspective: most web developers haven't a clue about what goes on in a large scale, high traffic, highly interactive site. There is a lot more to sites like Bebo than mere web design. The backend, the database servers and other services, accounts for most of the design of a site like Bebo.

If you want a good introduction to this kind of thing, there is an O'Reilly book "Building Scalable Websites" that is worth reading.

Now with quotes from Ireland of €50k for example, and then businesses on elance.com and rentacoder.com quoting roughly $3k for bebo/myspace/facebook clones, where is the huge gap in pricing coming in?

A: as other people have said on the thread, the difference in the cost of living.
B: A complete misunderstanding of what it takes to build a site like Bebo.
C: A Bebo like site isn't just web design but that's all you are getting quoted for.

Regards...jmcc

AARRRGH

tomED wrote: »

One of the biggest things people that approach us have in common is their complete under estimation of whats involved in developing a website. It's nothing new, everyone does it. Just look at how long it has taken yeraulone, who has a group of friends working on his site.

Yeah, this is a problem alright. Developing a complete website really is a software development project.

One of my sites, staff.ie, I've been working on that for years. Getting it just right takes a long time. It may look simple but there's tons of stuff going on under the hood...

Bob the Builder

dublindude wrote: »

One of my sites, staff.ie, I've been working on that for years. Getting it just right takes a long time. It may look simple but there's tons of stuff going on under the hood...

People do think that bebo is just a matter of programming a content system, a registration system and a youtube embed feature.

Thr truth is that there's a lot more to it than that, and that it requires concise planning even before the project begins. If you're doing a proper site, you're as well off to pay €50,000 for a perfectly finished website(or close to it), with good navignation and simple nifty features rather than a push around.

When you think of it, €50,000 is only a years wages for a reasonably well paid employee, because if the website takes off, you can expand and pay back debts, etc. whereas, a €3000 website will have more cracks in it than you could count.

Adam

nevf wrote: »

People do think that bebo is just a matter of programming a content system, a registration system and a youtube embed feature.

This is the part that bothers me. It's no more than this, if you build a small site with really clean, concise and efficient, DRY code, are you telling me that to get a scaling guru in to bring it up to speed for the masses you would have to pay 50k?

Not that I'm asking you specifically nevf, but everybody keeps saying "it's more than this, it's bigger than that" when in reality, nobody is in a better position to answer my question than I am myself, because nobody here that I'm aware of has built a site like bebo et al! So where's all this worldly knowledge coming from guys? At least I'm playing it humbly in that I don't know where the gap is. I'm not jumping around the place screaming mad figures.

Bob the Builder

Mirror wrote: »

This is the part that bothers me. It's no more than this, if you build a small site with really clean, concise and efficient, DRY code, are you telling me that to get a scaling guru in to bring it up to speed for the masses you would have to pay 50k?

Not that I'm asking you specifically nevf, but everybody keeps saying "it's more than this, it's bigger than that" when in reality, nobody is in a better position to answer my question than I am myself, because nobody here that I'm aware of has built a site like bebo et al! So where's all this worldly knowledge coming from guys? At least I'm playing it humbly in that I don't know where the gap is. I'm not jumping around the place screaming mad figures.

I understand what you mean, and just for the record, I have never ever built a website to that scale, but there are things like Search Engine Optimizations between accounts, and privacy settings/controls. Specific settings have to be built into the server so it takes less time to process. There are core things like comment feature, private mail, maybe a blog.

If I thought I saw an ugly looking, sh#t navigation website that i was supposed to use to talk to my friends, I would creak. I dont like websites like nimble.ie at all because of their bad navigation and the fact that you have to conform to messy graphics, etc. This was a website which was designed very well in terms of programming, but the preparation involved was dire.

Personally, imo, I believe that €3000 is too cheap and €50000 is much too expensive. If the website is as simple as you are describing there, I would imagine that €25-30k, would sort out everything from design to marketing to development without a problem.

You may already have seen it, but I'll just include it if you didn't:
The Anchor WATCH_YOUR_SPACE Survey
Survey of Irish Teenagers Use of Social Networking Websites
http://www.webwise.ie/GenPDF.aspx?id=1744

Good luck with it, and even tho my opinions can't be taken for true word, because i lack experience on the development part, but my experience of Bebo and Nimble, and even less used ones in Ireland like Myspace are cold and bitter in some areas. Be sure that I'll be on the site to test it out, and promote it if you decide to go ahead.

Regards,
Nev

Ziycon

I heard that bebo was a college project that they decided to put on the net which just took off!?

Y2J_MUFC

Bloody hell, thats a fairly handy website to build in fairness. 50K is extortion. People build social networking websites for their final year projects left right and center. Its a one person job for a couple of weeks, max.

Adam

Thank you for the realism in the last couple of posts, I'm starting to believe I'm not going mad now!

This is what I've been trying to say, yes folks, bebo is huge! But only because of the massive number of users on it! As I said to TomED in email, imagine if bebo only had two users. When I emailed him initially he reacted as though I was expecting him to recruit 40+ million users as well!

Bebo is not huge (for the most part if you drop the frills), it's standards with nice design. It's not using anything new, or any ground breaking technology, it's all tried and tested stuff that's been around a couple of years at least.

jmcc

Mirror wrote: »

This is the part that bothers me. It's no more than this, if you build a small site with really clean, concise and efficient, DRY code, are you telling me that to get a scaling guru in to bring it up to speed for the masses you would have to pay 50k?

A lot more. Good database administrators and architects are very expensive. This kind of project involves more design and optimisation than sticking a few things in a MySQL shopping catalogue database. The data is continually changing and the design of a good system for that kind of traffic would probably be much more than 50K.

when in reality, nobody is in a better position to answer my question than I am myself, because nobody here that I'm aware of has built a site like bebo et al!

There are people here who have built sites that are high traffic and database backed. Just what do you think boards.ie is? Some people here have worked on major webdev projects. Not everyone here is "just" a webdev.

So where's all this worldly knowledge coming from guys? At least I'm playing it humbly in that I don't know where the gap is. I'm not jumping around the place screaming mad figures.

Well I work with very large databases (com/net/org/biz/info/mobi/ie/eu/uk/de/fr/etc zones) and track the registration/deletion and movements of domains in those TLDs. I also run website surveys of some of those domains. In the past, I've built search engines and directories. My current web dev project has approximately 117 million webpages and covers the hosting history of every web hoster since 2000 and also includes the history of various domains. The bandwidth calculations alone are terrifying. And my web dev skills are far from brilliant.

But if you really want a good understanding of the issues of scalable websites read that book I recommended earlier. It is by the lead developer of Flickr.com.

Regards...jmcc

Adam

jmcc wrote: »

A lot more. Good database administrators and architects are very expensive. This kind of project involves more design and optimisation than sticking a few things in a MySQL shopping catalogue database. The data is continually changing and the design of a good system for that kind of traffic would probably be much more than 50K.

So in this case are others wrong to be quoting 50k for the whole package when you could pay this just for the architect?

jmcc wrote: »

There are people here who have built sites that are high traffic and database backed. Just what do you think boards.ie is? Some people here have worked on major webdev projects. Not everyone here is "just" a webdev.

And I understand that boards.ie is quite large in terms of scale, but when I said "here" I meant on this thread. Point taken though.

jmcc wrote: »

Well I work with very large databases (com/net/org/biz/info/mobi/ie/eu/uk/de/fr/etc zones) and track the registration/deletion and movements of domains in those TLDs. I also run website surveys of some of those domains. In the past, I've built search engines and directories. My current web dev project has approximately 117 million webpages and covers the hosting history of every web hoster since 2000 and also includes the history of various domains. The bandwidth calculations alone are terrifying. And my web dev skills are far from brilliant.

I think this merely reinforces my point. You don't claim to be brilliantly skilled. Nor do I. Solid I guess would be a sufficient word. And yet we can do things like this. I guess what's becoming sickeningly apparent is that if DIY is an option, it's the only way to go!

jmcc wrote: »

But if you really want a good understanding of the issues of scalable websites read that book I recommended earlier. It is by the lead developer of Flickr.com.

Regards...jmcc

Thanks for the advice, I will definitely be picking up a copy!

jmcc

Mirror wrote: »

So in this case are others wrong to be quoting 50k for the whole package when you could pay this just for the architect?

It is a question of the scale. If you want serious scalability (100K+ users rather than <1K) then the skillsets change. It becomes somewhat less of a webdev project and more of a database type project. Web dev is still needed but the concerns of running a high availability, high traffic db become a lot more important. In effect, the database is the site.

The most important part for now is the specification. You have to know what you want to achieve, how you want to achieve it and how much you are willing to pay to achieve it. The best thing would be to sit down and work out the general structure of the project. You have to be able to put what you want to do into hard figures. Otherwise you will get figures that are all across the spectrum and the project will drift aimlessly.

Thanks for the advice, I will definitely be picking up a copy!

It is not a how-to book but it does cover a lot of ground. Perhaps one of the best parts of it is that it defines the elements of scalability:
"The system can accommodate increased usage.
The system can accommodate an increased dataset.
The system is maintainable."

Regards...jmcc

Ginger

Just to give you an idea of scalability and design..

A company I used to work built a document management system for one of its clients. All through the project which took about 3 months, everything worked fine as the client expected and performed in tests ok.

Then when it went to production, stuff started to slow down. Initially when it was tested there was maybe 20 users at a time on it. Now there was 200 closing fast to 500 at a time. Not that much of a jump, but because the developer hadnt designed with this in mind, the system was starting to suffer. We had 90 seconds return times on resultsets of 111 records from a 4GB database.

The database wasnt designed or optimised correctly for these loads. It took 2months of redesign work to get it fixed and back to a working situation. Its now reacting at .5 seconds for the above resultset regardless of number of users online. That was 2001 and its been upscaled since with no issues as the design took care of that, it allowed it to be distributed as required.

The database design and data access layer will have to be optimal with no bottlenecks. The database will have to be able to take the load that you expect. As jmcc pointed out, when you start scaling into the bigger projects your backend and infrastructure design become more important. You have to start looking at failover and distributed loads.

Plus there is hosting charges for 24/7 99.9% uptime levels and also traffic charges on top of that for bandwidth depending on the package you choose. Does your provider have the infrastructure to cope with that.

With offshore development you also have to take into account time and cultural differences. Having worked with an offshore development agency, I found out a few things. It may be cheaper, but god help you if you dont have a good and dedicated project manager on this side. It can become quite frustrating. Reporting bugs can become a minefield. Specifications must be exact and not presume anything. Other factors such as documentation and source code security must be considered.

Its not all about price, its about quality of work and more importantly comeback/accountability in case anything goes wrong.

From my own experience, designing a web application for en masse consumption takes a lot of work and requires multiple skills such as DB design, network infrastructure,software development & localisation.

tomED

Mirror wrote: »

Thank you for the realism in the last couple of posts, I'm starting to believe I'm not going mad now!

This is what I've been trying to say, yes folks, bebo is huge! But only because of the massive number of users on it! As I said to TomED in email, imagine if bebo only had two users. When I emailed him initially he reacted as though I was expecting him to recruit 40+ million users as well!

Bebo is not huge (for the most part if you drop the frills), it's standards with nice design. It's not using anything new, or any ground breaking technology, it's all tried and tested stuff that's been around a couple of years at least.

Bebo isn't only large because of the number of users. If there were only 2 users, the system architecture would still be the same as if there were 2 million online at one time.

Bebo is completely scalable and built in such a fashion that it can handle millions of users. It doesn't matter if there are only one or two users.

Even if it doesn't have many users, the achitecture etc has to be the same. You replicate Bebo properly in less than 3 months. So you can see how I worked out the figure above.

We're currently working on a YouTube like site - this has take then guts of 3 months to develop - if this only have 1 user, the system architecture had to be good enough to handle thousands of users uploading videos, nevermind interacting with each other. When the site goes live initially, they will be lucky to have 10 concurrent users.

Another thing to point out is the "off-the-shelf" bebo and myspace clones. The most popular one being phpFox, even this commercial product is littered with a huge number of bugs that people are constantly complainging out. This system took time to develop, but it obviously wasn't planned or tested properly.

Mirror, from the start you have grossly underestimated what's involved in a project like this. €50k is far from extorsion for a bebo like site. Compare that to the €250,000 that it cost to build MyHome.ie - which system looks more complex to you?

aidan_walsh

Ziycon wrote: »

I heard that bebo was a college project that they decided to put on the net which just took off!?

Could be bebo, most likely FaceBook IIRC.

feltoncrikk

I can't believe there's so much money in it! I think I might start one myself. Actually I found this one by accident- ciorcal.com. It seems fairly new, still kinda sketchy... and empty. I joined it in the hope I might end up in cahoots with the owner, get a job, end up extremely rich. It's a long shot, but it will work! So where is the money? advertising?

EKRIUQ

feltoncrikk wrote: »

I can't believe there's so much money in it! I think I might start one myself. Actually I found this one by accident- ciorcal.com. It seems fairly new, still kinda sketchy... and empty. I joined it in the hope I might end up in cahoots with the owner, get a job, end up extremely rich. It's a long shot, but it will work! So where is the money? advertising?

Ah man your sorted!!, that's how all of us on the forum made our fortune, handy money. Your a bit behind the times but I'm sure that site will make billions.

Like theirs only 10 members 9 of theowners and you so your in there!!!, don't waste this great opportunity!!

Good luck

MiCr0

I just noticed that no one has mentioned hardware costs.
Given that a decent high availability load balancing system costs nearly 15k, all these prices seem very optimistic.
Also what about support costs? Redundancy? Co location ? Data replication?