I'm all for pay-as-you-go. I've seen two types of users: those who create only a few repositories but use a lot of disk space and those that have a lot of repositories, but don't use a lot of space per repository.

If you create a new repository, you also create a new blog, wiki, ticket system etc. This is worth some money. Pay for what you use.

If you use more disk space you should pay more, easy.

I think the different tiers show that you appreciate the business a user is generating. However, I think that hosting a project should have a fixed price.

I'm curious to what every one else thinks.

PAYG, charged daily, definitely. As you say, it's a differentiator if you do it that way, and that is always good.

The "by the day" pricing (with the discounts applied appropriately) although complex, will give people a lot more flexibility, and people like that (I know I would).

Finally, dealing with businesses, most will prefer the pay for what I used approach, as they always like to defer payment as long as possible ;)

Free open source hosting sounds great. But what would prevent me from open sourcing every one of my projects, and just take the risk that someone will fork it. Once my project is complete, and my client is satisfied, I can simply remove the project from Codaset again. Chances are that no-one will have forked it, and while developing I will have had a free online backup. I don't think there are many people around who just browse the open projects, and sniff through the code. You only get forked if you advertise your repo somewhere on your site or blog.

That being said, I like pay as you go. It allows me as a starter to have my costs for this service grow as my needs grow. But beyond a certain threshold, I think pay as you go will be more expensive as fixed price plans. As long as you don't use too much, pay as you go is fine. But once you have, say 75 repositories, it would probably be cheaper with a fixed plan. It's like a car: trying to build one by buying all the separate components will be a lot more expensive in the end than buying a complete model from a seller.

@Tom That is stupid and dangerous. Your client probably won't allow it for you to save $5.00/month. Also, if some one does fork code that a client paid you to develop, takes it, releases it or whatever, you can expect a huge claim from your client. It's just not worth the risk.

If you really need 75 private repositories, you need to pay for them. Right?

The other day I was looking at github to check out what they would charge me for private projects. I found that they have a complicated payment scheme. I just want to have 1 private repo (10 MB max) and share it with 1 or 2 other persons but their micro plan does not allow me to do this for less than $7/month.

And the other thing is: they seem to only allow me to pay with credit cards, but in Europe (I'm from Germany) most people don't have credit cards. When we pay online we usually do this with Paypal.

So I would appreciate if you could implement PAYG with charging for repositories, space, blog/wiki/ticket-system, contributors and I think you should make it as easy as possible for new customers to try out your service.

@Ariejan I know that is a stupid scenario, but it is a scenario nonetheless. There are always freeriders in the world. Probably not a business, but some individual trying to sell something on the side.

If you need 75 private repos, you should pay for it, that is correct. I was just trying to make a case for fixed price plans as a means to pay for larger quantities. Consider it buying in bulk.

So if you all prefer "pay as you go", then should it be pre-pay or pay for what you have used at the end of each month?

I'd go for post-pay, as that seems the only possible way to really pay for what you use. On the other hand, that again opens the door to abuse :)

My guess would be pay afterward.

I like Amazon's strategy here. You are presented with a detailed view of what you used, how long you used it and what the cost for each item is.

I don't like pre-paid. It forces me to pay for something which I may or may not use. With post-payment you can get a monthly statement and pay that.

Also, pre-payment puts up a threshold for new users to start using the service.

@Tom: Yes, but I don't think you should try to service 100% of the users. If people are so stupid to put private code in a free open project, let them. I don't care about a fixed price. If I know that for hosting 75 projects I'll loose $50-$60 a months, that's enough for me. With the pricing tiers you get a discount after all.

Another option Joel may throw in is 'reserved repositories', but maybe that would make the payment too complex. E.g. you pay $3 a month and you can create up to 5 repositories for a fraction of the cost, totalling in at $4.25 for example instead of the 'regular' $5.

@Tom: No, it doens't. The process should be like this:

1) You sign up
2) You leave your credit card or paypal details so we can capture a payment
3) You use the service as much as you want.
4) You are billed for what you used.

If I go with post-pay, then I will have to impose a minimum charge for each payment, but only if charges are due for that month. This is because of credit card and payment processing fees incurred by the bank.

For example, I couldn't accept payments of $0.10 or $0.16 as I would lose money on every one of those types of transactions. It would cost me more to take that payment, than the amount that I would be taking.

I like the $1/repo scale. That seems really reasonable.

@joel: You could impose a minimum invoice amount. If you are below the threshold your current balance is rolled over to the next month. Won't happen that often I guess, but it might.

@Ariejan: Then we risk waiting months and months for a payment.

There seems to be a lot of business sense in pre-payments.

@joel: I agree that from a business perspective, it makes more sense to go for the pre-pay. This way, you as a business are sure that you get your money. But as a client, it's not "pay as you go" anymore. I wonder how you'd do it if you used pre-pay. When I create a new repo, I get charged. But do I get 30 days charged, even if I will only use the private repo for 12 days?

For me as a client, it makes more sense to first create the repo, use it for 12 days, and then at the end of the month get charged for 12 days usage. That's true pay as you go for me.

If I were to go prepay, then it will operate on an account basis. So you would fill up your account with any amount you want (although will most likely be a minimum), then at the end of each month, we calculate how much you owe based on what you have used, then deduct that amount from your pre-paid account.

I think you've made things complicated just to use the charge per project model. Tier pricing? Groups? Charge per day? How to tally projects? This is all solved if you charge by space. In fact, $1 per project with 1gb of space will probably drive your business broke if people use up all the space (esp. if their versioning images/graphics).

You might need to think of space in another perspective because charging by space is similar to charging by project and probably more profitable as well. Space limits the amount of projects, but this limit is flexible for the customer and easier to understand.

For example, how many projects can you fit in 100mb? Probably 1-2 proper projects. Hobby developers can probably fit more but you shouldn't be overcharging them that much in the first place.

If you charge $1 per 100mb, this is almost similar to charging $1 per project. However the benefit here is the customer decides how to use it. You don't have problems with forking or using tier pricing for having many projects. The business also profits, since projects always grow in size.

Charge $5 minimum a month for 500mb if you pay monthly.
Charge $50 a year if you pay annually (discount because you save on transaction costs)
Then allow $1 per 100mb up to a total of 1gb. After 1gb, charge $0.50 per 100mb.
When they reach their quota, ask them to prepay for more space in their next invoice.

Check your repos and find out the amount of space used per project on average. Chart the amount of activity against the repo sizes. See if you can tweak the above and make a profit.

@Alexander: All fair points, but I think I have said this before, that I am not selling disk space. The value proposition here is the Codaset service itself, and the features that it offers you, which is why I think a per project fee is fairer.

Also, everyone knows what a project is, and what it contains. There is no unknown quantity with a project. This is not the same if charging on disk usage alone. A lot of users would not know how much space they use or may need, so by charging them for each project instead, they always know how much they will need to pay at any time.

It's much easier to estimate and calculate costs based on the project, rather than the disk usage.

As you know, I think that charging by project is a huge mistake, for all kinds of reasons, and, I believe, will simply handicap your service. All the social, tracking and management tools certainly differentiate your service, and you can charge more for them, but the differentiation is lost, in my opinion, if you charge for them on a per-project level. There are precedents: Google charges for ads, not for mail, but gmail is the differentiator; in the same way, you could charge for storage and the excellent wiki, bug tracker, blog tools, etc., are the differentiators.

I think you should go the "pay as you go" route and charge by overall storage for all private repos together. It's easy to understand, predictable for me, the customer, and flexible. Customers that don't need lots of projects won't have a problem and customers that do won't be hampered.

Charge ahead of time, on a monthly basis. I think it makes good business sense to require the customer to commit some kind of minimum to using the service. This also virtually eliminates non-payment issues. Let customers pay a year or two in advance as well.

Code hosting is not quite the same as cloud hosting, despite the similarities. Cloud hosting is much more "abstract", so paying in advance probably poses all kinds of tricky business issues. Code hosting is pretty specific, has a generally rising storage requirement and serves a very specific business and technical need.

To remove the entry barrier, you can give a free private repo (as you suggested elsewhere) or a free month or a free 100 MB...

Bottom line: I say, PAYG, by storage per month, in advance.

The per-project basis is the best one IMO. As for minimum payments, what is the lowest amount that is feasible to collect? Let's say $3.

It wouldn't be a huge deal to collect $3 from a developer with one repo every 3 months, that would be the minimum. The pay-per-hour basis is kinda cool, but maybe implement a minimum fee of $1 to create a new project, so there is no overhead for someone creating and destroying projects ad-nauseum.

jhnsntmthy is right.

just invoice all of your customers every three month.
if they added new projects in these 3 month, charge for these post-pay.
for existing projects charge pre-pay for the next 3 month.

"Pay as you go" would make it hard to calculate a budget ahead of time. If someone does something stupid or generates a massive amount of data in one day, the cost for the project could instantly become far higher than you were planning.

On the other side, this model would most likely give a lower end cost to users.

I've not used a model like this before and would certainly be interested to try it out. Despite that above detriment, it may be the most sound model for Codaset.

Invoicing every 3 months would not be good for business, as that leaves it open for abuse. Users would get up to 3 months of free and unlimited usage, without having to pay a penny. It would have to be monthly billing at the minimum.

I was only suggesting the "every 3 months" issue for the rare case where someone owes you less than $3. I would also like the idea of paying upfront into an account, and having the account deducted from when I create a new project. Having a $5 entrance fee, plus automatic renewals would be a good way to go.

If you are worried about complexity, i propose to make a bill-preview page where you can play with the figures.

I mean something where i could go someplace at "My Account" and then see what the next bill would look like with current usage limits, and where i can let it recalculate with, say for example, one deleted project and 300 MB more disk usage (or whatever you decide to charge).

For cost control in a pay-per-use scheme, i would also like the ability to set at least warnings/notifications or hard-limits on chargeable dimensions. Or, simpler, having an account which has to be precharged, and if used up /emptied, then my projects switch to read-only. Basically the same as prepaid-mobile-phone cards, which people are accustomed to (and which is btw my favorite scheme for this).

Also i would suggest to have the one freebie closed source project you mentioned be limited to only the owners access. If you want other people to have access to it, you have to go OS or pay for it. I think that is fair, as it allows to have your pet project, but gives incentive to start paying as it grows.

Guys,

This payment per project model makes EXTREMELY POOR business sense to me.

I am an independent developer who would like to create innumerable number of small projects.

Do you realize that at the end of the day, this service is just hosting totally open source SCM with some open source bug tracking software, and on top of an an open source OS?

Suppose I want to create say 15 private projects, does that mean I would have to pay Codaset $15 per month?

Do you know that for that much money ( $ 15 p.m. ) you could get a Virtual Private Server running your favorite distribution of Linux and with just the help of 'gitolite' or 'gotosis' utilities you could get your own totally private setup of Git hosting (plus your favorite bug tracker) and have more than 10 GB of private space?

On top of that you could even host your own website and everything else you want.

Of course that would take some bit of Linux admin skills to get that going, but once you have that, you could be the master of your own private code sharing site.

I HATE the idea of any kind of web host who thinks they can charge for 'virtual units of content' (Git projects and/or users in this case), while they are actually only paying for the raw disk space , and the raw bandwidth for sharing the content.

And they talk about preventing 'freeriders' while they themselves ARE the freeriders on the system!

I am hosting my private projects on RepositoryHosting.com for this exact reason. They charge about $5 (or $6) per month flat rage for 2GB disk space, UNLIMITED repositories, and UNLIMITED number of users and contributors, and UNLIMITED bandwidth.

The only problem with RepositoryHosting.com is that their UI and bug-tracking system (Trac) sucks. Otherwise the Git stuff works works exactly the way you work with GitHub.

In summary , anyone who would be paying more than US $15 per month to Codaset is either an idiot or has a billionaire father. Not to mention that services like these are there to milk the gullible customers who don't want to do their own research before signing up for overly expensive services.

-Rajesh

Wow! Thanks for that Rajesh. Let me please explain again as to what Codaset is and is not.

Codaset is not simply a code hosting service. It is so much more than that, and will continue to be more than that. Yes, Git is central to what Codaset is all about, but it is not what I will be charging for. The value in using Codaset is it's feature set. Things like tickets, blog, wiki, user permissions, etc. These are features that complement the source code hosting part of the service, and it is these features - and more - that I hope gives Codaset it's value.

Do you realize that at the end of the day, this service is just hosting totally open source SCM with some open source bug tracking software, and on top of an an open source OS?

Not true! Yes, some technologies that Codaset relies on are open source, such as Git, Nginx, Ubuntu, Rails and a few other bits and bobs, but no more than any other web site. The site itself that you are using right now is proprietory software. I built all the code solely for Codaset, including the bug tracking (which is not - as you say - open source).

Of course, you could all go out and get yourself a VPS from your favourite hosting company, and install everything you need on it. But if that was such an attractive proposition, surely we would all be doing it?!

I HATE the idea of any kind of web host who thinks they can charge for 'virtual units of content' (Git projects and/or users in this case), while they are actually only paying for the raw disk space , and the raw bandwidth for sharing the content.

I will not be charging for 'virtual units of content'. I am providing a service, that will never be 100% tangible, and can not always be measured. So you will not only be paying for disk space and bandwidth.

The only problem with RepositoryHosting.com is that their UI and bug-tracking system (Trac) sucks.

Well then I would see that as a good a reason as any to try out Codaset, or any other similar service.

I won't go on much further, other than to repeat again, that Codaset is not and will not be charging it's customers based solely on disk usage or bandwidth. If that were all I offered, then I would be a normal everyday web host. But obviously that is not true, as this site and what you can do on it is testament to that. Codaset is a project management application for open source and private software; not a web host!

At the end of the day, we all place a value on something based our own experiences and requirements, and I completely respect that, and appreciate your comments.

Anyone else feel the same?

I actually agree with Rajesh. Why not charge based no how much disk space we use in total? It's not like it really costs more if I'm hosting 10MB spread across two different repos instead of 10 MB in one repo, right? If it does, then you're doing something wrong.

Again, I am not charging for disk space, as Codaset is NOT a web host, and is NOT offering web space for sale. I will be charging per project, because that is where the value lies. [calm down Joel] ;)

I think Rajesh is off-base for the most part. There is more to just the raw host costs in running a business. Personnel are required for on-going maintenance, support, etc. At the same time, I use Codaset mostly just for the code browser. I use separate software or services for issue tracking, wiki, chat, and blogging and have no interest in moving any of that to codaset.

So, the issue I really have with the pricing model is that at the company level, the pricing makes a lot of sense (we have < 5 repos and will unlikely grow beyond that without revenue growth), but at the individual level, where I'm apt to make many more repositories, it makes no sense at all (more repos = more cost, but I'm making no revenue). It's the same deal with GitHub. It's also why I do run a VPS with gitweb and gitauth. It's not an ideal solution for me, but I can't justify $40+ USD / month on tinkering projects.

But, I've just had to come to terms with not being the target customer for these types of services.

At the same time, I use Codaset mostly just for the code browser. I use separate software or services for issue tracking, wiki, chat, and blogging and have no interest in moving any of that to codaset.

Yeah, I used to use several different services for managing my projects and code just like you do. And that was the main reason why I started Codaset in the first place. (again, another value add).

But, I've just had to come to terms with not being the target customer for these types of services.

Exactly! I will never be able to please everyone, and I'm not trying to either.

Anyone can install or build a project management tool. Most are relatively stable and come with most, if not all, the features we need. And if we install it ourselves, we can dev for it! What a great option.

What you forget is that this means you have the resources to expend for maintenance (so you have either a dedicated tools guy or someone who has their workflow interrupted to hash out anything stupid happening with your system) as well as the experience with said system. So if you do not have Sysadmin abilities, or you have ACTUAL work to do, then this is not a feasible thing to do.

Rajesh, I suppose you don't pay for Basecamp or a private Github account, or even a hosting service at all? You just host out of your own space, and colocate (maybe you own your own datacenter and produce your own electricity?). A service is here to provide you with the service. Maybe you can build your own project management system (it's a difficult problem. I know, as I am building one targeted to my team's needs). But you still need to build it. A freelancer has better things to spend their time on, like making money. Not supporting their own version of Basecamp for team interaction, or making sure their project management tool works. You don't have much extra time to make sure your development server works as well as it could, or that the datacenter has paid it's bills. You pay for these services because they allow you to make MORE money.

Be reasonable. If you have 75 client projects or similar, you probably aren't making $75 dollars. In fact, if you are doing it correctly, the Codaset fee for all your projects for a year could be paid by what you charge 1 client. Or you could just do half as much work and maintain your own version of Codaset.

At the end of the day, Codaset offers you a hosted, SOCIAL service, ready-made, already installed, with a pretty much dedicated support team. You can't do that. And if you are, you are spending more than a theoretical $75 dollars in terms of monthly maintenance, any necessary design, and the headache. Have fun with that.

/rant

I applaud Codaset and their opening up of the business model to it's users. Personally, I'd rather Codaset do the following for pricing:

Minimum $5 account, always gives you 5 repos and all that jazz.
Above $5 account, it's pay as you go (pay X amount for Y repos, according to the scaled payment scheme outlined in the Blog Post)

This way, Codaset can keep healthy in terms of operational costs and simplifies the fees for small outfits. Larger outfits can then decide what resources they need, but everyone knows that they are paying for AT LEAST x amount of resources. Just my thoughts on the issue though :)

We wholeheartedly agree with you Joel, your service is already as useful as github's, if maybe not as popular. If I want a real private repo, I already use my own VPS to host those projects. I just use codaset when I want the sharing features and issue tracking and everything else involved, it is happily worth the value.

Joel,

It's your product and it's your business. You don't need anyone's approval for whatever business model you choose. There are enough potential customers out there for whom your proposed business model is excellent even if they don't use CodaSet as anything more than a git host. I think this is obvious. In any case, the market will decide, and I'm betting CodaSet is going to make a very nice profit. If it were public, I'd buy stock. Seriously.

But even though you can't satisfy everyone (this is also obvious), you are asking the community at large to weigh in on your proposed business model. And amongst those responses that do not like your proposal, there seems to be one common thread (I've raised it myself, as well): this business model neglects what I believe to be a significant, mostly untapped, market.

Freelancers and other small development shops simply cannot be billed or constrained by the number of projects. This is a business constraint (for many reasons). That does not mean that we do not place value on all the features that differentiate CodaSet from the competition. On the contrary: we need all of the them because we don't have the resources, the time or the people that can do all of these things for us. We need the social tools to communicate with our customers (and their customers, not to mention their other freelance specialists). We need the integrated service CodaSet provides, but your business model, quite simply, puts them out of our reach.

Now I believe that there are many many freelancers and tiny development shops out there that share this business constraint. So if you find my belief in the size of this group to be unfounded, you can dismiss this market, this loosely banded segment of the community, and we will sadly take our business elsewhere. Or you can listen to this community and revise your business model to target us as well.

Obviously, it's your call.

Jose,

I can operate my own system for far less than $75 a month. In fact, it costs me effectively nothing. I just don't like it. I like Codaset. I'm also not a contractor/freelancer/consultant/whatever people call themselves these days. I create a git repo when I want to work on something because it's easy to do and it's the git way. I am a professional software engineer, but one of my personal hobbies is also software. So, I charge $0 per project and I make $0 per project (lets please ignore any hypothetical value I'm adding by skills picked up during any project). Also, for these projects, I have no other collaborators.

This use case -- the hobbyist -- is not suited to the pricing model. I've accepted it and am even okay with it (I've talked with Joel on IRC a fair bit); I mean, it sucks, but so be it. But, belittling people because they don't have billables associated with a repository is really not appropriate.

First of all, thanks to all of you for your comments and feedback, no matter what that may be; positive or negative. They all help with making decisions about my business, and improving the service to you guys.

@Noam: I'm not sure exactly how I am dismissing the freelancer/small developer market with my chosen business model? Would like to know more about what you mean, because even if I change the business model to the traditional selection set plans, you will still be billed or "constrained" by the number of projects. As each plan will have a limit on the number of private projects you can create.

@Nirvdrum: If you code as a hobby, then you should never have to pay for Codaset, as you can create as many public projects as you want.

@Joel Moss,

Well, hobby doesn't mean open source or public. I really don't want to go on a diatribe there. But, I only release something publicly once I've done everything I should: proper license file, proper copyright messages in headers, cleaned up code, good documentation, etc. Usually I don't bother with proper licensing until after the project seems to have some value. I'm sure others would argue with that approach, but I'm pretty active in the OSS world and feel it's the best way to go.

In any event, I'm not going to publicly disclose my dot files, my /etc files, my old class projects, etc. And if I need to run a local server for that anyway, there's little point in moving elsewhere. And as much as I like codaset, almost all my open source work goes on a GitHub because that's just where I started and I have followers over there, etc. I do like codaset's tools better, but that's a lot of friction to overcome.

I am absolutely shocked how many apologists and trolls Joel Moss has to deride my arguments.

@Joel: Codaset is on the WRONG BUSINESS PATH. Why don't you see that from a developers point of view? Maybe you have been able to fool yourself that listening to these MBA types preaching you the "Business Model" boool-sheeet is right for you?

Let me as you again:

1. Why is it so hard for you to understand that you are NOT providing any Out of this World (TM) service. All you are providing is just a place to store git repositories and an integrated bug/issue/wiki tracker + browser. Doing this does NOT take too much time or energy. In fact all you need is a virtual machine or VPS where you could install Git/Gitolite/Gitosis for handling Git repositories + Redmine or Trac hosted under Apache/Lighttpd/Nginx. With either of Trac or Redmine installed you get all of the "features" and "services" you claim to provide (Wiki/bugtracking/reporting/source-browsing/forums/news/RSS etc.). Want more exotic content management ? Just install RadiantCMS alongwith Redmine! And you are DONE!

2. I don't see why someone should even bother to pay a single dollar to Codaset when services that provide ALL of the features that Joel is claiming to provide is available on an UNLIMITED basis from competing providers ( e.g. RepositoryHosting.com)?

(FULL DISCLOSURE: I am a customer of RepositoryHosting.com but looking for a change (because I dont' like their Trac interface I have mentioned before). Other than that I nave no other affiliation with them.)

So Joel, please stop telling yourself the same lies that your apologists keep telling you;

please stop believing in the myth that you are starting a .COM and this is year 1999; and

please start looking at the competition and the markets who are looking for better value for LOWER price (and NO, your per project pricing is NOT A BETTER VALUE PROPOSITION for customers).

I know that this goes against your foolish belief that you can charge per-project with some of the "value-added" services you are bundling and make a zillion dollars out of this business , but you and your venture fund providers (if any) are soon going to be VERY disappointed.

Plesase WAKE UP before you go too far down the drain!

@Joel: The problem with your business model is that small/individual developers aren't going to use your service to store their personal private projects with Codaset with your current pay-per-repostiory business model. Rather than saying "I don't want your business! I can't please everyone!", why not rethink it? I would switch to Codaset in a heartbeat if you did something to make it more friendly to individual developers that want a ton of private repositories for their personal projects that aren't public yet. I don't really understand why giving users a set amount of space that they can use for however many repositories they want is a bad thing. That's something that would actually be really useful for me, and it would work for larger developers too because they could just purchase a plan with more space. It just means that users get more flexibility for how they use their account; it's not some badge of shame that will turn Codaset into like http://1dollar-hosting.com/! It'll be a kick-ass code hosting service that's accessible and useful to every single type of developer, whether they need 4 repositories that are each 2GB, or 15 repositories that are 0.25 MB!

c'mon Rajesh, it is one thing to want a company to offer you what you feel you want, and another thing to build the business that has the potential to attract attention. The reason you aren't happy with your current company is because they are selling disk space, and aren't building a tool or community around the code (they also aren't giving OSS developers any hosting for free either).

That said, any really great disruptive business is usually going to do more than 1-up the competition, but rather throw open the doors to do it better and cheaper than before.

@nirvdum Part of having a git repo means you don't need to have it up somewhere. It can stay local on your machine. If it really is a hobby, I doubt you need the issue-tracking/wiki/blogs etc. at that time. Personally, I have a CodebaseHQ (http://www.codebasehq.com) account where a project (in my case, stuffs) can be private and can have infinite repositories. They even have a free version of their app. Try it.

The point of Codaset (and Github/Bitbucket/whatever for that matter) is that it is social coding. If you are a hobbyist and don't wish to release yet/ever, just push to your own private server. You don't need ticket tracking or social coding. If you are a freelancer, you benefit by having all the tools you need here (maybe not time tracking, but I'm hoping you get that once it goes all profit-y). If you don't need that, then you don't need Redmine either. You just need to host your code. And you can just setup a private server to do that (I do it, it's not hard). But this is convenience.

If its a question of Github or Codaset for open source code, use whichever you like. They are basically the same, baring the obvious differences in style (where are the Codaset graphs!). And it will be free for Open Source stuff. If you want to live on the edge and host your client work in a free acount, YMMV. I don't think Codaset/Github will pick up the bill if your client ever sues. And if you are wondering whether or not you should use one or the other for client work, well, thats up to you. No one is forcing you to use either. But you can forgo the hassle of setting anything up by simply signing up for a service (like repositoryhosting.com if you have a hard-on for trac, or codebasehq.com or any other innumerable number of services).

And if anyone thinks I am apologizing for Codaset, I personally hate it's style. I'd rather use CodebaseHQ. But that doesn't work for my developer team's workflow. And Github would get too expensive at our volume. Trac sucks and we do php work, so it's unlikely I'll install Redmine. In short, Codaset is the worst project management tool, except for all the other ones.

@jhnsntmthy

The reason you aren't happy with your current company is

because they are selling disk space

WRONG! I am quite satisfied with the service I got from my current repository host. But being satisfied does not mean being happy. Yes I am looking for happiness (aren't we all? its even in the US Declaration of Independence), which is why I want to let the newer providers know what I want and how they can get me to be THEIR customer instead of someone else's.

It is my job as a potential customer to let the providers know what I from a code hosting (or any other) service. But it is then upto them if they want to provide the kind of service or not.

and aren't building a tool or community around the code (they

also aren't giving OSS developers any hosting for free either).

WRONG AGAIN: Most people pay for hosted code repository services because they want to host THEIR own code, with a small number of THEIR own developers. They are not coming to Codaset (or any other paid repository hosting) to "build a community around their code". If Codaset people believe these BS theories, then they haven't done enough market research and analysis. So they SHOULD NOT be in this business.

If I want to "build a community" around my code, I would do that on GitHub or numerous other OSS community + code-hosting providers. (FYI: In fact I have a number of projects hosted on GitHub where I would gladly accept inputs from the "community"). But despite taht I have chosen to subscribe to a paid Git repository hosting provider BECAUSE I have a lot of code that I CANNOT share with others (due to personal reasons).

That said, any really great disruptive business is usually going

to do more than 1-up the competition,

Yes. But 1-upping the competition is VERY DIFFERENT from ignoring the conpetition and potential customers altogether, which is what Codaset is on the path of doing.

but rather throw open the doors to do it better and cheaper

than before.

Human civilization makes progress whenever someone makes a widget or provides a service that is both CHEAPER and BETTER than what is already available to people. So there is nothign wrong with attempting to 1-up the competition while providing more features than the competion , and charging less money than the competition.

IMHO this "disruptive" business model of charging per project and/or per contributor/project STINKS of relentless and unreasonable GREED. Nothing more , nothing less.

Before someone starts lecturing me on greed, yes, I know the difference between good greed and bad greed. The business model that Joel is attempting to adopt for Codaset is bad greed, if you ask me.

I ultimately do agree with most of your points, my only position is that $1/mo for each private repo is miles BETTER than what I can get out there now from anyone else, and a better feature set too. So I would be HAPPIER on Codaset than any of the other mentioned sites.

Don't lecture me on market research, I am positive that you and your friends's opinions do not count as true "market research". This thread is all about people with different needs expressing their views on value. I have different ones than you. And Joel gets to pick which customer type he would rather serve. He may be going after numbers, profit/customer, or prestige, but it is his choice... We will vote with our money.

A more accurate measure of costs is probably bandwidth. Or that combined with disk usage.
But these are quantities that are for the most part unknown for a developer creating or hosting a project remotely. How often are you going to push, how many people will pull? These are unknowns until you establish a userbase.

I think per-project pricing probably makes the most amount of sense given all factors associated.
If people really can't afford the huge sum of $15 a month for their 15 projects, two things could happen:
1) Get a better sales person / pitch. If you cant afford $1 per month for hosting of your project code on an external system, how are you affording the development time, or the essentials of life, for that matter.
2) Maybe Joel will offer reduced bulk pricing in brackets for those that have a bajillion projects they need hosted.

Each to their own! We all have our own opinions, and we are completely entitled to that, and I for one, appreciate that immensely. But please remember that your opinion may not be right or wrong, or not what every other person thinks.

At the end of the day, I have to make a decision that I think is good for business, and most people will want. I know full well that some of you will not like or agree with any decision I make. And that is why running a business is so hard, and so rewarding. I've been there before, after having run a successful hosting biz for 7 years.

@Rajesh: You still don't seem to understand the concept of creating value out of a business. What is valuable to one person, may not be so to another. We can all go and build our own Codaset, or Github, or RepositoryHosting.com, etc. without any problem at all, but we don't all do that! And that could be because of a hundred different reasons. But I did do that, as I felt the current offerings were not good enough for me, and did not offer what I wanted. And some of the best products and services have been born out of that very same desire. I guarantee you that if you want or need something, there are also others feel the same.

They are not coming to Codaset (or any other paid repository hosting) to "build a community around their code".

Not true at all! I used to use Github [and now use Codaset] because it is a social tool, and I wanted to build a community around my open source projects. At the end of the day, that is essential to the definition of open source; it's all about the community.

@Rajesh: I think this is simply a matter of your reasons for using a code hosting service, differ to mine and others here. But obviously you are not alone. We all have different reasons for using a service.

Again, thank you to all of you for your comments. Please keep them coming. I am taking them all on board, and considering them all.

Jose,

I think we're on the same page now. Part of the point I was trying to make is I don't like having my OSS in one place (GitHub), my personal stuff another place (my prgmr instance), and company stuff some place else (presumably codaset). If I could consolidate everything to one provider that would be ideal. But, it's not looking like that's going to happen. So be it.

I agree mostly with cmelbye. But Joel's plan seems pretty well set now, so what this means is there is an opportunity for someone else who has the motivation, time, money, etc... to create a different kind of codaset service that caters to this type of user... someone like me... who has needs like this:
* Want to create a LOT of projects, mostly small and shortlived, but want to keep history of sourcecode and issues etc... online in a private space that is robust (backed up, accessible, etc...) * Comfortable paying based on actual usage (bandwidth, storage, etc...) * Want the interface of the system to be clean and easy to use * Want an API to make getting bugs in/out of the system easy/possible. * No crazy issues (like system wide id's being used for my project's ticket id's)

Joel,

I apologize for the length of my answer. I invested some thought in how to reply to your question, and felt the detailed answer was the only way to really demonstrate what I (and I think a bunch of us) have been trying to say. I know one of your concerns is value, and my reply goes to that as well.

You asked me:

I'm not sure exactly how I am dismissing the freelancer/small developer market with my chosen business model?

I'll give you just one example of a very common scenario you simply don't support with your chosen business model. I built a very small application in about 20 billable hours for a customer. This project benefits from bug tracking and some documentation. The project lay dormant for more than a year before the customer called back asking for improvements, which took another 20 billable hours.

So I have this business project which has been inactive for the greater part of 2 years. You could say that at the worst case this translates to $24 total for hosting, which really doesn't sound so bad.

Except it's for just one mostly inactive project. I've got lots of these tiny projects, some of which are still dormant, others eventually see repeat business and a few are on-going.

If I only had 20 such projects, totalling less than 2.5 GB of disk space, you'd be asking me to shell out $240 per year. That's almost 10 times what ProjectLocker charges, and it also gives me a wiki and a bug tracker, not to mention unlimited repositories.

Your chosen business model also makes my expenses less predictable and incurs additional expenses for completed projects (bought and paid for) post-production, which dips into my profit margin. I can't just dump the code because I can never know how long a project will be "alive" and you're penalizing me for this.

For various legal, technical and political reasons, I can't even pass this cost off to my customers: it's a permanent, increasing, operating expense for me. But I can't charge my customers for indefinite hosting for another critical reason: this is part of the service I provide, part of the value proposition I offer my customers because I want their repeat business. This is as close to "passive income" as I can get on freelance projects.

I do want better tools and I do need better integration and I also like CodaSet's look and feel. In many cases, social tools of various kinds have helped in my closed source projects. These things have a lot of value for me and for (most of) my customers. I'm willing to pay more than I pay ProjectLocker to get them because I can demonstrate ROI. But your chosen business model is clearly a losing proposition for me.

You also asked me:

Would like to know more about what you mean, because even if I change the business model to the traditional selection set plans, you will still be billed or "constrained" by the number of projects. As each plan will have a limit on the number of private projects you can create.

If you hypothetically offered the same plans ProjectLocker offers, a lot of freelancers would move to CodaSet in a heartbeat. Not because of unlimited repositories, and not because disk space is the measure of quality or indicator of value, but because, all things being equal, CodaSet is better.

I'm not telling you how to run your business or how to structure your plans. I'm telling you, because you asked, that if you adopt the model you proposed, a lot of potential customers like me will unfortunately walk away.

I'm also not suggesting you take the "traditional" route (assuming you mean "github"). In fact, I've suggested before that there are quite a few options for alternate plans (e.g., based on disk usage, bandwidth, feature-set, active/passive projects, and any combination thereof, to name just a few). My favorite is free and ad-based, but that's really a pipe-dream.

My second favorite, because it makes a lot of sense to me, is to charge by disk space. Both ProjectLocker and RepositoryHosting provide unlimited repositories and constrain disk space. The fact that the meter is based on disk space does not mean that I'm buying disk space or that their service is somehow less sublime that CodaSet. I'm paying for a code hosting service that offers many of the features that you offer as well. They're less cohesive and integrated, they don't look as nice, but they target me and they offer me value, measurable business value.

Repositories are a basic (and usually free) commodity for programmers like me. From my point of view, by charging per project per month, you will inflate the price of this basic commodity and de-value your service too much. It's like charging $10 for a single potato. By not charging per project, by making the basic commodity freely available, you will place value on all the other services. The high-quality features you offer - your wiki, blog, tracker, API, integration, graphs, look-and-feel, permissions, and so forth - will all acquire value and differentiate you from the competition.

@Noam: Thank you for your comments. They really do help me.

First point I want to make, is that Codaset is not the same as services such as ProjectLocker or RepositoryHosting. These services are simply hosted open source software, and simply host popular, freely available software for you. Usually Trac. While there is absolutely nothing wrong with that, and a lot of people are very happy with Trac, etc., I place such a service as less valuable or useful than a proprietory service such as Codaset, or Unfuddle, or Github. So although they are competition, I would not compare Codaset directly to such a service.

The reason why open source software hosting services - such as ProjectLocker - charge my disk space, and provide unlimited projects, is usually because disk usage and bandwidth are their main costs. So it makes sense for the business model used. They have very little development costs because they are using open source software developed by the community.

@Noam: How much space do you currently use on ProjectLocker and how many projects do you have? Are they all private?

Repositories are a basic (and usually free) commodity for programmers like me.

Yes you are right. But as already mentioned, I am not simply hosting code repositories. I want to get away from that with Codaset, and want to add so much more than that "basic commodity". I'm offering, or intend to offer some genuinely useful features for every project, that include a lot more than just code hosting. And that is why I will be charging on a project basis.

Again, I thank you for your feedback. I am actually toying with the idea of enhancing the pay as you go model with some sort of basic and advanced plan that could sandwich the pay as you go pricing, but I will not be charging for disk usage for the reasons I mention above.

Joel, I understand what you're saying about disk space and I won't urge you more than I (and others) already have to go down that route.

What I'm trying to say is a bit different though. Whatever meter you use to charge for your service will not be the factor people use to measure value.

Take restaurants as an example. Obviously the basic commodity is the food. While the price has some correlation to the cost of the "raw materials", the fact is that the price covers other aspects as well, including the salaries, rent, special knives, atmosphere, mending chairs and tables, etc.

So you could say that ProjectLocker is like a fast-food restaurant (cheaper operating expenses, lower salaries, less atmosphere, smaller ratio of price to cost) and CodaSet like a gourmet restaurant. I'd accept that comparison; quality comes with a price.

But all of that doesn't necessitate charging by the project. No restaurant I know charges per seat. The fee, the number customers see, is always for the meal, though there may be a Minamoto charge per person, and the differentiator is all of those aspects that give the restaurant value.

So choose whatever meter you want; it's unquestionably your prerogative. For many developers, there isn't any real distinction between the "seat" and the "meal" so the price comes out the same either way. But for me, this is an unfeasible proposition.

@Noam: How much space do you currently use on ProjectLocker and how many projects do you have? Are they all private?

Right now, I use about 300 MB on ProjectLocker for 7 projects. Two of these projects take the lion's share. But I only started to use a repository host about a year ago. I have close to 50 repositories for billable projects that I host myself (in Subversion) on my server. About 4 of them are currently active. I have a few more personal projects I'm tinkering with that are not for customers. I used an open source bug tracker (not trac), also hosted locally, and various solutions for documentation. I estimate roughly that I use about 2 GB of space all in all, not including the artifacts.

I've only migrated projects that saw repeat business or were ongoing at the time, but I will be migrating the rest of them to git eventually. I almost lost most of my repositories because of hardware failure and thankfully one of my backups worked. So now I'm making everything more redundant, and working with Subversion offline is impractical at the very least.

All my repositories are private, though I am considering making one or two public. I do participate a bit on open source projects, but none of them are mine.

After reading all of these comments. It's quite a bargain for a company to use Codaset, probably.
But, for individual developers, it's just much more feasible and flexible to either host their own code, set up their own system, or going open source.

I will probably only use Codaset for paid client/company projects or open source. Not for my personal and private code.

How about this for an idea? I don't plan on limiting the number of users or collaborators on any project, but perhaps as a way of appealing to individual developers who are the only user of a project, I could create a sole developer pricing or plan, which would be cheaper to run.

Another options could be to allow you to archive a project, which would simply make it inactive. WIll need to define exactly what that means, but there would be no charge for archived/inactive projects, and you can re-activate it any time.

Thoughts?

I like the pay as you go model, and would be willing to pay $1/project and include 1GB of space. Get an additional GB of disk space for $1/month. I think this would allow for "unlimited" repositories per project.

Unlimited users for either public or private is nice as well. The ability to achive a project (read-only mode) is a feature that most other providers offer, although it does go against the pay as you go model, since you are still using space.

Do you have any idea of when you are planning to go live? I just joined a new company and we are evaluating providers. I'd like to include codaset in the evaluation, but without pricing set, it's hard to compare apples to apples.

I hope to take Codaset out of beta by the end of the week. Full pricing will be announced on this blog in the next few days.

Joel,

I like all those alternatives for a single developer. I just hope it makes business sense for you to do so.

Joel,

Any plan that greets freelancers with open arms is welcome!

A reduced price for individuals sounds good, but the devil is in the details. I do think archiving projects is a great idea and a terrific compromise.

I have one reservation though (for both ideas):

Please distinguish between the developer and other users. The individual developer should be the only one with permissions to access the repository, obviously. But it should be possible to let other users participate on the project's other tools without "counting" as additional programmers. My clients have to be able to add tickets, modify milestones, and so forth.

The same goes for archival. Archival should not affect the ticket system. Customers are often hesitant and tentative before giving the go-ahead on any work, including repeat work. They also tend to wait until there's "enough" work to justify the expense. Allowing a customer to create tickets without having to commit to work would lower that barrier and give them a clearer view of the backlog.

Joel

Every now and again something comes along that's a real game changer. Love what your doing and your approach. Big fan. Watching closely.

The diddypiggy collaborative.

nice ;)

Why not both? I mean what would be the problem to allow both ways? Some likes flexibility and some ppl like "flatrates". Thats why most of us use internet flatrates, cause people like that way :)

so why not create a "flexible plan"

like you posted. And a flatrate per repo? something like 4 or 5 Dollars user. No support at all. And if people like support you could add another 2 oder 3 dollars. Well and a master plan support for 20 Dollar per month ( Just an idea :) )