[valdiorn]

I'd like to offer my opinion as someone who LOATHES the SaaS model.

The first and only thing you need to consider is who you're selling to (power systems engineering businesses, as you said) and what they do.

They make power systems that need to be reliable and last for fifty years or more.

You probably know banks use 30 year old COBOL software to make the world tick. Why? Because it's reliable and rarely breaks. When it breaks, they have people there to fix it. Same goes for power and manufacturing industries, they have old control systems that rarely break, because failures are very expensive.

So now you come along trying to sell them software as a service. Can you GUARANTEE that your service will be available for the next ten years? Of course you can't, you can almost guarantee that is WON'T.

So now maybe you understand why they want to BUY the product and maintain ownership of it; because that means they can manage the risk, they don't have to trust you.

And this is why I never, ever buy SaaS, because I don't trust that whoever is providing me that service is going to be there next year, or even next month (the only exception is recreation, because if Netflix shuts down tomorrow, I don't lose any value). That's why I don't use an online album to store all my photos, why I don't use SkyDrive to store all my personal documents, because I can't TRUST them.

tl;dr: SAAS == No trust (in my opinion). Either you don't trust your client (like Adobe are doing with Creative Suite) and the client can't trust you, because despite your best intentions you just can't guarantee that you'll stay in business for the next fifty years.

[jasim]

This is a valid argument. SaaS is a step in the wrong direction for use cases where continued access to the service is crucial.

I do not trust LastPass or similar third party systems with my passwords. Passwords are too important a part of my identity that I should have complete and exclusive control over it. (I use a mix of Text files+TrueCrypt+Timemachine and Dropbox for sync).

Same goes for email. I use IMAP to keep a local storage of my emails. But that is not a complete solution - my email address is still owned by a third party (Google) and they can lock me out of my identity any time they choose.

I'm currently cobbling together a couple of scripts to backup my pictures and media, mostly WORM (http://git-annex.branchable.com/backends/) data, onto multiple harddisks. The 1 TB storage of Flickr is enticing, but I'm not ready to exclusively trust a part of my identity which I want to indefinitely preserve, to yet another third party.

[sigkill]

>my email address is still owned by a third party (Google) and they can lock me out of my identity any time

They can lock you out of your past emails, but they can't take your identity away if you use Google Apps for custom domains or MSN Live Domains. You'll still have access to your email address which would probably be "[e-mail]" or something. Actually, if you make it a habit to schedule a weekly outlook download of your emails I don't think they can even lock you out of your email. I've taken steps to do the first part but I'm being lazy and not downloading all my mails onto my computer yet. [Effort vs. Payoff doesn't seem worth it yet]

[Harkins]

Gmvault + cron makes the second part pretty easy: http://gmvault.org/

[sigkill]

Holy sweet jesus batman! That is exactly what I want, and the Effort vs. Payoff has drastically reduced. I know what my computer's going to be doing for the next few nights.

Thank you!

[DenisM]

>Same goes for power and manufacturing industries, they have old control systems that rarely break, because failures are very expensive.

Out of curiosity, how much experience do you have selling to the manufacturing industry? I have no problems drumming up interest for my sales and marketing SaaS product for manufacturing industry. If you found otherwise we should compare notes.

[GFischer]

I don't know about the power and manufacturing industries, but for insurance, I have seen firsthand a company lose a multimillion dollar contract because they pushed a SaaS offering, even though their solution was very good.

[yourapostasy]

Would your position on vendor stability be mitigated if the SAAS was implemented as an appliance that can be imaged, which is placed under software escrow? If the vendor goes out of business, then you know that you will receive a discrete snapshot of what delivered the service (from the appliance, and the image can be dropped into a VM in your infrastructure), and you would receive all the code that the vendor created to implement the appliance.

You are no worse off than buying off the shelf software from closed-source companies, even large ones like Microsoft/IBM/Oracle that can be counted upon to be around in ten years. I see clients of mine get stranded by those companies all the time on just old versions of products that simply go out of support. In fact, if you get the appliance and the developed source code, you're in a better position because at least then you have the option to explore the feasibility of modifying the code to make it continue to work for your needs.

[lnguyen]

The reason those companies are still running the old versions is because even if they're out of support, they're known quantities. Implementing new versions will require development and testing to put in to place along with the possibility of additional upgrades (OS, other apps/libraries, etc.). Unless you have resources to handle that aspect in addition to any of your actual business, you're going to want to stay on a version for as long as possible instead of being on the upgrade treadmill that Microsoft/IBM/Oracle try to place you on.