SaaS v’s desktop apps: Celebrity death-match or horses-for-courses?

I was chatting to one of the founders of salesforce.com a few days ago about the merits of software delivered as a service over the web, negating the need for locally installed apps. I’m talking about the concept of SaaS (Software as a Service), and the reason the conversation came up is because izimi (my company’s software) is most definitely not SaaS. What is the way of the future, we debated.

Now when you’re talking to someone who has built one of the world’s most successful and fast growing software companies on a SaaS architecture you’d better listen, those words are not to be taken lightly. So how could I resolve this conflict, is there just one way that will win out, or is it more complex than that?

My view is that, for some applications, vendors will find themselves adopting the SaaS route. You’ll see as you read on that I think it’s exactly the route to go for some. On the other hand, I think we must be cautious not to automatically assume this works for all classes of application – you have to consider the commercial and technical drivers that are in play, as they both have a bearing.

The conversation went back and forwards for a few days afterwards by email, and this is the realization I came to. I’ll use the two real examples which were naturally the subject of our discussions: salesforce.com and izimi. Actually it’s a bit embarrassing putting izimi in the same sentence as salesforce.com – the former is huge, with hundreds of thousands of corporate users and a current market cap north of $5bn. The latter is an 18 person startup and certainly worth a little less at this stage. Nonetheless, they serve as useful examples of two different types of software application:

Salesforce.com
The market they’re in is CRM, enterprise software, large databases, and complex business rules/logic – traditionally speaking a potential management and deployment issue for any organization. Therefore SaaS (salesforce.com’s approach) is great here. What MIS manager wouldn’t see the benefit of outsourcing the provision of the software/service to someone else who will manage all of the deployment, development, upgrade, support, and backup issues? One of the biggest problems with the pre-SaaS client/server enterprise software is deployment and management at the users’ desktops – that’s just one of the problems and yet on its own it’s a major headache.

Enterprise software users are also generally considerate of their total cost of ownership when looking at new business apps. When they are investing in new client/server apps they have to consider an (often considerable) license fee, support costs, training… Saleforce.com takes all those financial concerns away, and avoids the upfront capital investment in software (which often is a pain to deploy, doesn’t work as it should, and is soon out of date).

Also worth noting is that many organizations that deploy client-server enterprise apps (as opposed to the SaaS approach) end up getting drawn into so much customization and integration that they subsequently require a good deal of bespoke consulting, design, and integration just to get and keep their apps running. All the major ERP and CRM vendors have huge services businesses to support those customers and it’s a big part of their business. Often the economic promises of integrating client-server business apps with other internal system have proved to have been outweighed by the cost/time penalty of doing the work required to get there. SaaS software like Saleforce.com’s products let organizations get running quickly without that big MIS overhead, PLUS with APIs like Apex those that do want to do the integration have a means to do so if they wish.

izimi
Something like izimi is a totally different product to the previous example. It’s a different application, different class of users, and different factors governing which architecture best suits. Firstly it’s not an enterprise product, so we’re not talking about MIS managers overcoming a bunch of problems or financial concerns to make a buying decision. It’s for individuals (though could be used by groups in companies) who want to instantly share any type of digital assets without any restrictions in terms of file type, size, etc. It doesn’t require the processing of complex business rules and it doesn’t require the management of large databases of business-critical data.

Now, the story gets more complicated (as these things always do) because, at present, if you want to share multiple files multiple times with multiple people you tend to use one (though probably more) of the upload-store-and-serve websites (these could be thought of as SaaS offerings). These sites work OK, but there are problems: having to wait for uploads (and sometimes media processing) to complete before stuff is available; uploading 5 files is simple and fast, but 300 files is much more time consuming; most upload sites restrict users on file type, file size, quantity; many upload sites do not preserve the original file format (e.g. many convert to flash media, destroying the original file format). izimi offers an instant option for these people and removes all the limitations currently imposed. So in this use for these users there is some benefit to having a lightweight app like izimi residing locally, and none of the MIS headaches that an enterprise app suffers from.

So what?

My realization is that it’s a case of horses for courses – I don’t see a death-match taking place across the board – within application classes, yes definitely, but across different application classes, I don’t think so. I can’t say that all software development and delivery will go one way. In some applications SaaS suits perfectly and appears to be the future (that’s why people like SAP and Siebel are feeling threatened by Salesforce.com). Simultaneously in others it doesn’t suit so well and alternative architectures may play better, hence the routes taken by products like izimi.

I think the trick is to consider the particular circumstance and ask does one route or another best help us overcome users’ problems. If one works better than the other for your class of application then use it, if it doesn’t don’t just follow the crowd.

Hey well, at the end of the day its all food for thought, and that’s the point of openbusiness.cc, lots of opinions count. What’s yours?