I recently wrote a post for InformationWeek titled Why Multitenancy Matters in the Cloud. I described in my post why multitenancy is the most direct path to spending less and getting more from a cloud application. The blog post was actively discussed on various forums, including LinkedIn and Google groups and several experts raised very good questions. One question that was debated vigorously was whether a single-tenant or a “hybrid” model would or wouldn’t qualify as a Cloud solution.

I think that you can offer Cloud/SaaS products as single-tenant, multitenant or both (meaning one model for Goldman Sachs and another for Peter’s Garage) and also label them the way you want … the main question is that of revenue and cost … or, more plainly, whether you will become a winner in your space in the long term.

The keywords are both “winner” and “space.” If most of your customers can be served through multitenancy, you will have a winning SaaS/cloud product since you will eventually win the cost/value battle over your competitors. If most of your customers, on the other hand, require customizations that make your model single-tenant you may eventually become a winner in the consulting/SI space but not in the product space (someone else will figure out how to offer a multitenant product at a lower cost).

It’s a well known fact that Marc Benioff’s early investors pressured him to also offer single-tenant versions of Salesforce and give customers a choice… we can all guess where salesforce.com would have been had he accepted that compromise. But let me also give you examples from experiences that I have lived (through Navatar).

Our cloud products for Financial Services on the Force.com platform are all multitenant. For instance, 70+ Private Equity firms that use Navatar’s Private Equity product run their own instances of salesforce (where the product is loaded), have made their own configurations (and customizations) however, they are still able to receive upgrades seamlessly from salesforce.com as well as from Navatar. That’s the architecture we were able to design (force.com surely helped). We sometimes have to stay away from potential customers who want heavy customizations that can prevent us from offering them seamless upgrades (even though we have a consulting practice) that’s a price we pay.

Another example: Our consulting practice specializes in building products (on Force.com) for others. One such product we recently built was for Points of Light Institute, a pioneer and leader in facilitating volunteering. They offered a client/server product to nonprofits, to manage volunteer activity within their community. The cost of maintaining and supporting their 300 customers/affiliates was too high and the model wasn’t scalable. The new Force.com-based multitenant product allows each customer to customize, create their own web pages etc, without losing the ability to upgrade to the next version (in development). However, it was the targeted saving of 50% in the cost of maintenance/support that made the move to the cloud worthwhile for them without multitenancy, instead of savings, they would have been building fatter maintenance and consulting practices.

Also, as someone pointed out, multitenancy is a necessary, but not sufficient condition for a profitable cloud product business – there are a whole bunch of others, and I complete agree. It’s also worth mentioning the blogs of those that actively participated in the debate and helped me enhance my own learning on all of this. Here are some of them that you may find interesting:

Customization: An enemy of SaaS, There goes the neighbourhood, Multitenancy in the Cloud, Principal-driven Customization, Trident Capital blog.