Business Relationship Software Feature Article

Matt Miller, a partner with WaldenVC, has an interesting piece on the RightNow Technologies Web site arguing that, in fact, software as a service isnt the thing for everyone. His opinions are lucid and insightful:

I have been watching software pricing models struggle from my time on the Oracle pricing committee in the late 80's to being a CEO with a subscription model in 2000, to being a VC looking for SaaS investments today, Miller writes, letting us know his wisdom is hard-won: I have a few thoughts of the topic that may or may not be comforting, depending on where you sit.

1. The vast majority of enterprise software does not lend itself to SaaS pricing. I wish it did, but it doesn't. I am a bit tired of meeting startups that think they can slap a subscription model on anything that moves. My thought is that the only time you are likely to succeed with a use-based model is when the vendor is continuously delivering new value and service beyond the traditional software upgrades.

Many enterprise apps categories don't really do this. Companies buy an app, customize it to their unique needs and try not to change it. This applies in general to Supply Chain (too much custom integration here to use a vanilla app) and ERP (also the need for secrecy and security favors a behind the firewall model).

Why has CRM succeeded for SalesForce? Because the best practices sales process is very similar across many companies AND the hosted architecture is a natural for geographically distributed sales teams. I have bought SalesForce for 3 companies now and barely had to modify it... perfect for Saas. Other software types which do not favor Saas are all forms of middleware or many infrastructure products like Network Management.

2. The cheapest alternative for the customer is often a license. When selling subscription software, I have often been told by a CIO that "we can start off on a subscription" but if this thing really takes off, I will need to bring it in-house." This is due to both security issues as well as a realization that if the CIO can lock in a price one time, he will likely pay less in the long run.

Hell, this is the way CIO's have been buying software for the last 20 years, so don't expect them to easily accept a perpetual stream of payments forever especially when many subscriptions are set to one third of typical perpetual license prices. That math is pretty easy. CIO's often say that they simply need to have perfect cost visibility over time. While I understand all the arguments against this view, most still feel uncomfortable signing up to big annual subscription payments that will inevitably lapse and set up a renewal renegotiation wherein the vendor has too much leverage.

3. All this will change, but very slowly. I have been looking for investments in SaaS over the last 2 years and have seen a few areas that truly do favor a subscription model. I have seen a few lead management companies like Blue Roads in San Mateo that offer a hosted system to track and distribute leads to partners. The hosted service is a perfect architecture for this problem, since the IT group really doesn't want 10,000 partner sales reps all tunneling through their firewall to get to their internal CRM system. There is high value in Blue Roads running this system 24 x 7 and for it, one pays a subscription fee.

Similar stories apply to hosted PLM companies that allow design partners or supply chain partners to collaborate on products or bills of material. My newest investment for WaldenVC is a company called Palamida. Every night, the Palamida system spiders the internet to find more new software components, many of them open source. It then allows software developers to identify what components they have in their own code bases and what licenses might apply.

As we all know, the dirty little secret of enterprise software is that no one really knows exactly what is in a large code base. Now there is a way to find out. Like an anti-virus system, the value of Palamida is only realized with a continuous process to audit code and look for newly added components. Thus the system is purchased as a subscription just like anti-virus. Again, the architecture fits the SaaS model

I hope I see more markets start to accept SaaS, but progress will be slow since many companies own their software outright and have fear about cost visibility or value. Look for SaaS to take off where it makes sense and in new areas where old investments are not still being depreciated.