Saturday, December 16, 2006

Why I disagree with Tony Zingale on the future of SaaS

A couple weeks ago, I attended the 13th Silicon Valley Annual VC/Entrepreneur luncheon organized by NVCA. The guest speaker was a well known and highly successful Silicon Valley veteran: Tony Zingale, President and CEO of Mercury Interactive. Great choice for the last event of the year. The theme was "the future of software".

Tony Zingale started his speech with an overview of his career and - leaving aside the comments on how to handle option backdating - his key advice could be summarized as: do sooner, assemble a great team, select board members wisely. He is a great speaker and I enjoyed the speech, but after 40mn, I was still waiting to hear about the future of software. At last, he addressed the subject and presented his perspective:
1) Growing need for application management and mapping software
2) Future is in Service Oriented Architecture (SOA) - more and more applications will be built from building blocks
3) Security will continue to remain a key element of the stack
4) Software as a Service (SaaS) business model will not pay off in the long run as the cost of sales and services will increase (main argument was that SaaS companies have to "resell" their service every year - even sometimes every month)

I was in line with his first three points, but the last one on SaaS really surprised me. It is true that the SaaS model faces some challenges like service reliability, data privacy, lower level of customization, integration with existing applications, vendor viability concerns... however, public SaaS companies like Concur, Saba, Taleo, LivePerson and Ultimate are growing 50-100% per year and trading at an average 5x trailing revenue multiples (I excluded salesforce leading the pack with 9x!) whereas traditional software companies of this size have a growth rate of less than 10% and revenue multiples of 2-2.5x.

Why? There are several reasons driving the success of SaaS:

From the customer standpoint:
- SaaS provides superior economics to the customer: lower TCO (20-30% less than the traditional software model), no/small upfront payment (pay as you go) and low ratio of upfront integration services
- SaaS takes out a lot of IT pain: it is easy to try and buy (and lots of services have free trials, so you can see what you will get before buying it), it is a predicable cost model for companies, it can scale up or down easily, customers do not feel "locked-in" by their vendor and it requires less infrastructure
- SaaS accelerate the pace of innovation: as traditional software company provides new versions every 3-4 years, SaaS companies have new releases every six months, and the upgrade is transparent for customers. In addition, SaaS companies can monitor the usage patterns of their customers and therefore better align innovation with customer needs

From the SaaS company standpoint:
- The service model provides more predictable revenues and cash flows
- Capital requirements are much lower than for traditional software development: a start-up can launch a product and acquire several customers with less than $1m
- The time to market - and therefore break even - is also shorter: 6-12 months for SaaS vs. 18-24 months for a traditional software company
- All the customers are on the same version of the application: this makes the maintenance and support a lot easier

To get back to Tony Zingale's concerns about the cost of sales, I am not sure it is relevant - the payment method is independent from the delivery model. Companies like Microsoft have developed subscription-based licensing contracts without a SaaS delivery model and SaaS companies are selling multi-year contracts for their services.

So, to conclude this post, I think that all the benefits provided by the SaaS model will overcome the challenges, in particular for small and medium businesses (SMBs), which are the most price sensitive customers. SaaS will give them access to technologies that were accessible only to larger companies in the past, giving them a new edge to compete. With time, SaaS will also penetrate larger enterprises (main challenges being integration with legacy systems, reliability and privacy), starting with applications that are not touching their core competencies. Salesforce has started to move up this path with large customers like Symantec.


Philippe Botteri said...

My colleague Sarah wrote a nice article on the value of data collected by SaaS companies. Here is a link to her post: Future of SaaS: don't forget the data!

Unknown said...

I heard the same speech, and I too was blown away that Tony felt that the SaaS model was doomed. I think he must have been biased by the infrastructure view of Mercury/HP, which admitedly will not be a SaaS early adopter market for many reasons.

However, I have to admit that I'm somewhat excited by the fact that brilliant software execs like Tony (and I do think he's among the best in the business) still don't get that SaaS is here to stay and is changing the market in fundamental ways. This means that the incumbants still aren't moving fast enough, and well-funded startups can run laps around the traditional vendors.

Unknown said...


I have to both agree and disagree with you.

First it’s hard to believe that such visionaries like Tony don't see SaaS as the future of software and hence why I agree with you. Without a doubt SaaS is here to stay because it makes sense to both parties.

Unfortunately I have to disagree on two of your value points from the vendor's point of view.

- Capital requirements are NOT much lower than for traditional software development, in fact they are much higher (70% to 100% more) for several reasons that you can read here

- Both time to market and break even are also NOT shorter than traditional software companies.

Actually these are 3 of the few negatives of SaaS from the point of view of the vendor.

Requires more capital because:
- It is harder to develop for many new factors like tenant data isolation, user management, subscription services, etc. (Same reason why time to market is longer)
- It takes longer to break even because revenue from customers is received in small increments (usually is equivalent to a 2 to 3 year amortization of a traditional software license)

All 3 of these downsides are being addressed by SaaS Platforms like Salesforce’s Apex and Apprenda’s SaaSGrid but none the less are still relevant to mention.

Other than that you were dead on :)


Anonymous said...