SaaS - new software model, new challenges

Software as a Service (SaaS), Software on Demand, Hosted Solutions...it goes by many names and is a new emerging software business model. The benefits to the buying customer are obvious...rent the software and hosting infrastructure rather than buy it up front. But, what happens to the software company business model? The answers are different for start-ups versus established software companies.

Start-ups need to get to revenue, and then to cash flow break even quickly...before the cash runs out. Any entrepreneur knows that cash is king. Selling up front perpetual licenses brings in the cash, but SaaS subscriptions provide a smooth predictable revenue stream that builds up over time.

Michael Skok of North Bridge Venture Partners says that it takes 70% to 100% more capital to fund a SaaS company to a liquidity event than a traditional perpetual license company. It also takes 2 to 3 times longer to get there. However, NBVP believes the extra time and money is worth it and pays off in higher market cap values.

Mr. Skok is not guessing...he has real numbers. NBVP has 8 funded SaaS companies out of a total of 53 portfolio companies. The estimates are averages for enterprise software applications. Consumer applications can be done less expensively, and get to market faster. The feature requirements, integration and customization requirements are also usually less demanding for consumer applications. NBVP actively invests in both enterprise and consumer SaaS companies.

Michael Skok was a panelist on a Mass Technology Leadership Council session on Software as a Service. He shared the following facts based on his investments in SaaS companies;

• SaaS companies need an average of $35M in VC capital, versus $20M for a similar perpetual license company.
• It takes 6 to 7 years to get to a liquidity event (IPO or acquisition)
• Public equity markets pay a 10% to 20% premium for predictable revenue streams
• SaaS companies move faster than big companies. They can introduce new features instantly versus waiting for the next major release. Think years.
• SaaS requires an architecture that supports end user customization
• Industry standards are critical for interoperability
• Steady state business models require 15-18% for engineering and 30-35% for Sales and Marketing.

Established traditional software companies must deal with "renting" their software for something like $5 per user/ per month versus selling it for $20K. This effects quarterly profitability and sales compensation. Established companies have the advantage of cash flow from their perpetual license business to sustain them while the SaaS business is growing, but the strategy conflicts are pervasive.

If you are a sales person with a $1M quota and you have the choice of selling a $100K perpetual license, or a $2K a month SaaS subscription, which would you do? Some companies are changing their sales compensation models and metrics to level the playing field. They might give quota credit for 24 to 36 months of the subscription and pay commissions as the revenue comes in. Or they might just pay commissions based on 18 to 24 months of subscriptions.

Engineering costs for SaaS software are higher for both start-ups and established companies, and it takes longer to develop. Engineers must design in user customization and remote development capabilities. It takes a unique architecture and extensive User Interface design to allow customers to customize and integrate a hosted SaaS application, especially remotely. Remember, this isn't shrink wrapped desktop software.  Most enterprise software is customized in some way for each customers environment. Traditional companies sell consulting services to do this. SaaS companies must build the customization features into the base software.

Hosting is a big deal, and a subject unto itself. I will not get into it here but suffice it to say that it is very expensive and complex to build and manage a 24X7X365 service that is always up. Most companies outsource this, but it is not free. It must be factored into your pricing model and cashflow model.

How do software companies price their SaaS offerings? Jim Geisman president of MarketShare, a software pricing consultancy, said most companies use a 3 year break-even formula for SaaS vs. Perpetual license prices. Jim makes his living consulting on these issues so I will not reveal all his secrets, but he did share this hypothetical example.

A SaaS company prices its product at $65 per user per month for 10 users. A competing company offers to sell you the same software for $17,500 plus the normal 15% annual maintenance. Which is the better deal? It depends on how many users and your time horizon. The break even point is 3 years. After that SaaS actually costs more than the traditional license. But, after three years you would probably upgrade your traditional license to a new release, so the amortization clock starts again.

The panel predicted that SaaS would continue to grow as a percentage of total software contracts, but that both models would persist for a long time. The estimates are that about 10% of software is sold as SaaS now and that over the next 5 to 10 years it will grow to 25%.

If you are considering the SaaS model for your software company make sure you focus on these things;

• Cashflow - VCs must be patient and willing to fund you to break even.
• Sales compensation - Look at metrics, total compensation, and cash flow.
• Engineering - Have an open architecture that supports end user customization
• Hosting - Partner with a solid, well funded, hosting company
• Legal contracts - Automatic renewals, termination clauses, SLA (Service Level Agreements), privacy issues, data loss, data export, and a variety of other issues must be addressed in your subscription license.

Done correctly SaaS is a very compelling and rewarding business model. Salesforce.com is the poster child for SaaS success. The stock market will reward companies that can get "over the hump" and build a predictable revenue stream. Smart VCs are willing to commit the funds necessary to get to liquidity.

Comments

[James Governor]

some UK innovation for a change - a company called smartfundit offers an online calculator to see how a packaged software company can smooth the transition to a SaaS company. The cool bit - it offers connections to VCs potentially willing to fund the transition.

check out smartfundit.com

funding as a service

disclaimer: they are a client

[Yoav Shapira]

Thoughtful, insightful, well-written entry. Thanks!

[Rami Hamodah]

As an on-demand CRM provider, I would have to agree with this article, but with some added notes:

- Upgrades to the SaaS model are more frequent and free, while on-premise are not. Take Microsoft for example, last year they had CRM 1.2, now they are releasing CRM 3. That will represent a 65% increase of license costs in less than 12 months.

- Software as a Service is actually less costly to produce, because of web standards available, look at writely.com for example, 3 man operation produced a SaaS word processing application that is actually better and more user friendly than Microsoft Office.
- The hosting of SaaS is not free, but it surely costs a lot less to host 10 servers for 1000 customers (SaaS) than to host 1000+ servers for 1000 customers (On-premise)
- There are so many types of businesses that will not benefit from on-demand solutions because of customer-sensitive information and government legislation ( I don’t think the FBI, or Bank of America will outsource enterprise applications to SaaS providers, therefore the End-of-Software slogan by Salesforce.com is just hype.

Also the risk for SaaS providers is greater since attrition rates can escalate if service is bad considering the ease in which customers can jump ship to other providers.

[SaaS Follower]

This comment is in response to Rami Hamodah's comment: "- There are so many types of businesses that will not benefit from on-demand solutions because of customer-sensitive information and government legislation ( I don’t think the FBI, or Bank of America will outsource enterprise applications to SaaS providers, therefore the End-of-Software slogan by Salesforce.com is just hype."

Here you are entirely wrong. To use your example against you, Bank of America pays millions for SaaS from Taleo (www.taleo.com) a pure-play SaaS vendor of talent management software. Taleo also has very large government organizations as customers greater in size than the FBI.

Do the research before you make such comments. As an on-demand CRM provider you have reason to be more optimistic about adoption of this model for large customers. Salesforce just closed a 6000+ user deal with Cisco. BusinessWeek also has a recent article with many more examples that prove this is not just hype:
http://www.businessweek.com/technology/content/apr2006/tc20060417_996365.htm?campaign_id=topStories_ssi_5

[Rami Hamodah]

That is true. But what I was trying to say is that at all times some organizations will not outsource certain data to vendors in our space.

Talent Management is not the same as Bank Accounts information! The average customer of Salesforce.com has less than 19 seats, please do the math.

Don't you think I would be a lot happier if SaaS was the only way to go? That would increase my business by 10 folds! But the truth is, it is not there yet, and it will take at least 5 years for SaaS to have a 50% market share of enterprise applications.

Also recent outages of SFDC certainly make it harder for us to convince some customers that SaaS is the way to go.

[John Martin]

The "3-year breakeven" argument for SaaS is a myth: the on-premise cost doesn't include cost of the infrastructure (servers, database licenses, application server licenses) for multiple environments (devevelopment, test and production), installation and customization (often many times the license cost), and operations (DBAs/SysAdmins, maintenance costs on the infrastructure components).

If these things are included, there is no breakeven ever -- SaaS is perpetually less expensive than on-premise.

Plus, SaaS provides much higher value:
- Fast time to value (days/weeks instead of months)
- Ongoing free upgrades (rather than waiting several years to justify a costly upgrade project)
- Top-to-bottom support (how many times have you heard from an on-premise vendor "upgrade to our latest release" or "the issue is due to your customization"?)

So if you have an option between a SaaS solution and an on-premise solution, there's really no comparison.

[Don Dodge]

John, The three year break even analysis came from Jim Geisman president of MarketShare, a software pricing consultancy. He had a pretty detailed analysis of all the associated costs, which I didn't get into in my blog post.

But, you are probably right about the cost of hardware servers, DBA's and system admin people. These are huge costs that many companies have carried for years as a necessary part of doing business. Some of these costs could be avoided if they shifted to more SaaS applications.

The customization costs of traditional software are probably about the same for SaaS. Most companies have integration and customization needs regardless of the application. SaaS doesn't change that unless they have spent the extra development dollars to build in personalization, customization, and easy integration interfaces.

I agree that SaaS is a popular licensing and delivery model, and that the software industry is moving in that direction. However, there are issues to consider for both the customer and the software company.

[Lloyd D Budd]

Hi Don,

This is a very interesting topic to me, particularly as it can be applied to building open source software and selling it as a service. Thank you for your usual concise presentation.

What is meant by "Steady state business models"? And in the context of "require 15-18% for engineering and 30-35% for Sales and Marketing." Is that a percentage of total operating costs? And that compares to?

What are some other examples in the consumer and enterprise spheres of successes and failures?

Thank you,
Lloyd

[DonDodge]

Lloyd, Steady state means the normal run rate or "state" after the heavy investment in engineering for development and marketing for the product launch. Once those big investments are done you get to the steady state spending model.

The 15% for engineering and 30% for sales is a percentage of revenue. This is a good business model for enterprise software. Manufacturing businesses, or computer hardware businesses have very different spending models as a percentage of revenue.

Open source software is a great match for the SaaS delivery model. The development costs are lower so there is less pressure to generate big immediate revenues before the cash runs out. VCs expect a liquidity event within 5 years of initial investment, and they don't expect to put more cash in, so...the pressure is on to sell traditional up front licenses. SaaS is probably a better long term business model.