Harry Debes, CEO of Lawson, created quite a stir in an interview where he pronounced that SaaS would "collapse" in two years. Many others have weighed in on this including Vinnie Mirchandani, Jason Corsello, and Sarah Lacy. They make the case for why SaaS will last and is good for customers. One of Harry's arguments is that there has only been one really successful SaaS vendor, Salesforce.com. He says:
"An industry has to have more than just one poster child to overhaul the system. One day Salesforce.com will not deliver its growth projections, and its stock price will tumble in a big hurry. Then, the rest of the [SaaS] industry will collapse."
There are arguably other successful SaaS vendors (especially in HCM as Jason notes). However, Salesforce.com has been the poster child. Joshua Greenbaum, who has previously predicted that demise of Salesforce.com, had an interesting post where he discussed Harry's comments. He made an important point. There are other factors at play in the success or demise of Salesforce.com beyond just the SaaS model. However, he is not a big long-term believer in pure-play SaaS. Joshua believes the future is a hybrid model (vendors offering the customer a choice of delivery models).
As I have said previously, I believe there is no such thing as a SaaS industry or market. SaaS is a delivery model that is used in many different software markets. So, I disagree with Harry on that point. There is no industry to collapse. Joshua (as well as Harry) is right though that SaaS is not necessarily the best answer for all software markets. However, I believe Joshua comes to the wrong conclusion. I do not think the industry will move to the hybrid model. I think the industry will continue to have multiple models. Different markets have different requirements. Pure SaaS solutions may be appropriate for some and not for others.
There is nothing particularly wrong with multiple models from a customer perspective. In the HCM software market, I find that that most customers choose the best-fit application first, then they evaluate the delivery model choices. If SaaS is the only option from the best-fit vendor, then that is usually not an issue for customers. If the best-fit vendor offers multiple options, then customers choose the best one based on their needs. Harry's own experience with Salesforce.com is typical.
"We use Salesforce.com, and I like it. But I would've bought the product even if it wasn't SaaS. The success of Salesforce.com, in my opinion, has to do with their product being good, not because it's SaaS."
So, are you with Harry? Is SaaS over-hyped? Are you with Joshua? Is the pure-play SaaS model untenable in the long-term? I am on the record here and here (Gartner subscription required) as believing that SaaS usage will continue to grow. What do you think?
Update
I received the following clarification today from Kathy Nottingham at Lawson providing additional color on its position on SaaS:
"While in APAC a couple weeks ago, Harry Debes was interviewed by ZDNet Asia. He made some bold, broad statements predicting the demise of SaaS which created quite a stir in the analyst/blog world. Harry stands by the general statement, but here are some points of clarification:
While Harry has strong opinions about the long term viability of SaaS, he and the rest of the Lawson organization are not trying to dictate application delivery options to the market, but rather respond to market demand. Travis White posted an "opinionizer" article on lawson.com clarifying Lawson's perspective on SaaS.
http://www.lawson.com/wcw.nsf/pub/Opinionizer
While we have not seen significant market demand for SaaS delivery/deployment options for our general ERP solutions, we have seen interest in SaaS solutions in the HCM space. Lawson is delivering SaaS HCM solutions to our clients and providing other deployment options as well. Larry Dunivan provides more Lawson perspective on SaaS in the HCM space in his latest blog posting:
Hopefully, these articles provide a more complete perspective on Lawson's SaaS position. If you have any questions about Lawson and pricing/delivery options that we offer to our clients, please contact me.
One of the things that Larry mentions is the ease of switching that will result in loss of growth and the crash of the stock price. For example, Rightnow technologies has a retention of about 88%, i.e., 12% of customers leave every year. This will only grow as more SaaS vendors start making a lower cost offering. The only way to get around this is to retain customers with advanced technologies. Contact me for details anandkarasi gmail com
Posted by: Anand Karasi | August 29, 2008 at 05:57 PM
SaaS will survive and you are right, it is not an industry. However you do have to be committed to the model and know how to make it work. Subscription marketing is very different than selling a license or a shrink-wrapped box of product. A lot of companies will declare it a bust simply because they don't understand it and can't make it work - not because it's a bad business model.
Posted by: Ian Gilyeat | August 29, 2008 at 10:22 PM
Interesting topic. I helped launch PSFT eCenter in 2000 which was their version of SaaS but in reality was an alternative option to hosting internally. I then moved to IBM in their Applications on Demand division which was the rebranding of the Corio acquisition. The value of hosting then SaaS to an enterprise is based upon the culture of core competency. Is I.T. management the business you want to be in? Or do you want to make a better widget? Back in 2000 most companies only considered Hosting/SaaS because the option was introduced by the vendor. Then a few years later Hosting/SaaS was requested in every RFP. Today, HR departments are approaching delivery with the mindset, why not SaaS as internal I.T. is simply another "vendor" option. I think the real challenge is on the vendor side. Profits are back-end loaded so the greater the volume, the greater the deferred revenue. As the industry moves to SaaS, vendors miss the upfront license revenue and if a software vendor decides to transition from a license model to SaaS, they better be ready to explain the revenue dip to shareholders.
Posted by: Scott Pruitt | August 30, 2008 at 09:55 AM
Retention rates are certainly important to a vendor business model based on SaaS delivery. It is a good metric to watch. There can be good business models based on SaaS and poor ones (delivery model does not equal business model). However, I do not buy the switching costs as being significantly different for different delivery models. If you have to implement a different product, you still have to pay to license that different product regardless of the model.
Posted by: Jim | August 31, 2008 at 02:32 PM
I think Jim is right on the money both in increased SaaS growth and the likely need for most major solutions to include hybrid offerings.
The decision is not cut-and-dried; there are customers who may clearly benefit from either SaaS or on-premise based on the situation.
The dynamics of low barriers to entry (low upfront prices, often tactical needs, and motivation to innovate) will run into higher barriers to exit (change costs, ecosystem growth, long-term costs) and dim the red-hot glow of SaaS at some point, but SaaS will remain a great idea and value for many, many customers.
The notion of SaaS just dying shows Mr. Debes to be totally out of touch- there are literally thousands of SaaS startups out there.....
Posted by: Martin Snyder (Official Recruiting Blogosphere Beadle) | September 01, 2008 at 11:02 PM
Personally I agree with both you and Joshua. I think that SaaS as a delivery model will continue and that some customers will like this delivery and some would prefer OnPremise.
What I think you should talk about though, is the cost to the vendor for multiple delivery models. In the end a hybrid is desirable but expensive, as both SaaS and OnPremise have different cost structures and support mechanisms needed. Not every vendor will be able to support that kind of flexibility, but good for customers if they can.
Posted by: Meg Bear | September 03, 2008 at 04:25 PM
Meg,
That is a good point. Hybrid SaaS is difficult. For Gartner clients, Rob Desisto has published extensively on this topic on (see "Hybrid SaaS Deployment Models Should Cause Concern" http://my.gartner.com/portal/server.pt?open=512&objID=219&mode=2&PageID=466517&docCode=146391&ref=docDisplay).
Posted by: Jim | September 04, 2008 at 08:57 AM
Jim,
I agree with your view that the SaaS model is more appropriate for some applications than others. SaaS works particularly well for talent management applications. Among other reasons, it’s an effective way to deploy and update an application that’s used by a very large and widely-dispersed population.
I would, however, enlarge on your assertion that SaaS refers to just a delivery model. Done well, the SaaS delivery model actually requires changes across the vendor’s entire organization including development, finance, legal, operations and marketing. In other words, the impact goes well beyond the delivery mechanism. For example, a nimble SaaS vendor who delivers frequent solution enhancements will service the customer, capture customer feedback, and develop enhancements to respond more quickly to these feedback loops in ways that are completely unlike on-premise vendors.
In one sense, I would actually agree with Harry Debes. Moving from a traditional on-premise model to a SaaS model puts a great deal of stress on a vendor. But just because it’s a difficult challenge for vendors, it doesn’t mean it’s not the right thing to do for customers. As Jeff Kaplan at THINK IT Services points out, “corporate executives who refuse to pay attention to fundamental market changes will pay a severe price.” You can read his full post here: http://thinkitservices.blogspot.com/2008/08/refusing-to-accept-change.html
Posted by: Tod Loofbourrow | September 04, 2008 at 04:26 PM
I was speaking more from the customer perspective than the vendor perspective. I did think about writing about the vendor side of it too, but the post was already pretty long. You are quite right that from the vendor perspective, SaaS is more than a delivery model. It fundamentally impacts the business model of the vendor. My previous response to Meg's comment that referenced the research Gartner has done on Hybrid SaaS is a good example of the potential business model impacts.
Posted by: Jim | September 04, 2008 at 05:23 PM
Interesting post Jim.
I'd like to add that there are really two pieces to what is referred to as a SaaS model, and only one of those is the distribution model. The second piece is the licensing model, and traditionally SaaS combines both On Demand deployment with a rental licensing model. Modern software companies are realizing this, and you are starting to see vendors offer both choices independently. I think customers win when they can choose the deployment model that suits them best, and the licensing model that suits them best. Good software vendors will evolve their business to give customers these choices.
I do agree that these issues are secondary to the primary issue - namely what solution best fits the organizations needs?
In regards to the hybrid or choice model and the impact on the vendor's business model, I think that many of the SaaS vendors in Talent Management a much higher risk than a solid profitable hybrid vendor. The good news for customers is that can take their data with them, so no matter their choice they can switch down the road if they need to, though they will incur some costs to do so.
Posted by: Sean Conrad | September 05, 2008 at 09:37 AM
Great comments here: I would argue from a vendor perspective that offering a hybrid model actually makes for a stronger SaaS solution for several reasons. 1) the software code may be more robust since it has to be tested for multiple environments rather than only the SaaS vendor's 2) It encourages vendors to create and maintain tools and processes to move customers from one model to the other, which gives SaaS customers key long-term options. 3) it focuses the vendor on creating software that is easy on HW resources and data resources that are easier to maintain because DBA's without direct knowledge of the data will be doing that work.
Lastly, it really tends to force both vendors and customers to focus on the actual cost drivers of any given solution, because the $ delta between the choices usually somehow derives from real-world cost/value relationships (rather than for example just SaaS market pricing forces).
I could not be happier that people like Jim are bringing these issues to the fore.
Posted by: Martin Snyder (Official Recruiting Blogosphere Beadle) | September 10, 2008 at 08:26 AM