SaaS and the Servitization of the Software Industry Part 2

In the first part of this post, I explored some of the historical trends that have led the software and IT  industry in general to be structured as it is today. The key point is that for most enterprises a gap exists between themselves as users and purchasers of software and the actual software vendors. The gap is filled by third parties such as value added resellers (VARs), system integrators and outsourcers as well as other elements of the distribution channel such as retailers and distributors and all the technology infrastructure that supports them. So the big question then, is how is this large group of intermediaries affected by such a fundamental change in the customer/vendor relationship that SaaS potentially brings?

To try to answer that question it’s necessary to break it down into a number of different elements. It’s also necessary to set some sort of context as terms such as enterprise customer and software vendor are quite broad definitions. The following sections try to provide this context and work through each key area of impact for SaaS.

The Scenario

Software vendors come in all shapes and sizes, as do their applications and customers, and the impact and uptake of SaaS will be different depending upon the demographics involved, so for the purposes of this exercise I will assume the following:

The customer is a mid size enterprise with multiple office locations and 5000 staff. They have outsourced most of their IT functions and assets to a 3rd party outsourcer but retain a team of IT strategy, governance and architecture staff.

The application suite they are thinking of migrating to a new SaaS vendor is currently implemented using an on-premise system that runs on on multiple geographically distributed servers with a client-server architecture and various customisations made over the years.

The Players

There are a number of participants in this, fairly typical, enterprise ecosystem both in terms of job functions and organisations. They are broken down as follows:

Enterprise Customer

These are the people using and paying for the software application and its supporting infrastructure. In previous years this would have included a sizeable IT department that installed, configured and supported the application environment but most of these functions have been outsourced to a third party now and the remaining IT staff are a small team who set the corporate IT strategy and framework.

The Software Vendor

There are actually a number of software vendors installed within the enterprise. These are traditional on-premise, perpetual license vendors whose software is a mix of server, client and client-server applications.

The Infrastructure Vendor

Like most enterprises, the customer has purchased and deployed a range of computing and infrastructure systems over the years and this has led to the building of data-centres in major locations consisting of racks of servers, storage, networking gear and remote access equipment – all to power and provide access to the enterprises software applications and associated data.

The VAR / System Integrator

This organisation has resold the applications and servers of the existing infrastructure for some time. They are a channel partner of the primary software vendors and offer a broad range of services to the enterprise customer including design, installation, configuration, optimisation and customisation of the existing applications. In recent years their scope and level of activity has been squeezed somewhat by the outsourcing organisation who now operates almost all of the enterprise IT infrastructure under a 5 year deal and is taking on many of the activities previously performed by the VAR.

The Outsourcer

The outsourcer is a large managed services business who specialises in providing a complete turnkey IT solution to enterprises whereby they take over the support and operation of the enterprise IT environment including the transition of enterprise IT support staff and potentially assets such as data-centre equipment (servers, network equipment etc).

The Transition to SaaS

So somewhere along the line, the enterprise customer decides to move to some/most/all (delete as required) of their application needs to a SaaS provider. In general industry terms this is going to be a slow but steady transition. All analyst estimates and current data show SaaS deployments out performing the market (in terms of percentage growth) by a very significant margin – albeit from a much lower base – so over time the transition will become more prevalent. Of course the pace of change and impact will vary enormously between particular customers and particular application segments but again for the purpose of this discussion lets assume it  happens nonetheless and focus on its impact.

The Drivers

Well you can look at any SaaS vendors list of benefits to see the claimed advantages of a SaaS model but here are some of the obvious ones for our example enterprise:

CapEx Savings

When an enterprise moves from an on-premise (lets assume server based) application deployment to a cloud based SaaS offering there is a significant transfer of infrastructure responsibilities. All of the server side hardware, software and related infrastructure is no longer required in-house (again this assumes a complete SaaS hosting scenario rather than some sort of hybrid model). This will result in CapEx savings in the following areas (but I will address the issue of current versus future savings later):

Servers

The server compute hardware is no longer required in the enterprise data centre for the SaaS migrated application. In addition to the hardware costs there are also things like OS licenses, virtualization licenses and so on.

Storage

Whether inside the servers or in external dedicated storage infrastructure, this hardware is no longer required when the application it serves moves to the SaaS cloud (there will of course be storage needs in the enterprise for some time to come though).

Networking

The networking gear that lives within and between the enterprise data-centres such as Ethernet switches, routers, internal firewalls, load balancers, remote access equipment etc. These are all in place to provide connectivity services to servers that are no longer required when the applications move to the SaaS cloud.

Data-Centre Facilities

This covers all of the facilities infrastructure to host the servers and networking gear such as racks, UPSs, air-con, cable systems and so on.

Lastly, but certainly not least, is the licensing costs of the on-premise software itself.

In summary, there is a lot of capital  costs that are directly impacted by a large scale SaaS migration. I will touch on the flip side of this issue later on.

OpEx Savings

There are again very significant operational expenditure savings to be made within the enterprise as a result of a large SaaS migration. These fall broadly into two camps – the server/data-centre operating/support costs (directly related to the CapEx saving listed above) and desktop support costs. The data-centre cost saving are more obvious and quantifiable but when you factor in the implications of a web delivered client for the SaaS application that is now the responsibility of the SaaS vender, there are potentially significant savings in desktop application support, deployment and maintenance.

The savings in OpEx then are primarily about people costs – the roles and responsibilities normally paid for directly by the enterprise for on-premise activities are now supplied by (and paid for indirectly to) the SaaS vendor.

There are also annual maintenance costs for on-premise software and infrastructure to be factored in and finally there are the utility costs of powering the various enterprise data centres.

The Constraints

While the benefits above may make the transition process seem like a no-brainer, there are of course a number of constraining factors to consider. Here are some of the commonly cited ones:

Existing Infrastructure Investment

While there are clearly major CapEx and OpEx savings to be made by shifting responsibilities to the SaaS vendor these are significantly diluted by the fact that the enterprise has already paid for and implemented its current infrastructure (even leased equipment will have term constraints). This blurs the ROI figures unless you take a longer term view. If the enterprise decides to migrate a large piece of its application infrastructure to a SaaS vendor its going to be left with a lot of redundant hardware and data-centre facilities, much of which is already amortised.

Of course by taking the longer term view the enterprise can see the savings realised both in terms of reducing future on-premise expenditure for growth and also legacy support. The difficulty is during the transition period when only some of the applications are moved to SaaS and so the data-centre needs may not be reduced as quickly or significantly as first thought – it probably means more empty space in the data-centre rather than smaller or no data-centres until a certain tipping point is reached in terms of SaaS migration.

The Costs Of Change

Another major constraint is the perceived and real costs of moving from one application platform to another. Inevitably, it is going to cost the enterprise time and money to assess, design, plan and implement any such transition. As in all such situations, the benefit of the new model must really appear obvious and significant to entice the customer to proceed.

In particular, when an outsourcing relationship is already in place there will be cost and contractual issues to be addressed if the SaaS migration will reduce the need for equipment and people that the outsourcer may have acquired directly from the enterprise.

The Vested Interests

The enterprise is not making decisions about the migration to SaaS in a vacuum. The various partners involved in the enterprise ecosystem listed earlier, particularly the VARs and Outsourcers, have an important role to play in the decision making process. Because a major SaaS migration will alter the CapEx and OpEx models of the enterprise, it will also have a major impact on those 3rd parties which form the CapEx and OpEx costs. This is in addition to the existing software vendors who will fight hard to maintain their business.

The Impact

Okay, so assuming that the enterprise navigates the various issues previously outlined and embarks on a major shift towards a SaaS model what are the likely impacts for the major participant groups? Using the same groupings listed earlier on, here are some thoughts:

The Enterprise

The impact on the enterprise will vary depending upon the nature of their existing application environment and the level of outsourcing they have already contracted in. If there is a largely in-house IT team then clearly there must be an impact on this group. The focus of the in-house IT organisation must shift from building and delivering IT services to defining them and measuring their performance and business impact. Many enterprises will already be some way down this road through outsourcing but SaaS will have an even greater impact.

There are some particular issues that must be addressed differently when the applications and data are being delivered from a SaaS vendor. These involve policies and procedures for data security, compliance, privacy, availability, redundancy and so on. The enterprise must ensure that the SaaS vendor meets and delivers on all of these items and implements the necessary contracts and SLAs to enforce them.

The enterprise must also calculate the long term costs of the move to SaaS and this must be done holistically by comparing the subscription costs of the SaaS vendor over a long time period against the savings in On-premise CapEx and OpEx.

If carefully considered and managed however, the enterprise has a lot of upside from this transition. There is almost an inevitability about this trend too. If you look at how the electrification process began in enterprises during the early part of the 20th century it has similar parallels – enterprises first saw electricity as a differentiator and built power generation plants next to their buildings for private use. The private power plant was seen as a strategic asset by the enterprise and the early power industry was structured around supplying, installing and operating on-premise power generation. It took some time for enterprises to accept the notion of some external 3rd party providing power on-demand from the grid. You could view the evolution of global manufacturing outsourcing in a similar way.

The Software Vendor

By this I mean the incumbent on-premise vendor. Clearly a major industry shift towards SaaS will dramatically change the structure and fortunes of the software industry and its vendors – over time. The implications for existing on-premise software companies and the appropriate strategic response is a tough call. But in many ways its a tough call in relation to timing only. How long should an ISV wait before implementing a SaaS strategy and then how quickly should it migrate to it fully. These are typical questions that any business operating in an industry undergoing radical innovation and change must decide.

But whatever the timing, the competitive pressures are certain. Again in typical fashion, new entrants are moving in with completely different business models than the existing software companies. Now many of the existing companies are also beginning to move, so that creates pressure from within the existing industry boundaries and beyond.

There is also something specific to consider in terms of timing and strategic options for ISVs. The SaaS vendors are not just offering the same sort of software that happens to be delivered over the web – they are in some cases creating entirely new (and often proprietary) ecosystems. Salesforce.com’s Force platform is a good example of this, by creating a rich platform and getting to a certain critical mass of customers they can now become more than just a vendor but a decision point for the enterprise customer. As the enterprise does more work on the Force platform, it becomes attractive for them to take more applications that are integrated with their datasets stored on Force than in another platform (particularly another on-premise platform).

So the more time that passes, the more external influencers that the ISV needs to accommodate (I guess this swings both ways though and at least the late movers have many more platform options available).

So the end result for the on-premise software vendor? Well in most cases they will need to embrace the SaaS model in the longer term so its now a question of when and how.

The Infrastructure Vendor

As SaaS becomes more dominant then the infrastructure vendors will find the purchase volumes shift increasingly from the enterprise to the service provider space (where the SaaS vendors will be hosted). As the server and storage needs on-premise decrease (or at least don’t grow as fast in the short to medium term) then all the related connectivity and facilities are impacted too. Instead, the enterprise application needs will focus on edge connectivity (and protection etc) with more bandwidth required and possibly direct (virtually) private connections to the larger SaaS providers.

The VAR / System Integrator

This group faces significant change. As SaaS brings the enterprise customer much closer to the SaaS vendor architecturally and commercially, many of the services provided by the traditional VAR / SI become less or un-necessary. The entire on-premise design, architecture, implementation, upgrading and related activities no longer apply in a pure SaaS environment. So is this the end of the road for this group? No, but it is a major fork in the road and they need to re-think their strategy and service offerings.

The strength the VARs have today is their customer relationships and the insight this gives them into their customer business needs and processes. This is of significant value. For most VARs they can adopt a two pronged approach (fork pun intended). They will obviously continue supporting and working with their existing customer base as they are today but can also look at the emerging SaaS ecosystem and value chain to see where they can play a role there too. If they get this right then they will not find themselves pushed out of the value chain as the SaaS migration continues. The SaaS vendors meanwhile, need to better understand their new customer base and can leverage the VARs to achieve this. Obvious activities include:

Customer Aquisition

This is a costly business for the SaaS vendor as can be seen by the level of investment they make in sales and marketing.

Customer Migration

Depending upon the application on offer there will be a level of migration from the customers existing systems to the SaaS application. This will involve the movement/population of key data and some level of customisation of the SaaS application to better align with the customer business needs. SaaS applications often differentiate themselves by offering a high degree of UI customisation and the integration of a range of optional software modules – a trend that will only increase going forward.

Onsite Training

Another major cost and concern of the enterprise customer is to roll out the new SaaS application to their userbase and a degree of training is likely to be required for this purpose.

In each of the areas above the VAR can play an important role in this process and it’s aparent today that many are engaging in this way through partnership programmes such as that offered by Salesforce.com and others. So in summary, the VARs have an important role to play in a SaaS future but it is different to the role they play today.

The Outsourcers

This group have probably the most to lose from a large scale migration to SaaS. They have built large businesses around taking over and running the on-premise infrastructure and related support functions. When the applications are no longer running in the customer data-centre, or the outsourcers data-centre on their behalf, their role is greatly reduced. There will of course remain a major role in the management and support of the front line IT services such as desktop PCs, laptops and printers but their backend role could go.

In response to this, the outsourcers can of course take an active role in the backend infrastructure on behalf of the SaaS vendors – as indeed many will given their current standing in the industry. But they are up against some serious competition from the likes of Amazon, Microsoft, Google not to mention the SaaS platform vendors such as SaasGrid, OpSource, Force.com and the huge variety of internet hosting companies who play in this space.

So in conclusion, there are both challenges and opportunities ahead for the software industry and the IT industry in general as it servitizes through a large scale migration to SaaS and it will certainly be an interesting transition to be part of.

Steve