ASPs: highway to heaven or hell?


Let’s limber up with a simple question – what is an ASP?

At its simplest level, it is an entity that hosts an application that an end-user can access. But this is actually the most complicated and confusing question you can ask because there is so much hype emanating from the ASP community.

For example, Oracle says it is an ASP, but it is in fact Exodus that is hosting everything, while Oracle just provides the software. And you have the likes of XTML, which is offering PeopleSoft. This means it has to be in bed with PeopleSoft itself, with PricewaterhouseCoopers and Hewlett Packard.

What is the ASP model good for?

Not what a lot of people think it’s good for, namely reducing cost of ownership. What it is good for is time to market, and time to responsiveness.

If you are sitting there thinking ‘we’re a bit behind the curve here, let’s put together x business functionality’, and you go out and get a systems integrator and a consultancy on the job, and it reports back six months later with thick reports, you’ll be 18 months off implementation.

But if you can go to an ASP with a specific business problem and it can solve it by offering exactly the functionality you need from its environment – that’s when the ASP model really shines.

What kinds of business problems can this model handle well?

Where it is working OK currently is at the commodity level. So, at the level of taking a company’s e-mail off its hands, and running Microsoft Exchange for it, it’s working fine.

Beyond that, you have to go back to the future. Ten years ago, we had the bureau model, where a company would outsource its accounting function, say. That kind of air-locking of non-integrated functionality is what the ASPs are doing now.

Where the struggle is, at the moment, is around ERP (enterprise resource planning), customer relationship management, supply chain management – areas that are highly integrated into the rest of the organisation.

The Gartner Group is saying that the ASP model is currently going through the ‘trough of disillusionment’, having come down from the ‘peak of inflated expectations’.

It is predicting an ‘ASP meltdown’ that will be more serious than the collapse of the dot.coms, because it will draw more bricks-and-mortar companies into the debacle. It also counsels users not to commit to ASPs for the next 18 months until the market consolidates on the ‘plateau of productivity’, after traversing the ‘slope of enlightenment’.

Where do you think we are on this pilgrim’s progress and will the ASP ever get to the Celestial City?

We are just coming off the top of the expectation curve. The dot.coms drove the ASPs, and the ASPs were seen as the enablers. Reality has now started to set in, and Gartner is right about that, but we won’t say to the end-user: ‘Don’t do it, wait for it all to sort out.’

Who are these ASPs?

Name the names. It’s a stack!At the bottom, there are the carriers, the providers of the fibre and the copper like UUNet, AT&T, and, nominally at least, BT Ignite, and so on.

On top of these bandwidth providers, you have the infrastructure suppliers, the Exoduses and Globixes of this world. Now, they do host sites themselves – for example, Exodus hosts Yahoo – but what they mostly do is look after the infrastructure for ASPs.

Then you have the NetStores and the iProvides, which operate at the commodity level of e-mail, (Microsoft) office functionality, and so on. Then, at a higher level, you have the advanced business functionality guys like XTML and CSC – which although it doesn’t advertise itself as an ASP, is hosting more business critical applications than anybody else in Europe.

Once you’ve decided that you have a business need that the ASP model can meet, what are the key qualities to look for in a provider, and what should you look to get out of a service-level agreement (SLA)?

If you’re an SME, it is a case of first steps – e-mail management and some data management, but not business functionality. Your SLA will be along the lines of: ‘If the e-mail server goes down for up to four hours, you get compensated with four hours of free service.’

As you move up the food chain and the ASP is hosting mission-critical parts of your business (an e-commerce site, say), then the SLA should recompense for, for example, estimated revenue loss, not just downtime.

Is that kind of level of protection common?

No. The vast majority of ASPs prefer to go for a service credit mode of compensation, where, say, your e-commerce site goes down for four weeks, and they offer one month’s free service! These service credits are worse than worthless.

You need to ensure your ASP provides business equivalent recompense.

That said, it’s important for the user not to crucify the ASP since the ASP model is fragile at the moment.

Does this mean we are in a cuddly network economy where we have to trust each other and collaborate for all our benefits?

The whole ASP model has to be based on trust. It’s hard work on both sides and getting it wrong is catastrophic. That said, it’s not all doom and gloom. Some 80% of ASPs will go out of business, but 80% of consultants will go out of business too!

For the user, it is a case of doing due diligence – making sure the ASP understands your business, that it does have deep pockets, that it’s got a plan B, that it can scale up as your company grows.

What’s your assessment of Microsoft’s turn to ASPs?

Well, it is very bullish. Microsoft has realised that to get the ASP model working properly, you need to re-architect the application for the web. A lot of the vendors will now find that Microsoft is working with them and is genuinely into an ASP hostable model.

Now, Microsoft is not doing this because it is feeling philanthropic – it realises .Net will have a bumpy ride and it’s got to make sure that BizTalk, and so on, gains good acceptance. But I’m fairly pro reharding what Microsoft is up to.

To what extent is the ASP model relevant outside the SME space?

Well, if you look at the market that Interliant is strong in – namely the provision of virtual private network solutions for large corporates – then, yes, the ASP model is already active outside the SME space, even if it isn’t labelled that way.

When you consider that the average Siebel implementation takes from nine to 18 months, and the average SAP implementation takes from 12 to 18 months, then large organisations are eventually going to say ‘enough is enough’, and will move to the ASP model because of its time to market virtues.

It seems that ASPs need to have considerable technological savvy along with a deep commercial understanding of their customers’ businesses. Isn’t that a tall order?

Yes, ASP companies do need to have a split personality – they need, for example, to be more secure than any of the IT departments they are talking to, and yet need to understand the commercial thrust of their customers’ businesses. And so we are beginning to see ASPs courting system integrators and consultants.

What are the career implications for the IT professional if the ASP model proves successful?

Real propeller heads who know how to set up an ASP environment will be at a premium. Business analysts within the IT department who are used to taking corporate requirements and translating them into IT terms will have to move decisively towards mediating between the general market requirements of their business and the ASP.

Code writers will have the real problem because if we do move to more outsourcing of business functionality, then there will be no more in-house writing of code.

Is there a bigger picture suggested by the ASP model?

Perhaps one of access supplanting ownership, as we move from an economy based on markets to one based on networks? There is a bigger picture here, and it raises big questions.

For example, if you take the question of the ownership of the application, then there are two opposing forces at play – the software vendor that wants the user to own the application and the ASP to own the provision, and the user, who has no interest at all in owning the application but every interest in owning the data. The ASP does create revenue problems and Wall Street problems for the software vendors.

What will determine the ultimate success or failure of the ASP model?

How it is perceived. There will be a lot of failures in this model, and breaches of security. It won’t be easy, but Microsoft reckons it will be successful in a three to five-year period.

I think that, providing the ASP market can spin-doctor its messages; there will be a broader take-up by the end of 2001 and the beginning of 2002 at the commodity level. That will lead to full-service success at the end of 2002 and the beginning of 2003. I’m even more bullish than Microsoft, but, for the ASPs, it will require spin doctoring and a separation of the wheat from the chaff.

Clive Longbottom is an analyst at research firm Quocirca.

  • This article first appeared on

Related reading