Image representing Ariba as depicted in CrunchBase

Image via CrunchBase

A recurring theme for me in my writing is the pain that traditional on-premises vendors feel when considering, or actually making, a move to SaaS. Recently I posted about the success Callidus has had with this very move. Within days of that post I was contacted by Ariba software, a vendor in the spend management space. Over the past few years Ariba has made moved from 100% perpetual licensing to the point where today subscription revenues make up nearly all their bottom line revenue.

Fellow SaaS commentator Phil Wainewright ran an interesting podcast during which he spoke with Ariba’s CTO, Bhaskar Himatsingka about the technological issues around the move. Director of Investor Relations at Ariba, John Duncan, was keen to speak to me about the bottom line impacts the move has made on the company.

Back in 2003, Ariba made the decision to move from a perpetual licensing approach to a subscription one. While SaaS has gotten widespread coverage now in 2009, back then it was early days. Even, arguable the originator of SaaS as we know it, was only four years old at that point. I was really interested to see what the move did to Ariba’s revenues, its sales model and it’s investors. It has been said many times that SaaS is difficult for traditional vendors, Duncan concurred with this coining the term “cocaine hit” to describe the quarterly impact of perpetual license feeds on a business. That quarterly fix is highly addictive and Ariba concurs that weaning an organization from it is a difficult ask.

Initially Ariba took the interesting approach of selling on-premises software, but on a subscription basis. This was an interim move until products were created or modified to run as true SaaS offerings. Thereafter it took Ariba close to five years to recover total revenues lost with the shift – the chart below shows perpetual, subscription and total revenues earned in the years after their decision to move to subscription.


(Note – the 2008 result has had the impact from the acquisition of another company removed from the figures also the 2009 figures are, of course, budgeted).

Duncan admitted the fact that the move created a revenue trough for 2-3 years that they’re only now beginning to recover out of. He did however explain the value that Ariba has seen from their “landing and expanding” strategy whereby they can sell a small on-demand module to a customer and gradually on-sell further modules. The modular approach their business have has really helped therefore to offset some of the more corrosive effects of the move to SaaS. Their share price similar suffered a lull in the year or two following conversion – this is a temporary aberration and good investor communication eased this in Ariba’s case.

We discussed the different skills needed to sell SaaS as opposed to on-premises software. Duncan told of the old days when top gun salespeople were feted at company events for selling the biggest ticket value application deal. Salesperson compensation was in fact based on annualized revenue whereas now they compensate based on true recurring revenue. Ariba have moved from seeking “SAP style” sales staff to hiring staff with experience selling subscription based products – proof that no matter what aspect of the tech world you’re involved with – engineering, sales or management, professional development is critical.

Sales personnel are measured on different metrics, the number of forward deals or the number of subscription customers renewing for example. Duncan described the vendor/client relationship for SaaS much more like a marriage than the “love them and leave them” perception that one time license sales unavoidably creates.

The investor reporting metrics that Ariba use today are also far different from what they did in the old perpetual licensing days, Ariba makes use of what they call a “backlog chart”. This shows committed future revenues – ie those that are for contracted future sales. We discussed the fact that the backlog chart commentary needs to be articulated in a different way to ensure investors and analysts aren’t confusing it with the murky (and generally highly fictitious) sales forecasts of old.

Similarly Ariba places much importance on their subscription renewal percentage. This has grown from around 80% when they first moves to subscription billing to close to 90% today – proof that they’re moving to a more responsive approach when it comes to customer support.

CloudAve is exclusively sponsored by

Ben Kepes

Ben Kepes is a technology evangelist, an investor, a commentator and a business adviser. Ben covers the convergence of technology, mobile, ubiquity and agility, all enabled by the Cloud. His areas of interest extend to enterprise software, software integration, financial/accounting software, platforms and infrastructure as well as articulating technology simply for everyday users.

  • What a great post. This sounds like a story of an alcoholic that was able to beat the habit and triumph. I applaud the persistence of the Ariba and Callidus executives to live through the 2-3 year revenue trough and do the hard thing that was right for the market. Both markets were going On-Demand. Ariba and Callidus did what Siebel couldn’t. Interestingly though both markets are relatively smaller as compared to CRM and mostly competed with custom in-house rather than packaged applications that were similarly priced. Also Ariba and Callidus are much smaller companies as compared to IBM, SAP, ORCL.

    That said, I cannot imagine companies like IBM, SAP and Oracle to do the same in the BI space. On-Demand software will still be on the fringe in these companies. The are three ways these companies will be able to offer On-Demand alternatives: acquire someone, spin-off and create a separate company or the Larry way (fund the creation of a / Netsuite). BI is too big a market with too big a revenue impact for these companies to do what Ariba and Callidus achieved.

  • Ben,

    Just my two cents here. The missing important “bit” to me is that the business logic of Ariba apps had to be moved to a web native model to allow the SaaS model to establish itself. In effect, being SaaS is not just about being a monthly subscription model, rather it entails the app to have enough services embedded for business users to subscribe to, from their browser. The technological leap being expensive up front is the sweat to wear until the business starts collecting the honey. Let’s be real, Ariba decided this move because?? Legacy applications are losing trail, it’s the whole discussion of on premises vs off premises..Nevertheless, Ariba needs an accolade for they took the technology leap, kind of…they followed the steps of an early innovator-aka salesforce-, most likely because they liked the economics behind the model: it is the future of technology.

    Secondly, going “SaaS” means adopting a different attitude to cash-flow, marketing, promoting, selling and supporting apps. More than ever before, your business app is out to…”Service” your customers…and this is the catch! The most important word in SaaS is not software, but Service…thus turning a technology company onto a service company is a real detox…but can the management lead it forward??

    Perhaps, Ariba should focus on sustaining their ability to deliver the services their customers need which entails asking them question, gathering their feedback and co-creating the apps functionality road map.Is it a fair comment to say that for as long as investors (by the way, the hands on investors or money in-out investors?) and analysts feel comfortable Ariba ask their customers what they want and Ariba delivers on it, maybe they will see the value coming out of a browser?

  • This is a great post. I can relate to Ariba as our company went through the same hoops as he. We provide hosted CRM and subscription business is definitely the way to increase your bottom line if you have the producet that you can depend on and the customer service that is exceptional.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.