My blog has moved!

You should be automatically redirected in 4 seconds. If not, visit
and update your bookmarks.

Thursday 1 March 2007

Software as a Service (SaaS) business models

The Irish have always been blessed with the “gift of the gab” – a propensity to talk a lot. We even have some data to confirm it, according to the mobile phone operators here in Ireland. Ireland has consistently had one of the highest “average revenue per user” (ARPU) figures internationally for the mobile operators. Now, I can hear some cynics amongst those of you here in Ireland saying the ARPU figures are in fact driven by higher pricing in Ireland, thus “ripping-off” the Irish consumer and business communities. However the operators assert on the contrary it is all because we talk a lot - and they have told the Irish Government so!

In the mobile operator industry, there is continuous competition to first acquire customers, and then retain them and minimise “churn”. You aim to maximise ARPU whilst remaining sufficiently competitive in the market so as to maintain or even increase market share. Optimisation of tariffs is critical: understanding and measuring the tariff efficiency so as to offer attractive tariff plans to your customers, whilst maximising ARPU for their typical usage patterns.

A very naïve approach would be to offer a simple fixed price per month – regardless of actual usage – to each subscriber. Whilst very simple to understand and explain to consumers, I personally doubt whether any operator would survive in the market by offering such a tariff plan. There are a large number of tariff parameters which can be mapped against usage patterns: set-up charges, roaming tariffs, premium rate charges, rollover periods, billing band thresholds, average call rates (including bundling and discounts), flexibility in varying the size and scope of bundled services, rates for calls to other wireless and landline networks, SMS (text) charges, and MMS (multimedia) charges. Operators analyse all of these parameters, in conjunction with the observed traffic patterns generated by customers, and then adjust tariff rates to retain them and acquire new ones.

Many of us in the software sector are closely examining the “Software As A Service” (SaaS) business model, in which software is billed in proportion to actual use, or on demand, rather than the traditional pre-paid perpetual license. The software concerned is usually remotely hosted on an internet server farm, although as an aside to this discussion, I contend that the business model can also be applied to on-premises in-house installed software as well – provided of course the appropriate protections are built in to prevent fraudulent use. The essence of SaaS to me is paying for software in proportion to its usage, regardless of the actual location of the software installation. For enterprise customers – including publicly quoted companies and government systems – secure and verifiable business processes and practices are essential for regulatory approval (such as Sarbennes-Oxley and section 404 in particular) and the physical location of business critical software can become an issue. Remote hosting has benefits and disadvantages; in-house on-premises likewise.

Anyway. End of the aside and back to the main point. My own view is that SaaS is still at a relatively early stage of adoption across the industry, despite the momentum behind leaders such as In particular, I believe that many SaaS vendors currently offer a very simple charging model: per-user, per-month (PUPM). The number of staff within a customer who access a SaaS offering is measured – or agreed in advance with the Saas provider – and the customer is billed on a monthly basis in proportion to this number.

Wouldn’t it be great if your mobile (or landline) operator charged your company some fixed monthly fee as a multiple of your number of staff using company phones ? Regardless of how much they used their phones and who they called, perhaps on the other side of the planet ? Maybe some operators do have such business models, please let me know (and just to save your keyboard, I already regularly use both Skype and jajah).

Now, perhaps you might argue with me and tell me that the functionality offered by a software package in general is more complex than the range of functions which a telecommunications operator can offer, and therefore it is not fair to compare their respective approaches to tariffs. You might say that a mobile operator needs to offer more sophisticated tariffs than simple PUPM because the industry is so competitive, and thus intelligent and clever bundling of services (SMS and MMS and call rates and roaming charges etc) is critical. A SaaS provider can by contrast just use PUPM, because a vast range of software functions can be offered, and that is the basis for competing, rather than the tariff rates. If you are inclined to that view, then I suspect someone like Nick Carr would disagree with you since he argues that software is rapidly becoming a commodity – like telephones – and that competition based on functionality alone is insufficient, since most software packages of a given genre do pretty much the same thing anyway. Maybe the 80/20 rule also applies too: 80% of the customers in practice only use the same 20% of the functionality available, and are actually pretty happy with it.

I believe that raw PUPM is not good for many customers, and its also not good for a SaaS provider. It is not good for many customers, because the intensity of usage of a SaaS offering may vary considerably across different grades of staff (line staff, field staff, mid-tier managers, and executives); at different time periods (end of quarter and end of year may drive more activity than at other periods); and in different functions (for example marketing campaign vs. pre-sales and lead qualification vs. negotiation and sales vs. customer-support vs. solution fulfillment). The same considerations also make PUPM bad for a SaaS provider: I strongly suspect many SaaS providers in practice leave money on the table as they negotiate under pressure to discount on their standard PUPM rate under such circumstances.

In the telecommunications industry, billing and rating engines are used to capture and describe an operator’s tariff policies, as a series of rules applied to traffic activity records in the network. There would seem an opportunity to apply the same general approach to the SaaS business model: can a SaaS provider use billing and rating technology applied to software activity records (business transactions, business events, session activity, etc) to implement a more sophisticated tariff policy than naïve PUPM ?

In fact, that is precisely what LeCayla does, and that is why I was delighted when I was asked to take on the non executive Chairman role last November. In my view SaaS providers are increasingly unlikely to be able to compete on functionality alone – such software is becoming a commodity, in Nick Carr’s terms.

As Conor Halpin, the LeCayla founder and CEO says: “If you want to maximise the value from each customer then you need a pricing model that captures marginal value.”

The ability to compete on tariff structures - by understanding what customers usage patterns actually are, and devising and applying tariff plans based on business events and activity - is likely to be the basis for a successful SaaS provider, thus increasing value to customers, and increasing market share.

After all, mobile operators have honed their skills to a fine art. They understand what customers usage patterns actually are, and have devised and applied tariff plans based on customers events and activity, thus increasing value to customers, and increasing market share. They understand how much we talk here in Ireland.


The Lal said...

the simple PUPM makes sense in the consumer and low-end SME/Soho market and I believe is the reason a lot of these (re-emerged ASPs) models work.

The more complex pricing (per module/feature etc..) is likely 2 make more sense in a larger enterprise but don't underestimate the surge in the new hosted model (and the simplicity it brings via PUPM)
I've had the hassle many times of trying 2 figure out "true" charges for large and complex software/data centre systems and often its someones FT job (and unnecessarily so).

My 2c

keithbohanna said...


have you come across examples/stats showing the adoption of SaaS by consumers or on the associated business models - recurring subscriptions?

I hope LaCayla is going well for ye - Conor is a good bloke.

thanks, keith