Skip to content

The Trouble with Traditional Software Pricing and Subscribing to Configuration

I’ve been thinking about the idea of charging for a subscription to “configuration.” Here, I mean “configuration” in the broadest scope possible: as in, “the settings that software uses to determine how it should run.” For example, in the area of:

  • Compliance: the configuration required to to tell the software how to enforce governance and what to do when non-compliant events are encountered.
  • Systems Management: monitoring thresholds, management reactions to bad states, and even process workflows that staff and systems would use. Collaborative Systems Management information too.
  • Development: best practices that were programatically enforceable.

On the software side, subscribing to configuration requires a platformy and network aware architecture. Your application needs to be platformy in the sense that how it runs must be programatically configurable. It needs to be network aware, ideally, so that configuration updates are automatically downloaded. Updates could be shipped on CD, DVD, or tape, but that’s probably only required for the most paranoid or tragically low-bandwidth shops.

No More Cash Upfront

While there are, no doubt, already people who are and have tried the above approach, using it as a revenue model for OSS approaches seems promising. In OSS — unless you’re dual-license — you’re dumping your up-front licensing take. This isn’t a problem, it just means that you have to figure out how to get the same amount of money (or more) by charging for something other than copying the software.

Those Damn Telcos

To this point, I like the tell people the cellphone service pricing anecdote (I don’t remember where I picked it up, but it’s not mine). I’ve had “wireless” (pagers and cellphones) service for pretty much the last 15 years. As I recall, the first monthly charge I paid was $20-30 for an alphanumeric pager. Now I pay around $100-120 a month for service. During that time, the carriers have offered more and more minutes and other services without cannibalizing their revenue. In fact, they’ve grown it. That is, the raw price per unit, so to speak, that carriers are charging has decreased while the amount I pay has increased.

Now, I’m no telco expert, so maybe I’m analogizing out of my back-side, but to me those two data points mean that the telcos companies have figured out how shield themselves from revenue loss as they decrease the raw value of their products and services: a 2006 minute of voice is worth much less than a 1996 minute of voice.

Zero Value Software, Service, For Pay Configuration, and Pricing

In essence, one effect of open sourcing your software is to set the price of the software at zero dollars. This is an even more dramatic version of the shrinking value of cellphone minutes. Indeed, to further analogize, your software becomes the cellphone that you give away for free because you know you’ll be able to suck $1200+ a year out of suckers …I mean customers…like me.

If you’re giving your software away, already selling support and maintenance, chances are, you still need to recoup that initial license sale you’re no longer getting. I’m highly skeptical that customers are going to be tricked into lumping that license price into their support and maintenance contracts: “wait, last year I paid $10,000 for the license and $5,000 for support…and now that you’ve OSS’ed it, I pay zero for the license and $15,000 for the support…uh…what?”

Tools Don’t Pay the Mortgage When the Tools are Free

This leaves customizing the software, or “service.” Many companies (like BEA and Sun) are servicing agnostic, while others (IBM, Accenture, JBoss) make loads of doe on services. If you’re going to be a commercial OSS vendor, it’s time to get religion and take services on whole-hog. Otherwise, there’s going to be a nasty drop in revenue, and you’ll just be handing those dollars over to consultants, SI’s, ISVs, and IGS.

The primary dirty secret of enterprise software is that the majority of your losses are incurred during support. Every vendor has a list of trouble customers that have long since over-stayed the welcome that the initial license and support contracts bought. In reality, servicing is what enterprise software has always been about: building software that 10-20 giant customers want and then figuring out how to sell it to the other customers.

Unfortunately, the methodologies to develop and support that software aren’t always geared towards that cycle. Instead of focusing on keeping the cost of the whole relationship with the customer at a minimum, most development methodologies focus on making the cost of that first copy as low as possible. Breaking that mind-set is a large part of what I’m hoping see emerge in Enterprise Agile; that’s some thinking for another post.

The point is this: lots of costly (to the vendor) service and support is the reality of enterprise software, but few vendors have oriented their process and pricing to take advantage of (make money from) that reality. Instead of wishfully thinking that the model is to make that first copy as cheap as possible, they should shift to a model that makes maintaining the paying relationship with their customers as cheap as possible. From there you get bigger margins, which means more profit.

This strategy demands new models for software, and subscribing to, or paying for configuration is one possible model.

Paying for Configuration

Paying configuration is another way to recoup the cash lost from dumping the up-front license. For example, you might have a COA driven product that provides all the services needed to monitor and enforce regulations, laws, and other governance. As the regulations, laws, and best practices for how to enforce those change each year, month, or day the vendor sells the appropriate configuration updates. Instead of having to re-tool their compliance infrastructure, the customers “simply” update the configuration.

In both cases, the customer pays the vendor. But in the configuration subscription case, there aren’t license charges. Because there aren’t license charges, whether the software is open source or not is irrelevant. Meaning, we’ve got a way to make money with an open source stack.

Essentially, this is what TurboTax does. They’re not open source, but the reason you buy their product each year (online or desktop) is because they’ve updated the tax code configuration and the associated workflows (new and better ways to pay less taxes, or “best practices). If those “configurations” didn’t change each year, TurboTax would be much less successful.

Selling How the Software Is Used

Enterprise software vendors have been climbing higher up the value-chain from their tools roots. Stories like BSM, BPO, ILM, and even development methodology families like Agile are more about people and process than the code that comes on a CD or from a download. SOA, of course, is the great emperor of this trend.

After all this positioning and actual work (the emperor is wearing cloths…though not always the exact outfit everyone thinks he is), most vendors are well poised to sell the streams of configuration required to deliver on those visions. They’ve primed people with the story that the exact software you use isn’t quite so important as the way you use the software: standards, right?

Companies who’ve built up experience in consulting and servicing, like IBM, are best positioned to sell configuration. On the other hand, people like BEA who’re quite proud of their service agnosticism could be well positioned from the other angle of the problem: their platforms have to be fleixble enough to be customized by their partners and ISVs, meaning they could end up being (near) flexible enough to support configuration subscriptions.

Software is, by nature, quite static: until true AI’s come about, people are required to update it and tell it what to do. More importantly, the “knowledge” of how to best use the software isn’t always in the hands of the end-users. So, that knowledge is something customers might pay for.

Disclaimer: IBM, BEA, and Sun are clients.

Technorati Tags: , , , ,

Categories: Enterprise Software, Ideas, Marketing, Programming.

Comment Feed

One Response

  1. “Now, I’m no telco expert, so maybe I’m analogizing out of my back-side, but to me those two data points mean that the telcos companies have figured out how shield themselves from revenue loss as they decrease the raw value of their products and services”

    That’s the Jonathan Schwartz model. A minute of CPU is worth less than it used to be, but we need lots more minutes. Maybe time to buy some of that SUNW.