Way back when, before most companies had computers of their own, computer service bureaus ruled.
If you wanted to do a statistical analysis, for example, it was either pen, paper, and slide rule (electronic calculators weren’t around yet), or send punched cards to the bureau, pay big bucks and
wait until a courier came back with a fat wad of continuous form printout. Oh, yes — and pray that your inputs were correct, because if something was wrong you’d repeat the process — and double the expense.
Then, users paid hundreds of dollars for a few minutes of processing time. Now, the same tasks can be accomplished by a modest personal computer, and cost almost nothing.
Yes, we now must licence the software —again, at a fraction of the price the old mainframe service bureaus paid — but any program that gets used a lot ends up almost free on a cost per minute basis.
Now vendors are again touting the notion of software as a service, as though it were a new concept. Guess what — it’s service bureaus, revisited.
That’s right – you will not buy a licence, and be able to use a piece of software as much as you want. Instead, you will purchase the use of it for a period of time (as a monthly subscription, for example), or perhaps, as with the mainframe bureaus, pay per minute of use.
Thanks to the Internet, we won’t have to ship decks of punched cards, or key in programs on a dumb terminal communicating with the mainframe over a pricey dial-up line. We’ll talk to Web hosts, which will probably accept our inputs in several convenient ways. We’ll be able to print output directly to local devices, and save files on our local hard drives.
The one thing that will be retro is the notion of paying by usage.
Software will cease to be a capital item and land on the expense side of the ledger. Users will no longer be able to simply double-click on a file and know it will open because they have a perpetual licence for the software that created it. Instead, it will depend on what usage they have purchased. I’ve seen several models suggested; one will put a viewer on the local machine so, for example, if you bought Microsoft Office as a service and decided not to renew the subscription, you would be able to open existing Word documents and read them, but not edit them or create new documents.
Some analysts suggest a subscription model would give software vendors an incentive to keep customers happy so they will renew. But unless files formats are readable by other products, customers are still stuck — if they want to perform a function, and use their existing files and data formats, they will have to renew at whatever price the vendor demands.
For some programs, this scheme may make sense. It will work for rarely-used software, especially specialized, expensive programs. Users who are loath to spend money on a perpetual licence when they only need the program for one project could use it short-term, earning the vendor revenue it otherwise would not have had, and letting the user get the necessary results without a huge cash outlay. That’s a win-win situation.
But consider what could happen if office suites or databases were sold as services. Instead of a one-time licence payment, companies would be continually pouring cash into vendors’ coffers. It would, to be sure, make vendor cash flow easier to predict, and vendor bottom lines happier places, but despite the pundits who insist customers are demanding software as a service, I’m not convinced there really are a lot of benefits for the user.