War of the worlds

War of the worlds

Summary: As if the software world was not facing enough threats, the hardware world now threatens to undermine software business models left and right. New chip architectures and other hardware trends now raise the prospect of software industry upheaval.

TOPICS: Tech Industry

As if the software world was not facing enough threats, the hardware world now threatens to undermine software business models left and right. New chip architectures and other hardware trends now raise the prospect of software industry upheaval. "The real losers," according to a recent piece in The Economist, could be companies such as Oracle, SAP and IBM, "whose industrial strength programs are the bedrock of business."

It is not just SOA that threatens established business models -- and is forcing software titans to reinvent themselves (a la SAP's "NetWeaver," Oracle's "Fusion" and and assorted other SOA initiatives that vendors have announced).

Turns out, dual-core processing platforms from the likes of AMD and Intel promise to create a wave of dilemmas for software companies. Dual-core processing dramatically expands the power of the chip (enabling it to perform like 1.3-1.8 single-core processors). That leaves software companies feeling that customers are getting a "free ride." Which raises an important question: Will software firms try to charge more for the enhanced core (and risk the customer's wrath) or must they continue charging for the processor (and, potentially, sacrifice a great deal of revenue)? While Oracle is apparently upping the price at this point, IBM and Microsoft have sought temporary market advantage by promising that prices won't change.

Then, there's "virtualization," "partitioning" and "rapid provisioning" -- trends that threaten to further erode conventional software business models. "While virtualization may be great for hardware ROI, it wreaks havoc on software licensing policies," the publication states.

Some folks believe subscription-based (Software-as-a-Service) models hold the answers for an industry that is struggling to put a price on value. However, recent research suggests that buyers "overwhelmingly" prefer single, one-time payment methods (per-user or per-processor) -- a surprising discovery given all the complaints in recent years about being gouged by vendors.   

Could it be that software-based value models are all doomed? Could it be that software companies will have to begin charging for -- dare I say it -- business results? That would require them to develop new results-driven, business models that involve the sharing of risk and reward with their clients. New SOA and open source approaches might enable them to aggregate existing software capabilities rather than pouring all their capital into proprietary ones. If that happens, one wonders if we will even continue to call them software companies. What do you think?

Topic: Tech Industry

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • IBM

    IBM has begun this migration! Their "Global Services" group is getting more and more of the business, while they slowly get out of the OS business (Linux migration). They also are pushing middleware products like Websphere, which would be the basis for a SOA strategy. In the end they will sell the hardware, give away the OS, and charge you for management and support functions of your software stack. They don't HAVE to make money on DB2 . . .
    Roger Ramjet
  • That would be a big change

    Given that every EULA on the planet (a slight exaggeration...) is an exercise in deflecting responsiblity, the shift in mindset this would require might very well cause executives' heads to explode.

    But, anything that pushes software companies to actually accept some responsiblity for how their software performs (whether in the "benchmark" sense or in the "delivers value" sense) is a big win for the businesses that use the software.

    But I wouldn't hold my breath waiting for this to happen...
  • changing times

    the software industry was concieved 20 years ago, which in computer years might as well be 200 years ago. It goes without saying that something new will come to replace it. With the present state of the internet, I doubt that consumers will give up standalone capability for SOA since it depends on bandwith and trust issues that have not been resolved. The internet needs to undergo a changeover to some new situation, perhaps even beyond the capabilities of internet vII.

    As for charging "by the processor core" or whatever ridiculous means the manufacturers want to try and sell new computers by, I think that in coming years specifications for performance will become more relevent as more computer saavy consumers spend on the next round of computing power.

    the clustering power of multi-core processors, further innovations from the graphics side, and higher available bandwith are threatening to obsolete most legacy software, but as a consumer I would want to pay for software and hardware as a fixed cost and not as a "pay as you go" scheme, since I have no way of containing the ultimate cost to myself (much like cell phone pricing at first)

    I believe that modular software with some sort of open sourced base will become the norm, with algorithms, functions, subroutines all available as comodoties (think all that patented technology) allowing groups (or individuals) to create custom "user environments" by purchasing the components and assembling them, but on a much more granular level than is done now (except by programmers and IT professionals), much like custom cars, aided by software that checks and debugs configurations.

    I'm sure that vendors will fill voids by creating various configurations that my become "the norm" but doing so creates a larger risk of tampering by hackers. Diversity is our best friend in the struggle with thieves (if every home lock was the same how long would it take to make a set of keys?)

    As more intellegent software becomes available to "create software" (environments) IT professionals may have to find a different niche. I believe every home will eventually have it's own webserver/website to aggregate content created by cell phone/pda/notebook (and probably share files as well)and with the increased "mainframe" capabilities of the home computer thin clients for the home will also become commonplace finding their way into toasters, refrigerators, and all those failed "internet appliances" able to be controlled from a common point.

    This of course leading to food creation robots, localized manufacturing of various small items that are made of plastic and be recycled in the home, and of course, larger robots for protection, housework, child rearing, and any other function that we find too taxing to deal with.
    • Techno competence

      Do you really believe that the vast majority of the world's population knows anything at all about how computers work and what goes into a program? It is very wonderful that the software industry does not pander exclusively to the lowest common denominator of technical expertise, as many other areas of commerce seem to, but at which percentile of the techno-savvy curve is the mode of available cash to purchase your dream toys? Also, how much of the real motivation to develop a revolutionary creation in software processes comes from making another buck, and how much from the joy of having your thoughts spring to life? Which kind of programmer do you need to hire? Which kind of customer is more rewarding to sell to?
  • A pricing model where customer and vendor share benefits

    Customers don't buy computing cycles or processing power and aren't even interested in how many user seats.

    Customers use computers to complete business transactions.

    A price per successfully completed business transaction would work for the benefit of both the customer and the supplier.

    Obviously the more transactions are completed successfully the more revenue the customer makes. The supplier also makes more revenue and shares in the customer's success. If the quality of the software goes down then the customer doesn't pay for the bad transactions (wrong invoices or downtime that causes the customer to lose business).

    The pricing by processing cycle encourages the supplier to be as inefficient as possible. Pricing by processor leads to suppliers overspecifying. Price per user seat presumably encourages the supplier to make their software as unintegrated and unfriendly as possible so the customer needs more people to get the job done.

    The price per business transaction locks the customer and supplier in a relationship of mutual benefit where the success of the supplier is intimately connected to the success of the customer.
    • RE: customer and vendor share benefits

      Certainly, there is value in the idea, especially if the up-front cost is very low. But I really don't see it as being a "silver bullet" for this problem.

      For example, a transaction fee could encourage the customer to artificially minimize the number of transactions or handle larger ones offline (if the fee is based on the monetary value). This naturally defeats much of the value of the software.

      In addition, how do you value a transaction? $5.00 flat charge? One percent of the transaction value? If a supplier gets this involved in the guts of a business, it seems like there could be an awful lot of negotiation over this, potentially poisoning the relationship.

      For example, do I really want my cell phone company to be "intimately connected" with the success of my phone calls? If I make a multi-million-dollar deal in a five-minute phone call, should I pay for 5 minutes of air time, or 1% of the deal?
      • There are no silver bullets

        but my point is that the per business transaction method reflects more accurately what a business is actually buying in terms of IT services rather than processing cycles, number of users or connection times.

        Sure there is a lot of room for negotiation, but this would be the case in any large contract. I think this method is beneficial in that it ties the supplier and customer together in having a mutual interest in the long term success of the customer.

        No, I don't think your cell phone company should have any interest in the content of your calls, but then you are quite free not to accept such contracts as you wish.
  • Software's Business Model

    I'm not sure that dual-core processors are really that important in the big picture. Software's inflection point might also arrive on 64-bit or fiber or virtual storage, or...

    The software companies themselves have recognized the need to address their business models. They did not invent the Software-as-a-Service concept, it was the Application Service Providers' (ASP) invention, but they were very quick to jump on the bandwagon.

    Also, witness the spagetti that most software companies have made of their licensing regimes. They are trying to meet the current demands of customers (one-off payments) against a need to introduce different pricing models in the near future.

    It seems to me that the software industry has to balance two models:
    - End-User Licenses; and
    - Service-User (ASP) Licenses.

    The way in which End-User Licenses are calculated is not really important - it just has to be onnerous enough to create a commercial incentive for end users to consider software as a service.

    The Service-User License (the license terms between software writers and ASPs) can more easily be measured by subscriber numbers, time used, facilities and functions used, data volumes, etc.. Also, ASPs have the infrastructure in place for these things to be measured easily and cost effectively.

    This needs more co-operation between software and service providers which is nearly always lacking. Software companies have an incentive to do this - though they often appear blind to that incentive. The incentive is that by developing a Service-License software companies will:
    - Promote their own software-as-a-service businesses;
    - Base usage on commercially measurable criteria, allowing the business value to be measured (and calculated within a business plan);
    - Simplify the base for usage calculations; and (in many cases);
    - Create a commercial layer of abstraction (the ASP - even if that ASP is another internal division of the same company) between softare as a block of functions, and software as an integral part of a customer's business processes.

    I believe this blindness comes from familiarity with the license model, and fear of change. Account Managers and customers of software have a perverse incentive to stick with licenses because of the continuous negotiation involved. Thus, most software companies' boards hear (from their sales and marketing people) of their struggles to maintain license agreements.

    However, senior software execs ought to be able to rise above this daily 'noise' and see the bigger picture. If they do, then it is not often obvious. There may be a fear of current-customer alienation here. I don't know why software companies feel they can't do this sales job - but it seems clear that they are unable to change quickly.

    ASP, Web Services, Industry (or Community) data standardization drives that result in XML documents (and/or schema), middleware developments, open standards, and Services Oriented Architectures all favor commercal terms for software where charges reflect actual use - because actual use is directly related to business value and all of these industry trends are aligning ICT to businesses needs.

    If software companies do not get this sorted very soon then they will feel even more heat from Open Source (OS). Because OS does not 'front load' business plans with big payments in the first year, it is easy to see why CIOs prefer to run new ICT projects on OS. But, think about this, using OS typically requires higher skill levels of internal staff, more contractors, or both. This is the basis of the ongoing debate over TCO models.

    However the current TCO debate, as such, is not of interest here. Rather, think about what happens when you employ an ASP. You employ people from outside to provide ICT infrastructure (incl. software) which is tailored to your needs and charged for on a delivery to schedule (at a flatter, and more predictable, rate than the buy-big-software-license-and-implement model), and it has easier to reach project-tune and project-off switches.

    Perhaps this is why patents have become such a big issue for Big Software. They cannot change course (change their license regimes) fast enough to favor S-a-a-S - where ASPs will standardize niche markets on the software that is developed fast and supported consistently and in depth (keeping the ASPs costs down). So, they have tried to slow down the pace of innovation to create the breathing space for Service-Licenses to get traction?

    Whatever the strategy being employed, it isn't working. Big Software has never looked more vulnerable than now (as the Economist recognized). They are still only half way, if that, to the dual approach I have suggested is their best future - and I can't see an alternative being touted anywhere...
    Stephen Wheeler
    • Who buys software?

      What about consumer-personal use purchases. I havent made a penny in about twelve years, but occasionally buy software for fun and life management. I also get Open Source stuff because the price fits my pocketbook. Microsoft began to a large extent by selling to technically not-so-competent consumers. Software companies seem to be mesmerized (hypnotized) by the big chunks of money that sales of commercial packages offer per transaction. Is the consumer market going to become completely Open Source, and how does that contribute to the ability of future employees to operate software setups with large annual prive tags, one-shot or monthly, etc?