The Increasing Cost of Ownership

By Martyn Day, editor, CADserver
September 9, 2003

Recently, I was asked to give a presentation to the Computer Suppliers Federation (CSF) here in the UK on pertinent trends in the CAD industry. While there are many trends to choose from, I decided to conduct a questionnaire of a sample of our mechanical CAD (MCAD) magazine readership, on a range of design software-related topics to get a quick “vox-pop.”

One of my questions was, “What’s the weakest part of your CAD system?” I had supplied a range of six possible answers: interoperability, 2D drawing layout, ease of use, choice of third-party applications, ability to handle large models and cost of ownership. I fully expected the traditional gripes of interoperability and large model speed to top the poll but they didn’t; the clear winner was “Cost of Ownership” taking 33% of the vote. This was followed by large model capability 27%, ease of use and choice of third-party applications around 16%, interoperability at 14% and the 2D issue coming in last with only 11%. It was clear that readers were concerned with how much their chosen CAD systems cost to run and maintain.

Unlike the majority of AEC customers, MCAD engineers are used to paying subscription along with the initial purchase fee of a 3D CAD system. Traditionally, this subscription for streamed new features was not sold optionally but was built into the original pitch. As the battle for the mid-range MCAD market heated up, products like SolidWorks and Solid Edge lowered both the entry point for 3D design and the typical annual maintenance cost. With Autodesk joining in the battle with Inventor, it produced even more aggressive pricing in the market. While this has been a tough lesson for existing high-end players like PTC (Parametric Technology Corporation, the creator of Pro/ENGINEER), which lost market share, there has also been a major influx of users moving from 2D to 3D who were unused to paying maintenance on top of an initial purchase price. The market dynamic means that high-end users can see these mid-range products are cheaper to run and are unhappy with the maintenance they currently pay, while some of the new users are not particularly happy about paying maintenance for software that is clearly still in development.

Moving away from the mechanical market, the CAD industry in general is moving from the box-selling volume model to a subscription-based one. While the high-end vendors, such as Dassault, EDS PLM and PTC, have run long-term maintenance (patches and service packs) and subscription programs (new functionality as well as patches and service packs), it's now the volume players’ turn to get in on the subscription act. The main reason for this is that the CAD software market has matured; most people that need it already have it, and those that have it are mainly looking to buy adjacent seats. The cost has gone up and the volume of sales is at best flat. Subscription is a way to get more revenue out of your existing users and have more direct contact with them, in return for streamed product improvement.

The biggest player in the volume market, Autodesk, is keen to move its enormous (4 million plus) user base over to subscription. At the moment users have upgraded their AutoCAD every other release, leading to revenue that comes in fits and spurts. By getting users onto a subscription for a yearly fee, they will be entitled to yearly wrap-up releases of functions. While the yearly fee will be lower than the cost of a single upgrade, if a high enough number of users sign on, then Autodesk’s revenues will increase; instead of buying one upgrade every four years (around $500), users will be paying twice as much in subscription revenue over the same period (4 x $250 = $1,000). To assist the subscription drive, Autodesk will be offering subscription deals including the use of other Autodesk applications as well as financial penalties for those who don’t opt for a subscription. In the past it has been possible to upgrade from a version that’s two or three releases old for pretty much the same price as those who have kept up to date. I understand the intention is to offer a more incremental costing to this, making subscription more appealing. The key factor here being that if Autodesk is successful in converting its user-base over to subscription, the company will see its income rocket.

There’s nothing wrong with a subscription-based model, so long as customers feel that they are getting value for their money. Those that already run subscription-based businesses, like Bentley, operate on very different criteria to those who survive on box sales. Subscription is all about catering to users’ needs and demands and supplying non-disruptive installments. To do this successfully software companies have to be close to their users to assess their likes and dislikes and must avoid forcing customers to adopt technology they don’t want or need. Another danger inherent in the subscription model is that users may perceive the subscription as too expensive and switch to another system.

Software is considered dead if it isn’t regularly updated. There appears to be three ways to satisfy this: incremental updates of existing features, introduction of individual new features, or adding bought-in technology. Although the first two methods have been the industry staples, it’s becoming increasingly popular to create bundles of core CAD and vertical applications (document management, piping, analysis tools, etc.) to create better ”value offerings” at a higher price.

Bundling is becoming the norm in the mechanical CAD world, with players like Autodesk and SolidWorks constantly on the lookout for small applications to purchase and add-in. The AEC industry is different, in that a single firm will rarely do all the design work for a project; instead, there are specific and separate products for each task, such as steelwork, concrete detailing, HVAC, site design and analysis. While there is some bundling of AEC functionality with the single building modelers (such as Revit, Architectural Desktop, ArchiCAD and TriForma), this market has yet to really take off and will probably move more towards purchasable modules.

In my view, although some of this bundled functionality is welcome, it can lead to a quantity - as opposed to a quality - approach to software design. Bugs that need fixing or features that need finishing don’t get addressed, as the concentration of the release focuses on broadening its capability into other vertical areas. Here, Revit comes to mind. Before Autodesk bought it, the company had only a few customers and they were dictating the development shots, so one early release featured scaffolding and cranes, when the software didn’t know what a lintel was!

It’s also hard to evaluate heavily bundled releases; SolidWorks Office and Inventor Professional are becoming the main “bundles” in the MCAD industry and while they may provide something extra, the real questions are how well do these extras perform and are they really necessary? The CAD software industry seems to be concentrating on developing “Professional” versions of established products; if only the new versions are “professional” one wonders what the previous versions were Autodesk, for one, appears to have upset many of its Inventor customers by quietly introducing a “Professional” bundled version. In time the software vendors expect that most users will migrate, upgrade or purchase these higher-end flavors.

The competition among CAD software vendors is heating up as they aggressively build up their offerings to outdo each other. This is leading to a more reactive market based on short-term views, deviating from any established long-term product vision. This unfortunately doesn’t relate to the state of the market, where existing customers are more important than new ones.

While I'm on the subject of continuous development, there comes a time when a software vendor creates a “Next Generation” of product, which in some way is fundamentally different or allegedly better. There is only so far one can modify a product's architecture through new code level or bundling vertical application sets. The typical lifespan of software architecture of a product is 6 to 10 years. While customers may feel that it's a natural continuation of the product they bought and subscribed to (for it usually shares the same name), they find that the software vendor sets a higher price tag. This also usually leads to legacy data and irregular end-user adoption.

Traditionally, there have been two routes to market: selling direct and selling through a dealer channel. Those that have tended to sell direct have produced the more expensive high-end products and the “volume” solutions have sold through distribution to dealers. As the market matured and new customers have become rarer, the vendors with dealer channels either killed them or decreased the amount of money they got for selling boxed items (margin). By lowering the margins to dealers, dealers can offer fewer discounts and therefore the street price rises.

The other consequence of lower margins is that fewer dealers can exist, again limiting the number of dealers in any one market and reducing the amount of competition, which impacts the street price. When the vendor realizes that it cannot cut the margin to dealers anymore without removing its route to market, the only other option is to raise the price of the software to the customer.

Direct sales have survived but mainly sell into existing large-seat corporate accounts. The fact that the sale is direct removes the losses incurred through using a distribution or dealer channel. However, due to the volume of seats sold, this can be heavily discounted.

The other possibility is that of retail, where software is purchased over the Web or from a brick-and-mortar store. While companies such as Autodesk have dominated this market with products like AutoCAD LT, the company now plans to increase the price of the software to place it well above the typical “retail” pricing. Over the past few years, LT has seen several price hikes, and it is rumored in the UK that LT could cost as much as £1,000 ($1,500) a seat by the end of the year. This is not the typical cost of a boxed software package available at your local CompUSA.

LT was originally released to fend off competition to Autodesk and AutoCAD; at around $750 it successfully fought off Visio’s IntelliCAD clone and is now one of Autodesk’s biggest volume products. It offers 80% of AutoCAD 2D functionality for 20% of the cost, and this issue has not been lost on customers looking for native DWG compatibility. During recent upgrade programs some AutoCAD customers chose to opt out of upgrading because for the same price they could purchase a copy of the current version of LT as a replacement and do so at their own pace, not Autodesk’s. So with LT pricing on the way up, LT is no longer performing the “sweeper” defensive role, leaving a pricing gap for a potential competitor.

All these issues add up to mean that the cost of owning CAD software is on the rise. Meanwhile, the cost of powerful hardware has substantially decreased. There was a time when the hardware cost more than the software. It’s now possible to buy two or three workstations for the price of one mid-range MCAD CAD license.

The lack of competition at the low end, the lock-in of proprietary formats, the reduction in dealer margins and dealers and the drive to subscription means that the annual outlay required to maintain existing investments, or add additional seats, will go up.

While in the past CAD companies have tended to compete on features and functions, today’s mature market looks set to change that dynamic towards one of cost of ownership. It’s like buying a nice sports car. In addition to the not-insubstantial purchase price, there’s the insurance, the maintenance and the fact that it only does 12 miles to the gallon. CAD systems have a joining fee, then there’s the yearly maintenance, the cost of hardware and training, the cost of support and the cost of upgrading (in downtime). The equation isn’t straightforward but I’d advise all engineering firms to make themselves very aware as to the total cost of their systems.

When the total cost is established, only then can you really appreciate the value you are getting for your money. Subscription makes payment for upgrades easier, but it also means you get more updates more frequently and only a percentage of these updates will be of any real use. Once begun, a subscription makes it notoriously difficult to temporarily retire seats; many vendors require back payment of subscription to re-activate seats.

From what I can see, the escalating cost of ownership is the result of a combination of factors, mainly to do with software vendors’ drive for revenue and increased profitability. It’s not because software is getting any more expensive to develop; Microsoft’s .NET language and skilled Indian/Russian programmers are, if anything, helping reduce the cost of development. Killing or reducing the channel lowers the software developers’ route to market and buying/bundling applications is cheaper than developing them in-house.

As the cost of finding new seat sales increases, as the competition becomes fiercer and as profit margins shrink, the obvious next move is to wring more money from each user, and more often. If the system’s new features aren’t enough to lure you to a subscription then the penalties for eschewing subscription might do the trick.

The smart users will be the ones committed to running a diverse portfolio of CAD products. Such diversity will help to protect you from the whims of the next revenue-generating business strategy. In some respects PLM and all the marketing hype around the benefits of system integration will lead to a reliance on the products and services of a key vendor. However, some companies, like Volkswagen, openly operate a mixed environment, to alleviate software vendor pressures and retain internal control of its design systems and integration. I think there's an old adage about eggs and baskets and not keeping them all in one place!

A final point on software pricing trends: as the cost of ownership increases, breathing space for developers and new solutions slowly open up in the “budget” section of the market - in our industry that's anything below about $1,400. At this price point, good products like Alibre Design, Robert McNeel & Associates’ Rhino, IMSI's TurboCAD, IntelliCAD, Gestel's solidThinking and IronCAD's Inovate, could all benefit from end-user attention seeking respite for the escalating costs of the traditional industry staples. While the market is mature and new users are hard to find, it doesn't necessarily mean that the current vendor market shares will stay at the present levels because the underlying trend in the industry has always been “more functionality for less.”
 

About the Author

Martyn Day is group editor of MCAD Magazine and AEC Magazine. For more information, visit the CADserver website.

More Select CAD Industry Articles