Where is Autodesk Heading?

by Martyn Day, CADserver, UK

A recently announced AutoCAD 2002 and a move towards a subscription-based model and general emphasis on the company’s marketing output, Martyn Day analyses the future direction of Autodesk, AutoCAD and its other products.

With reference to AutoCAD 2002, Autodesk’s move towards a subscription-based model and general emphasis on the company’s marketing output, Martyn Day analyses the future direction of Autodesk, AutoCAD and its other products.

Autodesk has long been accused of being a one product company. The sustained success of AutoCAD in the engineering and architectural markets has made it the envy of the design automation industry. Three million legal copies of AutoCAD in service (Autodesk’s figures) have created billions of archived drawings in the company’s proprietary DWG format.

AutoCAD doesn’t come cheap, yet the several thousand dollars per copy fee did little to deter AutoCAD from becoming the de facto standard for a technical design, culminating in the seminal release of AutoCAD R12. At the same time, Autodesk realised that as its product rapidly replaced the world’s drawing boards, it was also saturating the market. The company started to dabble with some diverse development ideas like 2D Animation software and 3D Virtual Reality. Already doing well with the high value, low-volume AutoCAD, Autodesk was also keen to dominate the high volume-end market. This led to the purchase of a budget CAD solution called Generic CADD. Unfortunately Generic CADD didn’t really integrate particularly well with AutoCAD and its file format, so eventually Autodesk shelved it and developed a cut down version of AutoCAD, called LT.

Desktops

Autodesk then reorganised into veritical market groups. The idea being that by using AutoCAD as the core engine, Autodesk development teams could build market-specific 2D and 3D solutions. These various Desktops, as they came to be called, could be considered “AutoCAD on steroids” and were tipped to dramatically increase the benefits of CAD within mechanical, AEC and mapping industries. The first Desktop offering out of the stable was Mechanical Desktop (MDT), which utilized the ACIS solid modelling engine and the object-based ARX development language. This was followed by Land Development Desktop (terrain modelling), Architectural Desktop (Architecture Engineering Construction) as well as several enhanced vertical versions of AutoCAD, namely Map and Mechanical.

The problems and lessons learnt in the development of MDT on top of the problematic R13 may have been the reason for Inventor - in my eyes, the first major new code stream from Autodesk since AutoCAD. The main contenders in the MCAD (Mechanical CAD) field all had pure 3D interfaces, while AutoCAD was very much 2D and 3D enhancements and, in fact, it was sold as a migratory solution for those that wanted to transition to 3D. With an unstable revision of AutoCAD to build on and quickly advancing competition (notably SolidWorks) , one can imagine the pressures on the Autodesk MCAD division’s VP, Dominic Gallello, who decided that a new ‘from scratch’ modeller should be developed.

To date, Mechanical Desktop has the largest installed base of any of the mid-range modellers and its continued success has even astounded its creators who now have to try to move the MDT users to the more than capable Inventor. Autodesk’s marketing of MDT has declined and future roadmaps of releases, past the next release (V6), are hard to come by.

Architectural Desktop (ADT) owes its heritage from Autodesk’s purchase of the US developed Softdesk product Auto Architect. One of the earliest object-based developments on top of AutoCAD, Softdesk’s solution managed to get a dominating position within the US market. Here in the UK we had our own longstanding development, AutoCAD AEC, but this was shelved in favour of ADT. While ADT is selling well, it’s still not clear if customers are using its Single Building Model capabilities in anger. My experiences with ADT have left me with the distinct feeling that I’m wading through treacle and I wondered if the AEC division was going to develop its own ‘Inventor’ style product and replace ADT?

“My experiences with ADT have left me with the distinct feeling that I’m wading through treacle…”

A year ago Autodesk previewed some technology which it called Project Nora, which became known as StudioDesk, a 2D/3D modelling package that was extremely easy to use and not based on AutoCAD. Developed by Autodesk’s CTO who thought up Inventor, Scott Borduin, StudioDesk appeared to be a revolutionary conceptual modelling tool. Now named Architectural Studio, the product is slated for launch later this year. Whether if it will replace ADT at some point is still unclear. However, in an article on www.cadserver.co.uk (******************search the feature section for The Technology Odyssey ************), he commented:

“In the future, ADT will be flexibly linked to a distributed and shared project-based building model database. Other purpose-built solutions will be working on appropriate elements of the same project information system, providing secure access to the right information by the right people at the appropriate time.”

This concept of a centralised database crops up again later in this article and appears to be an undercurrent to Autodesk’s ‘behind the scenes’ developments. For now I conclude that Architectural Studio will be the package which will aid the conceptual and unconstrained part of the process, where ADT, or something like it, will be there for the detailing.

So the Autodesk suite offers LT for intermediate DWG compatible drafting, vanilla AutoCAD for customised 2D and 3D design, the Desktops for industry-specific AutoCAD-based solutions and now is in the process of evolving new 3D code streams -- Inventor for MCAD and Architectural Studio for AEC solutions. Currently the pressure within Autodesk is to push the Desktop products to its current and potential AutoCAD customers. Not only do they do more than vanilla AutoCAD, they also provide better profit margin to the company.

2002 and extensions

And so to the new release of AutoCAD, R2002. Autodesk had publicly stated “no more big Rs” (referring to their release nomenclature of R13, R14). In some ways, mainly with its distinct lack of content, 2002 fits that bill. AutoCAD will be enhanced gradually through the year via the release of a series of ‘Extensions’ and every year there will be a release made up of these Extensions plus any core changes to AutoCAD (an equivalent of a big ‘R’). These releases will have to be purchased. However, those who sign up to an Autodesk subscription will get all the Extensions for a fixed yearly fee. Up until now, a new revision of AutoCAD has appeared around every 18-22 months with AutoCAD and its dealers spending marketing money attempting to sell the upgrade to the installed base. Subscription plans would smooth out Autodesk’s earnings, making it more predictable for the stock market, reduce the efforts required to transition users to one common release and benefit the end user by delivering features over time rather than in one big lump – though potentially causing training headaches and deployment issues for large users.

Users who have paid money for AutoCAD have really bought a perpetual licence giving them the right to use that version of AutoCAD forever. Subscription won’t change that, in fact it will open up more options where it will be possible to rent or lease AutoCAD seats on demand. This option will help in those crunch times where more copies of AutoCAD are needed, coming in 30, 60, or 90 days usage. In the future Autodesk will also offer to host AutoCAD and deliver it as an Application Service Provider (ASP) over the Web, either from its servers or on yours. It won’t run on a local machine, just interact with the designer over the web.

There will be management tools available with the extensions later on, these will allow CAD managers to deploy the extra functions provided by the regular Extension updates over a large number of workstations.

In January Autodesk will officially cease support for the popular AutoCAD R14 and this will mean that if you fail to upgrade to a newer version, you will have to pay for a whole new copy of AutoCAD to get an update. It will also be deleted from the sales catalog so no new copies of R14 will be sold. In many respects the lacklustre feature set of 2002, combined with the unchanged DWG file format, aids that transition providing few barriers or problems should users want to upgrade.

It’s at this point that one decides to either pay for upgrades every 18 months or go on subscription. Autodesk’s stated intention is to get all of its customers on subscription eventually. However, to do this it has to provide compelling reasons for customers. While I can see the benefit for subscription for companies with lots of seats of AutoCAD, smaller companies tend not to find subscription as appealing.

”One of the best models for subscription is Bentley’s Select program”

One of the best models for subscription is Bentley’s Select program. Over 80% of Bentley customers paying for MicroStation updates on a subscription basis. Here extra functionality comes via the web or on CD (users can even compile their own suite of functions to be burned onto CD at Bentley and delivered to the company). On the technology side Bentley has certainly talked up what’s in R&D but has fallen short on several occasions of not delivering on its own deadlines - one of the problems with talking up future technology - but I’d rather know what was coming down the line if I was already paying for it. One difference between Bentley’s customer base and Autodesk’s is that Bentley has, on average, over 25 seats per site, while Autodesk has a figure closer to 5. It may be the case that Autodesk’s typical customer may not be as eager to subscribe as a typical Bentley one.

At the launch of 2002 I asked if Autodesk could tell me what new features they were working on that will be streamed through Extensions over the next year or so. I was told to look back at what Autodesk had provided in previous releases.

While R14 became popular as it was the definitive Windows version of AutoCAD, 2000 and especially 2000i have added few useful features and enhancements. The majority has focused on the Internet and with the relatively poor acceptance of these versions it’s obvious that these features haven’t been compelling enough for AutoCAD customers. Which leads me to wonder just what else can be done to

AutoCAD in the coming years to make subscription worthwhile? According to Ralph Grabowski, Editor and Publisher of the Upfront ezine (www.upfrontezine.com), there has been a trend towards providing an ever decreasing number of new AutoCAD commands in releases since R14:

* Release 14 introduced 86 new commands (52 of which were "bonus commands"), but removed 11 commands. Net = 75.

* AutoCAD 2000 introduced 65 new commands, but removed 32 commands (many of which renamed or combined commands, like DDLModes). Net = 33.

* Between them, AutoCAD 2000i and 2002 introduced just 29 new commands, but removed 4 commands. Net = 25 over two releases.

Autodesk needs to provide a more robust vision of exactly where AutoCAD, as a product, is going. I want to know what features are in development? Could Autodesk spell out how specific future Extensions will enhance the productivity and ease the lives of designers? Before you ask a customer-base to pay in advance for new features, it’s only fair to give them some insight as to what they will get for their money. It seems odd that I can write more about how subscription will improve the business predictability and cash flow of Autodesk than I can write about the benefits of subscription to actual customers! Autodesk is a company that tells us more about investment opportunities with its stocks than sharing information on the technology path beyond the next release.

Goodbye DWG?

To get a vision of the future of Autodesk’s development, one has to read between the lines and extrapolate from conversations with key Autodesk personnel. The problem is mapping the current product portfolio to some of the inferred directions. From the Bernstein AEC article, there seems to be some work on a centralised data model and while talking with Scott Borduin, Autodesk’s CEO, on the subject of proprietary file formats, he made the following statement : “We are going to get to an era when the CAD application, this great big mammoth chunk of code, will interconnect with all of these new architectures and force you into a multi-tier paradigm with swappable components between these tiers and even from the application standpoint, you will have a lot smaller subsets of things going on – you won’t just fire up one CAD environment to add all of the information from the geometry to the detailing, to the cut lists and structural information. Every one of those will wind up being different small applications living at the top tier of all these architectures.”

In short this means that Borduin’s current view of the future of CAD is a central model database with small specialist programs, specific to certain market areas (architectural, services, land, QA, civils etc), accessing the information that is only applicable to their fields. All this will work over the Internet or intranets. If this is the real aim, then Autodesk has to work out a way to evolve AutoCAD and the Desktops into that type of solution, unless it chooses to deliver a ‘from scratch’ product like Inventor. One would assume that it would be easier to start with a clean slate than to tie the hands of the developers with legacy issues. My guess is at some point, perhaps as little as three years away, DWG and files in general may well become antiquated notions.

Conclusion

For Autodesk it’s clear that subscription is a global goal and the benefits that it will bring the company in predictable revenues will benefit it greatly in the coming years. To do this Autodesk has to switch to delivering piecemeal functionality improvements and deliver value throughout the year. An important part of this will be to engage its customers to keep them informed and implement their development wishes. I am sure that most of the Autodesk big cheeses can reel off a list of impressive corporate customers but very few AutoCAD customers that have five seats or less, yet it’s these that make up the lion’s share of the installed base. One other worry here is the US-centric mentality, which is still pervasive within the company. Many of its recent products start as US-only distribution, e.g. RedSpark, Buzzsaw, the Extensions and Streamline to name but a few. This makes me wonder just how much input the European part of the operation gets in these products? Perhaps the US is better ‘Internet prepared’ and more savvy in some of these areas but if the original target market of these products is the US, when does the European point of view get factored in?

I started off this article stating that Autodesk was perceived as a one product company and to a larger extent, irrespective of the Autodesk spin, the company is still a one product company. Vanilla AutoCAD is still providing a hefty amount of bottom line revenue, together with the AutoCAD code stream acting as the foundation for all its current Desktop and budget (LT) solutions. In a business sense, the proprietary AutoCAD DWG compatibility issue has been leveraged well throughout the product line and is utilised as a strong sales pitch. In the business world it’s not good to be seen as a one-product company, and while on the design and drafting side this is pretty much the case, the purchase of discreet (high-end Digital media, 3D Studio etc) did help diversify its portfolio.

Relatively recently, starting with the development and introduction of Inventor, Autodesk has embarked on a new journey, a place where its traditional comfort zone of DWG compatibility, armies of trained users and industry-wide familiarity cannot be played. The AutoCAD products lead to the development of the Desktops and while the foundation AutoCAD code has been redeveloped, there seems to be only so far you can really evolve a product. Applying modern programming techniques to redevelop a 19-year-old bit of software, while tying the hands of the developers to maintain legacy and historical functionality is like trying to re-build a bi-plane using the latest aerospace technology – whatever you do, it’s still going to be a bi-plane, albeit a jet powered one! For me, Inventor, an outstanding new product, is proof that starting from scratch is the way forward and reading between the lines, the forthcoming Architectural Studio looks to be the AEC’s equivalent to Inventor. It also seems Autodesk’s technologists are looking at ways of removing the limitations of ‘files’ in this interconnected and networked era, together with the centralisation of the design database. For Autodesk, the managing of this customer migration from its ‘legacy AutoCAD DWG’ world to the next generation vertical products will provide the biggest challenge.