The subscription yoke

2402 0


Originally, we were promised that the move from perpetual licences to subscription would lower the cost of software, be easier to manage and give more regular updates. But for Autodesk customers it hasn’t really turned out that way, as Martyn Day reports

Historically speaking, purchasing products like AutoCAD was akin to joining a club. Buying a perpetual licence for several thousand pounds (or dollars) would give you your own copy of the lingua franca of the CAD world. Multiple floppy disks came in a substantial box, along with a dongle (software lock), an overlay for a digitizer tablet and a pile of user manuals.

From 1988 to 2003, Autodesk was not hard-set on producing a yearly release, with some intervals being between one and three years. In that time, users would typically upgrade every other release and the company’s key marketing drive was about trying to get customers to upgrade with rafts of new features. Changes to the DWG format, which happened when some core new features were added, gave an added incentive to upgrade. Incompatibility between different versions made it harder to work with clients.

Autodesk then introduced the concept of what became known as an ‘obit’ (as in obituary), which meant that Autodesk would drop support to a version older than three releases back. If a customer still wanted updates and patches, they would have to upgrade. Letting a version of AutoCAD lapse into an ‘obit’ version meant there would be an additional fee to bring it up to date, and you might even completely lose the right to upgrade, having to acquire a whole new licence. Autodesk has always had ways and means to keep users reinvesting in its draughting tool, whether they want to, or not.

Moving to subscriptions

While the construction market may change its business model once in a millennia, the software industry is constantly evolving to get more money out of its customers. Software developers have become experts in ‘squeezing the lemon’. One of the most successful is Adobe, who moved from perpetual licences of Adobe Creative Suite to Adobe Creative Cloud.

Selling perpetual licences is an inherently ‘lumpy’ and unpredictable business, as customers are usually very satiated by competent, old versions of mature products. Adobe bit the bullet and went subscription only, which meant owning its ‘Creative’ software no longer required a heavy up-front price tag and could be paid for with relatively small monthly fees. Subscribers got access to the full suite of Adobe tools and all future updates.

Subscription also helped reduce pirate copies and increase revenues. The only downside was something called the ‘trough’, which is when revenues tank for a period due to the loss of big perpetual licence ‘joining fees’ but are then replaced by a torrent of income from monthly subscriptions. The subscription model also gave Adobe the additional benefit of its customers ending up with a higher cost of ownership, paying more compared to the old product upgrade lifecycles, where they might skip a release or three. The future renewal process also vastly reduced the cost of sales to Adobe. The software industry could only look on in awe.


Autodesk was most certainly a disciple, but the rewards would be potentially even higher. Unlike Adobe, Autodesk had a high-cost route to market, with distributors and Value Added Resellers (VARs) all taking a slice of the big ticket software sales. By moving to subscription Autodesk could get rid of the ‘lumpy’ upgrade revenue, lessen the piracy somewhat, lower marketing costs, and reduce the number of external distributors and VARs. Today, VARs now earn vastly reduced margins in return for ‘the right’ to service Autodesk’s clients. I recently met a VAR that makes £1 for every AutoCAD LT sold online.

Added to this subscription move, Autodesk increasingly went direct to large customers to put together massive Enterprise Licensing Agreements (typically lasting three years), cutting out its VARs again. The net result has been more money to Autodesk per sale. Pre Coronavirus it established a share price of over $200 (historically traded $40-$60 per share).

Product-wise Autodesk followed Adobe’s path and created ‘suites’ of products, as well as individual versions of popular tools (AutoCAD LT, AutoCAD, Revit etc.). Bundling the software in this way certainly appealed to customers. When it first launched its suites, Autodesk’s then CEO Carl Bass said “Ironically, we now make more money by giving more software away.”

Customer coercion

Since the 1990s, Autodesk has had great marketing, together with a vice like grip on its VAR channel. Most years, Autodesk would change the contracts with dealers, typically lowering margins and raising targets but also targeting specific products or markets.

As for customers, it has always deployed price coercion to either incentivise users to upgrade or upscale, or price disadvantaging those that upgrade on their own timescale. Pre-subscription, I’d regularly visit architectural practices that were contractually locked into an old version of AutoCAD (or AutoCAD AEC, ADT) for a three or four year project and I was shown the cupboard where they kept all the newer versions of AutoCAD which they had to buy to maintain their initial multi-thousand pound investment, but couldn’t currently upgrade within their project’s design lifecycle.

The net result is Autodesk has established a money-grabbing reputation amongst its customers, after years of price and upgrade deadline manipulation. The unwritten message to customers was keep current, keep investing. I would go so far as to say, too often in its history, Autodesk’s business model and financial aims have been more visible than the software development, on behalf of its customers, especially as products have matured.

As products age, there are less earth-shattering new features to add to them. They lose what I call ‘development velocity’. To some extent, 2D drawing is pretty much a done thing; it’s a commodity and has been for a long time. New features that Autodesk added to AutoCAD Release 14 in 1997, for example, would certainly have been more persuasive than anything added in the last few releases. This is probably one of the reasons why moving to a subscription model is good for developers as they don’t need to overcome customer inertia caused by feature satiation. But this isn’t so good for users as they are mainly paying to use the software as opposed to gaining an edge through innovation.

In theory, commodity products with a low development velocity should be relatively cheap to subscribe to, yet a single subscription for a seat of AutoCAD for one year is a staggering £1,890. In context, Adobe Creative Cloud is a professional suite of applications costing £50 a month (£596.33 a year) for a suite of 15 applications which all share the same interface and work together pretty seamlessly). Autodesk’s Construction Collection is £3,420 a year per seat.

However, the opposite is true for relatively new Autodesk products like Fusion 360, a 3D CAD/CAM/CAE tool used for product design and engineering. With stronger competition than it faces in the AEC market, there is much development work to be done and users are pushing Autodesk hard for new features. As Autodesk is the ‘underling’ taking on the market dominator, Solidworks, you also get a hell of a lot functionality for your £54 a month (£438 a year) subscription. This is undoubtedly bloody good value.

In the past, software companies had to create new versions packed with new features in order to entice their customers to part with their hard-earned cash. With subscription, software companies don’t have as big an incentive to do this, as they can just sit back and collect the money. This essentially means customers lose their right to vote with their money to upgrade in return for good quality, beneficial work.

Subscription is more about the right to temporarily use a software tool that may or may not be significantly upgraded over the term of the subscription. Autodesk tends to do new feature drops twice a year – September and March.

In the old days of purchasing a perpetual licence, the value was built into the original proposition. It was a one-time justification, it was an investment. The problem came when customers wanted to treat their software like an asset and, when needed, sell off excess licences. Autodesk did not like this. It was against the End User Licence Agreement (EULA) and the company tended to police sales of licences. There were also issues when a company was acquired or it changed its name. Even though the software cost two month’s wages, customers were never directly told they didn’t own the software, but they were buying the right to use it. When is an asset not an asset? When it’s software.

For subscription, the value proposition is constantly ongoing, usage and access are temporary and based on need. While the investment in end-user skills is a constant, the software is not a company asset but an operational cost. This does have advantages in the accounting department. However, in stopping sales of perpetual licences, Autodesk transitioned customers to pricing models that meant they would pay more to Autodesk over time than they averaged under the perpetual licences. The core benefits being: no big upfront fee thereby easing cashflow, access to the latest release, access to past versions, removing the need for big budgets in the future for whole company upgrades, plus all the development in the subscription time. This undoubtedly is a beneficial simplification for CAD managers having to manage licences and products over medium to large firms.

Changes to subscription

Prior to the Covid-19 crisis, Autodesk got in touch to describe forthcoming changes to its subscription licensing. Previously, Autodesk had promised that it would never take away perpetual licences from customers but this year the company has some bad news. Perpetual licences will lose the right to have maintenance. This might as well trap all the perpetual software licences Autodesk has ever sold in tree sap, rendering them amber-encased fossils in waiting.

Autodesk’s previous attempts to cost coerce hardcore perpetual licence owners to move to subscription and let go of their historic investments included punitively increasing annual maintenance fees by 5%, 10% and 20%, sequentially over the previous three years. This pushed up the cost of ownership in an attempt to make subscription look a tad more attractive. Now it will ‘obit’ future maintenance for these perpetual licences.

While Autodesk has kept its promise to not stop customers from keeping and using perpetual licences, they will have to do so without updates or upgrades. This is the point where we all do impressions of the ‘Monty Python Parrot Sketch’.

At the same time, Autodesk has decided it doesn’t like offering Network licences, where licences are pooled for dynamic allocation, so these are also going to be removed. In the future all software will be subscription and there will be no dynamic allocation of licences. Autodesk customers will need a licence per user. As Covid-19 started to bite, Autodesk delayed this happening to 7 August 2020 and moved the retirement to 7 August 2021. The company also enabled free commercial use of BIM 360 Docs, BIM 360 Design, Fusion 360, Fusion Team, AutoCAD Web and Mobile, and Shotgun until 31 May 2020, to all customers as part of its expanded access program.

Autodesk is also getting rid of ‘Serial Numbers’ and implementing ‘Named Users’. This means Autodesk wants personal logins for each user within a company. Autodesk promised me that it will not abuse the user data and operate under strict GDPR rules, globally. This potentially gives Autodesk direct access to all of its users, as opposed to perhaps knowing a contact in accounts that ‘owns’ the purchasing.

Once Autodesk has knowledge of every user, it can build a direct relationship model and possibly bypass its channel completely, reducing its cost of sale further and growing its business without increasing the overall sales volume. Based on this new granularity ‘new features’ are going to include learning each user’s techniques and offering training on features per individual.

The cost of new subscriptions has already gone up 5% but renewals have remained at the same prices as the previous year. Autodesk is also ending three year deal discounts and it put up the price of networked licences a whopping 33%.

The deal

Autodesk has an offer for Perpetual and Network licence customers that this change impacts. I am using AutoCAD here as a base product, but it applies throughout Autodesk’s product range. For perpetual users, Autodesk is offering an eight year deal in recognition of giving up a perpetual licence. Instead of the current full price subscription, pay the current maintenance fee with a 2.5% (compound) increase each year and thereafter pay the full subscription price. For Network licence users, Autodesk assumes that typically one licence was shared between two users so, for the next eight years, customers will get two individual licenses at 50% of the yearly maintenance fee for a single licence, each with a 2.5% increase each year.

Given this is going to happen and customers either stay stuck in 2020 (not a good year so far) with the current incarnation or migrate to the new deals, some cursory maths would indicate that over the eight year period both offers provide considerable saving over the subscription prices. AutoCAD subscribers would typically save £5,000 per licence and Network AutoCAD customers would save £14,000, so long as only two users were sharing that licence. It’s actually a pretty good deal based on subscription cost. However, the big question is, how comfortable do customers feel giving up their perpetual licences and giving complete control to Autodesk?

There is a reason why all these customers have held on for so long. There is a fundamental issue here and I am not sure Autodesk is listening. It’s hard to remain in control of your budgeting and costs when the pricing and licence flexibility is decided by a company that seems to be committed to driving up cost of ownership and does little to improve the value proposition. Outside of AutoCAD, every Enterprise Licence Agreement holder I have spoken too has complained about the huge increases they have faced in their last negotiations, while the development velocity of their core BIM product Revit appears to be languishing vs Autodesk’s heavy investment and work on its Construction Cloud. There is a widespread revaluation of design technology underway, with firms looking at their budget and what capabilities they need and where they want to be next. Post virus, this will be even more prescient.

Software debt

With subscription now at the core of so many software firms, AEC companies are now managing to accrue considerable monthly, yearly or more expenditure. For front-end authoring, to back office IT, the complete tech stack is now paid for with a mass of ongoing subscriptions. Many of these are in forward looking contracts and is actually a ‘software debt’ that needs to be built into their own business models. In some ways, it’s become similar to termed car leasing, and this nullifies one of the key benefits – should a company need to scale down, it still has to service the software lease.

I wonder how many firms build in this forward-looking software debt? Profit margins need to take account of this and profit margins have been trending down in the construction industry. Those using token-based systems are asked to predict usage three years into the future and get penalised for getting it wrong. It’s an impossible task. Software firms have benefitted from eradication of distribution and minimising VAR margins.

Now subscriptions are being jacked up, software development has slowed, token users asked impossible long-term questions and financially penalised if they underestimate. We now have IT directors and CAD managers worried about the cost of actually using their design system of choice, concerned about a lack of clarity in their management of project costs. The alleged benefit of knowing your subscription costs, or having a three year deal, is only the start of estimating a company’s software debt. The minds of IT Directors are now being focused even more. Concerned about a serious post Covid-19 recession with the cancellation of projects, potential layoffs and pressures on margins, those who recently signed three year contracts may be seen as a source of regret, committing to software usage and costs over and above any actual usage or staffing levels.

Alternative arrangements

With escalating prices and increasing cost of ownership, customers of all professional-level tools are looking at alternatives. As mentioned earlier, companies like Adobe have moved from desktop to cloud and dictated subscription to its customer base. For large firms, giving all designers Adobe Creative Cloud adds up to a significant cost. This has enabled an opening for competitors such as Serif’s Affinity products with Image editing (Designer), photo editing (Photo) and layout (Publisher) to come in and compete on price, at £23.99 each for perpetual licences. Firms that become the ‘de facto standard’, which charge a premium and fail to build happy communities by adding value can get into trouble.

The more I look at the design technology landscape, the more I wonder why the hell 2D draughting still costs so much. We are all 3D users. 2D may well be an output but the value is in making model definitions. AutoCAD is the brand to beat and to a degree AutoCAD LT has fended off the competitors, but in a subscription world the price differential could not be more stark. There are a number of alternatives that offer professional capability at drastically reduced rates per seat.

Graebert: Based in Berlin, Graebert is probably a company that many people don’t know. It’s the oldest AutoCAD clone developer and, having mastered DWG ages ago, has gone way beyond providing the capabilities of standard AutoCAD. Its core platform, ‘Ares’, enables 2D drawing on Windows, Mac, Linux, Android and iOS (smartphone and tablet) as well as in a browser. The web and mobile versions contain 1/3 to 1/2 of the features of the full desktop flavour. This is a unique achievement. It opens up all sorts of new possibilities for collaboration and the system is joined up with an Amazon AWS backbone. Meanwhile, Autodesk’s development of its online version of AutoCAD appears to have stalled.

Ares Trinity (Desktop, Mobile, Web) is just €250 per year or €695 for three years. The company also does Flex licensing for networked users. Currently access is free till the end of May, to support home working. A single three year subscription to Ares Trinity would save £4,497 on the current advertised price of an AutoCAD subscription. If you want a perpetual licence of Ares Trinity, it’s a one-off payment of €795.

Graebert provides its technology to numerous software developers and underpins many professional applications. The cloud-based product design application, Onshape, for instance, utilises Graebert’s engine for all drafting. CorelCAD is also based on its technology.

The company’s biggest OEM success is Draftsight. When Solidworks (owned by Dassault Systèmes) was relatively new in the manufacturing market, it discovered that many of its customers were still maintaining their seats of AutoCAD for editing and creating drawings. Solidworks cut a deal with Graebert where it licensed a 2D CAD system, which it called DraftSight and then literally gave it away to customers and anyone who wanted a copy to ‘monkey’ with Autodesk in shared customers. This product is now Dassault Systèmes’ biggest volume product, with millions of downloads and customers worldwide.

The software is underpinned by Graebert’s continued development but is no longer free. Dassault now charges £79 (Standard), £159 (Professional), £399 (Premium) per year for various flavours, running on Windows or Mac. A single three year subscription to Draftsight Premium would save £3,909 on the current advertised price of an AutoCAD subscription.

Bricsys, the Ghent developer of the DWG-based BricsCAD, now in the Hexagon group of companies, is another competent source of 2D and 3D design tools. While the company has historically been concerned with developing its DWG base platform, it recently turned into a high velocity developer of full on BIM, civils / Infrastructure and 3D MCAD. The company’s core advantage is that all of this is done on the same platform – LISP/ ARx API, 2D/3D constraints, ACIS Solids, rendering, deformable modelling, Tin surfaces, sheet metal and BIM, all in a DWG compatible design tool with the same interface. The company also offers a collaborative portal called 24/7, which provides document management, versioning, viewing and reports. It costs £175 to £525 for 20 users with 10GB or 30GB of online storage.

Prices start at £320 per year for entry level ‘Classic’ drafting and range up to £1,280 a year for the Ultimate 3D MCAD/ 3D BIM/ Drafting. The very interesting BIM flavour, which we have reviewed several times, is £1,130 a year. A single seat, three year subscription to Bricsys Pro would save £3,676 on the current advertised price of an AutoCAD subscription. A perpetual licence of Bricsys Pro is a one-off payment of £910.


This article has largely focused on 2D CAD, as it’s become such a commodity and offers the biggest potential savings for AEC firms. However, I have also heard from many architectural firms that are re-evaluating their core BIM tools. Software developers including Graphisoft (ArchiCAD), Vectorworks, Allplan and Bricsys (BricsCAD BIM) are sensing a big opportunity to pick up business.

While Revit is widely recognised to have lost development velocity, and there are potentially big cost savings to be made, swapping out a BIM tool is not as easy as changing a 2D CAD system. BIM workflows are not easily replaced and firms also need to consider deliverables, collaboration, contracts, training and third-party tools. This is something we will return to in future editions.


There is no arguing that DWG and 2D drafting is a commodity. DWG is a known entity and pretty much every design tool has the ability to read/write it fairly competently. The question is, why should it cost so much? If we are all investing in BIM, and costs are rising, can savings be made on software that offers the least bang per buck?

One of the defensive arguments against any of the DWG alternatives is that clones are ‘not professional grade’. I have heard this time and again. The hard reality is that drafting technologies from developers like Graebert are not only used by major companies but underpin platforms like Onshape and, especially, Solidworks / Dassault Systèmes who are global leaders in manufacturing.

Japan seems to be the one of the core territories that is willing to question its reliance on historical products such as AutoCAD. Having talked with Graebert and Bricsys, Japanese firms have been first to evaluate and replace seats of AutoCAD with their DWG-based tools. CTO Robert Graebert, explained to me how they would get regular updates from their customers such as Takenaka Corporation (£10 billion annual revenue) to tell them how their software benchmarks against its competitors on a feature by feature basis, with directions as to what needs to be improved. The Japanese take benchmarking to a whole new level. Bricsys has Mazda and Takamiya and Nippon Steel.

Returning to the whole issue of subscription, it all boils down to value for money and necessity. Autodesk is making the final move to kill historic perpetual licences and, to be fair, is giving a big discount for that. But the level of unease I sense within its customer base is palpable. The cost of ownership is escalating, irrespective of profitability within the industry. Doing things like jacking up prices 33% out of the blue, on one way of accessing licences (subscription networked licences) and then planning to remove them as an option, just adds to the narrative that the company has established and feeds fear amongst users that feel they have no control. Aggressive sales and price hikes for three year deal Enterprise customers have also spooked longstanding customers.

Autodesk needs to do much better in providing value for money for users of mature products and has to do something it has rarely achieved, by building community and working out how to improve customer relations beyond board level. Up till now, customers have been reticent to move elsewhere but in 30 years of covering this industry, I have never seen so many customers seriously evaluate competitive replacement software for 2D drawing, BIM and collaboration. Subscription has to work for everybody.

Customers have options. Making a budget stretch, it is not necessary to buy all the software and services from one expensive vendor. Savings can be made by opting for lower cost commodity software for generic tasks such as 2D drafting, while spending their budgets on the high-end BIM tools. This would be akin to going to Lidl or Aldi supermarket for the basics, like the very topical toilet roll and Marks and Spencer for the quality items. Also, expect to see more workflows around products such as Rhino and its ecosystem.


I began writing this just as Covid-19 was starting to appear to be a bit of an issue on the horizon. Within a matter of weeks, we have seen a deeply saddening human crisis, along with a combined shock to all of the economies of the world, which nobody has ever witnessed before. As we all lock down and try and work from home, it’s highlighting how cloud-connected working and the new breed of cloud-ready tools can go from being ‘a nice to have’ for occasional times to becoming absolutely mission critical.

The second thing that comes to mind is the flexibility of subscription in times of trouble. Those paying monthly could just stop, cancel their subscription and restart in better times. Here, a historic perpetual licence is really just as flexible. Those on three yearly enterprise subscription deals will not benefit from that part of subscription as it’s a future software debt. But those on a token system, will keep their pre-paid tokens for usage when things hopefully return to normal.

And finally, in the last downturn, 2008, Autodesk’s revenue dropped from $2.35 billion to $1.65 billion, as users stopped upgrading and buying new seats. It is this data-point that drove them to shift their business model more to subscription, so they could flatten and slow the revenue reduction curve.

The AIA report reminded us this week that in 2008-9, US firms reduced staff by as much as 60%. Already since the crisis, US architects are reporting a 50% decrease in new design project enquires during March.

It’s highly-likely that subscription may flatten this curve for a period, but it is inevitable that firms will utilise the subscription contracts to shed costs associated with fewer designers being employed. What Autodesk will be hoping is that the revenue curve will be more ‘V Shaped’ than the ‘L shaped’ curve of 2008 to 2015, as firms find subscription easier to buy back into than buying new licences.

Subscription positives and negatives

Positives of subscription

• Lower cost of access vs traditional big upfront purchase price
• CAPEX (for perpetual) versus paying with OPEX (for subscriptions)
• Lower cost to expand usage and shrink usage • Access to previous versions
• Streamed updates
• Easier to account for software and services costs per-project
• Maintenance fees were a kind of a subscription anyway • Use software away from licensed premises

Negatives of subscriptions

• Once you stop paying, you no longer have software to open files
• Increased cost of ownership – over three years you will spend more than previous models
• If you cancel your subscription you have nothing left for that money
• One glitch and all customers get it
• If locked into a multi-year deal, flexibility of subscription is lost
• Software developer might not add features in a timely fashion
• Software developer can change its business model on a whim.

If you enjoyed this article, subscribe to our email newsletter or print / PDF magazine for FREE


Leave a comment