Autodesk’s recent admission that there’s been a multi-year lack of development of Revit because of the concentration on cloud and construction has left many frustrated architectural practices wondering what the future of this 20 year-old BIM tool will be. Martyn Day examines the tea leaves
Autodesk acquired Revit Technology Corporation (RTC) with its Revit application in 2002 for $133 million cash. While RTC made a lot of noise, and its core parametric technology was innovative, it did not generate much revenue and, in fact, had few customers. Up to that point Autodesk had been trying to morph AutoCAD into a full-on architectural modelling tool with the unfeasibly complicated ‘Architectural Desktop’ variant.
One of the underlying key driving factors to the acquisition was that five years previously Autodesk dropped out of buying mechanical CAD (MCAD) tool Solidworks (Dassault Systèmes paid $310 million) which then went on to dominate desktop MCAD as users moved from UNIX to Windows.
Autodesk’s CEO at the time, Carl Bass, didn’t want to make the same mistake again. Had Autodesk bought Solidworks, it would not have spent millions on developing Inventor and then Fusion and would have been the number one player in desktop MCAD. The Revit acquisition protected Autodesk dominance in AEC and eventually proved to be a savvy investment.
The excitement of innovative code and doing things much better than previous generations of software is typically reserved for the early phases of development. Updates can be frequent and big; the potential can be seen. However, this furious initial phase typically slows down as a product ages and competition subsides. There is a lifecycle of software and to understand this better it helps if you can think a bit more like a software developer.
To generalise, the CAD software industry, on the micro scale, appears to have evolved to deliver the yearly evolution of authoring tools, executed on rolling 3 – 5 year development plans, along with equally evolving business models. For instance, we had perpetual licensing, where customers bought the perpetual right to access a version of software. Now the trend is for subscription licensing, where customers lease the right to use the software over a set time. The future looks to be heading towards a pay-per-use or subscription hybrid.
However, eventually software applications age; years of new layers of features compound to make products cluttered. The internal ‘wiring’ gets messy and fundamental changes can cause ‘regressions’ bugs in features that used to work but now don’t. At this point software companies need to make important decisions on how to maintain their success.
In the software world, a ten-year-old software program is deemed old. If a software package is extremely successful, the common-sense thing to do is to stretch it out and to do enough to retain the loyal base and keep the competition at bay. If you do a good enough job, there are few, if any, competitors and software is ‘sticky’. But even here, at some point, software developers usually take the risk of writing something new from scratch, or they acquire a start-up that has done that work for them as it’s better for them to kill their most successful products, rather than have a competitor do it.
In this case, Autodesk decided to buy the next generation to kill Architectural Desktop, and switched BIM horses.
If you want to understand where your chosen CAD product is in its lifespan, look at the development velocity. The less that is added each year, the more likely it is very mature. If new features appear to be ‘onion skins’ to the application, as opposed to fundamental core rewrites or big new dollops of functionality, it’s most likely that the team don’t want to dig into the product’s guts for fear of playing Buckaroo with all the dependant code.
Another option is to redevelop applications on the fly, rewriting core features sequentially and doing a lot of under the bonnet work while keeping the same interface and file format – a case in point is ArchiCAD which had good multi-core support added at a platform level and AI added to rewritten features like the stair tool. While this may keep the code refreshed, fundamental principles remain the same and this can come unstuck if industry workflows change or a new upstart comes out without decades of baggage.
Moving to the macro level, every ten or twenty years the fundamental underlying technologies are renewed, and that could be hardware (like processors), operating system (DOS to Windows NT) or the change we see today, to cloud. At each of these technology boundaries, firms and market leading products can lose market share, even go extinct, if the software firms fail to manage the transition by having planned for change and being the best to leverage the new technology platform. This is why we are seeing so much concentration on, and development around, cloud-based apps, despite the majority of users being fairly happy on their desktop PCs.
Historic examples that spring to mind are PTC, which was the leading desktop MCAD player with UNIX-based Pro/Engineer. It was severely mauled by the new Windows upstart Solidworks because PTC failed to get to Windows quickly enough and kept its prices high. Autodesk had a serious wobble when it went from AutoCAD R12 to AutoCAD R13, from DOS to Windows NT, completely rewriting the code in C++, adding in solids and then trying to build vertical applications on constantly-moving, buggy base code.
Autodesk was lucky that nobody was there to really pick up on the company’s self-inflicted errors, but the effects lingered long into AutoCAD R14 and partially led to the decision to develop Inventor, which was Autodesk’s first new in-house developed code-base since AutoCAD to tackle the MCAD market and specifically Solidworks. These generational changes are major hurdles for software firms to negotiate and not everyone gets out alive.
The open letters
Coming back to where the current AEC market is within these various development cycles, and we find ourselves at one of these key inflection points, as the platform transitions from desktop to cloud. The prime BIM authoring tool in many geographies is Revit, which has revolutionised AEC design, democratising the adoption of model-based methodology. However, it is 20 years old, has a low development velocity and doesn’t really make much use of modern multi-core computer architectures.
Revit’s yearly development velocity for the past six years has disappointed many of its biggest fans / most mature users, despite there being no shortage in requested features to add. This has long concerned Revit customers, especially when they see Autodesk spending over $1 billion developing and acquiring technology for its ‘Construction Cloud’ offering as an attempt to digitise and sell tools and services to the currently in vogue construction market.
Devoted Revit customers have been frustrated by this slow pace of development of the A in AEC from Autodesk. Prices have also risen and licence models changed, too frequently, which have increased cost of ownership.
On 22 June 2020, something snapped. 25 UK and Australian practices wrote an open letter to Autodesk’s CEO, Andrew Anagnost to raise the issue of Revit’s lack of development, the increasing cost of ownership, its lack of interoperability, Autodesk’s constant licensing changes and treatment of customers with audits.
On publication, it was quickly shared on social media and in the architectural press (including this magazine’s in-depth coverage ). Then the group started to be contacted by other ad hoc architectural groups from around the world which had similarly banded together to ask Autodesk what was happening with Revit development and to complain at price hikes.
Architects in Australia/NZ, Hong Kong and South Africa had all attempted to engage on the same subjects. It seems that in private Autodesk had more than enough warnings that its best and most mature Revit customers were all feeling very frustrated and sought a collective response.
The Open Letter movement now stands at 166 global firms directly signed up at the board level and as we were going to press, Brazil’s National Association of Architecture and Consulting Engineers (sinaenco), representing a whopping 30,000 firms, also released a similar open letter after its members took part in a questionnaire which backed the sentiments of the original UK/Aus letters group.
It cannot be said that these grievances are just the feelings of a few practices in the UK. This is a pandemic in the mature Revit base. These formal and informal groups are now talking and sharing experiences and opinions. We are essentially seeing the ‘unionisation’ of BIM-focussed architects and engineers as they openly address issues with Autodesk’s product development and a range of business practices.
We have seen a number of public responses from Autodesk, mainly from Amy Bunszel, Autodesk vice president of digital engineering products, and Andrew Anagnost the company’s CEO. Firstly, there was recognition that architecture has not been a strong focus of development. Anagnost stated, “the pace of Revit development has slowed over the last five-plus years,” admitted there were “some areas where we need to improve and take steps to fix,” and “our architecture functionality didn’t progress as quickly as it should have.”
Anagnost gave a vehement rejection of the criticism of Revit price increases, stating Revit pricing was ‘reasonable’, producing a PDF of historic evolving Revit prices, given in constant USD (income free foreign currency movements).
The positive is that Autodesk made some admissions on its lack of development of Revit and gave a commitment to address that. Its other defence, that it needed to allocate resources to BIM 360 and Construction Cloud, was perhaps less persuasive to architects.
Autodesk engaged directly with the Open Letter Group and has since had two meetings, one with the executive team, the other with the Revit development team. The first executive call was described as a listening meeting, where the Group presented their grievances on a wide range of topics (pricing, licence changes, interoperability, Suites/Collections ‘product stuffing’, development, non-compliance audits) and the second revolved around wishlists and features. Both are ongoing conversations which will hopefully have some positive impact on licensing, product development, productivity enhancements and non-compliance ‘audits’. Firms that sign up to support the letter can be involved in this process and be kept up to date on the progress.
While I do think Autodesk is listening, the one area in which I think there will be no change is that of cost of ownership. It’s the one area Autodesk has pushed back on and in conversation with Wall Street analysts Anagnost described the complainants as the ‘privileged 20%’ as they had ‘moved from maintenance to subscription and have really pretty deep price protections, relative to the rest of the base.’
The first thing to point out here is that this privileged minority who moved from maintenance to subscription would by definition be long paying, mature customers who would have made significant investment in Autodesk, its products and services by buying perpetual licences – licences which Autodesk first priced up and now is seeking to move to Subscription.
The second point is the PDF generated by Autodesk are in constant USD, hiding the currency fluctuations and corresponding price hikes that happened in countries like the UK, and is not representative of the historical cross-grading and upgrade costs provided by the Letters Group.
I think customers can work out for themselves if they are paying more and if they are seeing value. The concern at the increasing cost and the value of Revit is something that is felt globally, from single users to Enterprise Business Agreement customers – it’s not just the privileged few. I don’t think there will ever be agreement between users and Autodesk on how much is too much, until customers leave for alternative platforms.
In talking with the group, one of the most interesting aspects of discussion is ‘what is the future of Revit?’ It’s something that we here at AEC Magazine have been trying to prise out of Autodesk for many years now. After years of low level development and Autodesk’s focus being elsewhere, customers have a right to wonder what’s the plan? Will the current Revit be rewritten to make use of modern processors and GPUs? Will Revit be able to handle larger models? Will there be a Revit 2.0?
We have seen in the MCAD space Autodesk already make another new modelling tool called Fusion, which will eventually supersede Inventor as its feature set grows, utilising the cloud as a data backbone. Unlike applications that run in the cloud through a browser, so called ‘thin client’ applications as the executable on your machine is small, Autodesk has settled on ‘thick client’ technology which provides a bigger sized desktop app which synchronises with the cloud host.
In this market Autodesk competes with Onshape, now owned by PTC, and while Fusion and Onshape offer much promise, they are still a long way from replacing their more mature and powerful desktop competition (Autodesk Inventor, Siemens NX, Dassault Systèmes Catia). Given Fusion was launched in 2012, this move to the cloud for data authoring in MCAD is taking a long time and AEC has yet to start the journey.
In Autodesk’s Q1 2021 call with analysts Anagnost explained, “We believe that relatively modestly sized to thick clients with a really robust cloud backend are the future. Fusion has a thick client. But it has a very, very, very fine-grained multi-tenant cloud data infrastructure hidden behind it. Fusion’s cloud will get thinner – client will get thinner over time. You could also see an evolution with Revit that’s similar to that. That’s going to take a little longer.”
In 2017, Autodesk announced the development of Project Quantum. The idea was a centralised cloud system, in which all AEC participants would work and be able to see the real time geometry changes of other users. This was a great concept and the dream was that all project participants could work on a single model, a data centric approach.
The problem is it seemed to vanish, with the excuse being made that the technology was in some way re-appropriated to a different team. Talking with Anthony Frausto-Robledo, the owner of Architosh, he seems to have an Autodesk insight that Quantum was little more than a ‘PowerPoint deck’ at the time.
Roll forward to 2018 and Autodesk has Project Plasma, which built on some of Quantum’s core principles, expanding to embrace digital fabrication. Here the future was a central cloud database, with thick and thin clients accessing the data simultaneously and while catering to the needs of current design and build workflows, it would be designed to drive digital fabrication, offsite construction and modular.
I got the impression that Revit would somehow first be hooked up to the cloud, the database load would be lifted off the desktop app and eventually discipline-specific apps would evolve to dissolve the need for the historic, monolithic Revit desktop app we have today. Much of this is based off the work Autodesk has put into Forge, where applications can be quickly built from web components such as viewers, clash detection tools, document management, generative design, DWG tools etc.
When we talked to CEO Andrew Anagnost in October 2019, progress had been limited. It was clear to hear his frustration at the slow development and he stated that the essential data backbone work had still to be completed and would take till Q2 of 2020. The good news is that this work seems to have happened as in another Wall Street call, Anagnost hinted that at this year’s AU virtual, something would be shown from the Plasma development (and that it was not called Plasma anymore).
The question will be what is the focus of the new system – architecture or construction? From conversations with Autodesk developers the primary focus will be on construction, not design, and this does have good reasoning. To design a workflow, you need to know the output and when defining a software application for AEC for the next 20 years, digital fabrication must be at the core. It’s not something that can be added in as a patch later on.
A single design system that can take customers from conceptual (loose geometry) to 1:1 accurate models to generate Gcode, with everything in between, is a big ask. To some extent Autodesk already has the architecture market with Revit, but it wants to concentrate on cracking construction. As there isn’t anything out there to buy, Autodesk is developing itself.
There is no Revit 2.0
Autodesk has been clear; the development work it will be doing to Revit, based on the current roadmap, does not involve a bottom up rewrite. There is no next generation rewrite of Revit coming but there is a commitment to improve the roadmap to include some of the input from customers.
The question will be how deep into the guts of Revit will the development team go to improve performance and functionality. ‘Low hanging fruit’ enhancements that do not require a lot of re-engineering work will be more appealing. The question is will this be enough to keep mature, advanced users happy? Will the development of Revit be rapid enough to feel like it has velocity once more?
The Revit that is available today is the only Revit that there will be. In a recent ‘Inside the Factory’ video, featuring the Revit product leads, Kyle Bernhardt, director of building design strategy at Autodesk, was asked by a customer “where will Revit be in ten years?” He replied, “It’ll probably be running on a Windows machine of some variety, Revit, as you know it today, in a much more inter connected cloud ecosystem of tools, and will be able to deliver projects in a way that no system has ever allowed them to do so before. Revit is not going away, it does some very useful things for the world, like AutoCAD does some very useful things to the world today.”
He added that Autodesk is exploring how to decouple some of Revit’s sub services to work better with Autodesk’s Construction Cloud. He then went on to claim that Revit is being ‘modernised’ and ‘broken up into parts’.
Reading this, Revit customers could be excused for wondering what on Earth is going on? Unless there has been a dramatic change of heart at Autodesk in the last week, there is no next version of Revit, no new generation, no fundamental rewrite. The Letters Group, while their input to the development has been welcomed, have had their expectations tempered to not to expect to see any radical rewrite of Revit.
Autodesk’s historic lack of Revit development seems to have many reasons, a switch to construction, a switch to developing cloud services, while maximising the amount of money it can get (and value for shareholders) for doing as little development work on Revit’s old code.
Thanks to the industry collectively calling Autodesk out, there will be some redress to the development and roadmap situation, a near-term win. This does not answer the long-term question of how can skilled architecture practices move beyond what’s possible today, when they feel limited by their 20-year old tool?
The lack of development and focus gave the distinct impression that Revit was in, at best, maintenance mode and raised the question of its imminent demise. Conversations with users tended towards reciting the Monty Python ‘Dead Parrot Sketch’. The reality according to Autodesk, is that it was ‘only resting’.
While ‘Plasma’ architecture might have a data back bone and new name now, it will be shown in some capacity at AU this year. However, this seems to be very early days and may well take a significant period of time to reach the market and then mainly for construction firms, not architects or design. The speed with which it can be brought to market will be a test for Autodesk’s Forge development system, which should enable accelerated delivery of tools and workflows, but as I said, this is a massive undertaking.
The Autodesk plan seems to be to try and keep Revit customers happier, in the meantime build the Construction Cloud, under the hood integrate Revit desktop subsystems to better work with its cloud services. Then over time, replace Revit with thick client applications for architecture, MEP and structural. My estimate is this could take a timeframe of six to eight years to complete. I do have serious concerns how the current Revit desktop, even with more development, can be stretched out to keep the attention of demanding architectural users.
There is no Revit 2.0 and the future appears to be some kind of slow absorption into a cloud-based construction system with new thick client applications eventually replacing the single monolithic Revit application. Autodesk is technically ‘gapping’ between generations but is now desperate to try and reassure the Revit base that all is well. I’m not sure whether promising them that the same tool will be on Windows 2030 is an incentive to stay or go?
Autodesk’s development focus is clearly on construction; the plan for architectural seems amorphous and not fully fleshed out. Without the intervention of the Letters Group, I doubt we would have seen the Revit product owners so visibly engaging with their users.
With regards to licensing and costs, Autodesk has picked up that customers want pay per use (PPU). Looking at their current Token system, after 8 mins you get charged a whole day’s worth of tokens. 8 mins is the same as 23 hours, 59 mins 59 seconds. Most firms have Revit models that take upwards of 20 minutes to just load. So it costs a day’s worth of tokens to see your own model in Revit, let alone do any modelling.
In PPU I’m hoping that Autodesk is looking at better granularity on this, but I have concerns about cloud and PPU in general. Software companies will charge you to create your design, to store your design and then will make money from your supply chain accessing your design and using their tools on your data. Your data becomes a commodity in a pay to play system. This has all sorts of data protection and monetisation issues which need to be addressed.
This moment is clearly a time for architectural firms to reassess BIM strategy, its usage, the workflows, the workarounds, partnerships, deployment, product mix. I am seeing clear trends where architects are reclaiming the design portion of the process and pushing BIM tools to the documentation phase, almost to the extent that it becomes a discrete process.
BIM is too constraining for early stage design and seems to have not really delivered in automating the drawing process which was its main objective. From talking with the Letters Group it’s been really thought provoking discussing what comes next for design.
This is something that the Letters Group will cover in detail in an unmissable panel discussion that will take place during AEC Magazine‘s NXT BLD Virtual conference next month (8-14 October 2020 – get your free ticket here).
In talking to Autodesk’s competitors over the last few weeks, many asked me, if Revit customers were so upset, why did they decided to write a letter and not just move to another system? It’s a good question. One of the firms involved in the open letter, described it as ‘having dug a hole so deep, that when you have serious concerns as to the direction you have travelled, and the costs now involved, you look up and see how far it is to get out, and the question is, whether you should climb out of the hole and start again, or keep on digging, in case it gets better.’ In many respects, the multitude of issues that the Open Letters group addresses take this beyond wanting a better Revit, it wants a better Autodesk.
If you enjoyed this article, subscribe to our email newsletter or print / PDF magazine for FREE