[ADSK, Dassault, PTC] Simulation, CAD, and PLM: Part 2

Autodesk, Dassault, and PTC were all founded as CAD companies in the ‘80s but have since become much more than that.  PTC first integrated PLM and is now branching off into industrial IoT and augmented reality.  Dassault (which also moved into PLM and simulation) and Autodesk are no longer just supplying software tools to solve narrow design problems, but also spinning up collaboration and data aggregation hubs to coordinate workflows across supply chains, a major strategic push that has been accompanied by a shift in revenue model from perpetual licenses to SaaS and subscriptions.

(Aside: keep in mind that a “subscription” model is not necessarily the same thing as a cloud hosted “SaaS” model…the former is about how you pay while the latter is about where and how software is run.  This can be confusing as management teams often speak about subscription and cloud interchangeably.  They are not the same thing.  60% of PTC’s revenue comes from subscriptions, only a small part of which is true multi-tenant cloud hosted software accessed via web browser.  Likewise, Dassault talks about 3DExperience as a cloud platform, but in fact nearly all customers consume it through on-premise licenses.  Autodesk is also predominantly on-premise, sourcing less than 10% of its ARR from cloud apps)

ADSK

Whereas PTC and Dassault are known for high-end 3D CADs geared mostly toward large manufacturing and industrial enterprises, Autodesk sells predominantly to “Architecture, Engineering, and Construction” (AEC) companies (think companies design and build rail systems, commercial buildings, and bridges), with whom its flagship AutoCAD and Revit software commands a dominant presence¹

AutoCAD is a general purpose 2D CAD developed in the1980s to replace physical drafting boards. Revit, which came onto the scene 20 years after AutoCAD, is a more sophisticated 3D CAD that embeds data elements specific to designing buildings. So in AutoCAD you might draw some lines to represent a wall and calculate the area. But to determine the cost of painting the wall, you’ll have to jump into Excel and specify the cost of paint, and changing the wall’s dimensions in AutoCAD will mean also updating the paint cost calculations in Excel. That design and data elements are stored in two separate applications is especially inefficient when there are several designers working on the same AutoCAD model.

“Building information management” (or BIM) software like Revit makes things a lot easier. A user can specify the cost and weight of materials and all the calculations dynamically change with a structure’s dimensions. To maintain design consistency, architectural firms will standardize on Revit, which in turn compels universities to train their students on it. Widespread adoption by architects also compels usage downstream. The building developer will usually interact directly with the architect, who runs point on the project and hires the engineering firm. So if the architect is designing in Revit and storing documents in BIM 360, the engineering firm will also license those tools to make it easier to coordinate and ensure design consistency.

The common refrain you hear from customers is that there is no other practical option beside Revit. According to this Architect Magazine article, an estimated 90% of large US architectural firms use Revit.  Revit is so deeply ingrained and mission critical that a group of architectural practices, rather than switch to a competing software provider, could only decry Autodesk’s (allegedly) aggressive price hikes and torpid product innovation² in an open letter.  Of course, Autodesk balks at suggestions of monopoly power but, per Zero to One logic, that denial is probably a strong indication that the company does, in fact, de facto possess it.  

(Management reports revenue from its flagship AutoCAD application separately even though AutoCAD is used mostly by AEC professionals).

Relative to PTC, Dassault, and Siemens (via their acquisition of UGS), who got their start at the very high-end of the market, among large OEMs who required the most sophisticated design tools, Autodesk targeted smaller and mid-sized businesses³ with a relatively low-priced offering.  To reach customers of that size, the company developed a wide network of resellers (Autodesk has 1,700 partners and resellers who account for 70% of its sales; by contrast, PTC derives ~70%+ of its sales direct) and created a product that was easy to use and deploy. The combination of broad reach and a more intuitive user experience – the kind of user experience that high-end players have been trying to create for years – becomes an increasingly differentiated asset if the goal is to put design/data sharing tools in many non-expert hands.  

Autodesk’s mid-market emphasis has resulted in widespread adoption of its DWG design file format.  Today, there are some 10bn proprietary DWGs floating around, practically incompatible with third-party CADs, reinforcing AutoCAD lock-in.  Testament to the lock-in power of ubiquitous file formats, when creating cloud hosted versions of Word, Microsoft, rather than simply creating a new web-based replicate of Google Docs, embraced the complexity of creating a “Word Web App” that was interoperable with its widely adopted legacy on-premise version.  From this post by Terry Crowley:

“The final decision to build the “Word Web App” rather than “a new web-based word processor from Microsoft that is not fully compatible with Word” (and similarly for Excel, PowerPoint and OneNote) was strongly driven by the belief that the file formats continued to serve as a critical competitive moat with immensely strong network effects”.

The captive power of file formats also applies to Autodesk’s Revit, the main subject of complaint in the open letter referenced earlier. From Architect Magazine:

“Recently, when the tech leader’s licensing agreement with Autodesk came up for renewal, they considered switching their firm to ArchiCAD, a BIM platform developed by GraphiSoft. In terms of features, they thought ArchiCAD seemed comparable to Revit, perhaps even better in some ways. Financially, ArchiCAD’s licensing cost was less than the $1 million-plus they spend annually with Autodesk—even after factoring in the price of retraining hundreds of staff and redeveloping BIM content. But when their firm started getting serious about the deal, the tech leader began reading through past contracts with clients and realized that many required projects either to use Revit or to deliver data using Autodesk’s proprietary .rvt file format. Their firm couldn’t afford to lose this work so they stayed with Revit—and with Autodesk...

‘With the move to subscription,’ Bunszel says, ‘it’s easier than ever for customers to vote with their feet.’ In a sense, she’s right. Theoretically, firms could cancel their subscription with Autodesk at any time and jump to another BIM platform the same way a consumer might cancel Netflix and signup for Hulu. But in my reporting for this piece, I couldn’t find a single large American firm that had ever transitioned off Revit…practices are so invested in Revit that it can be hard to escape. Companies didn’t just buy a Revit license— they trained staff to use Revit, developed Revit content, and conditioned clients to expect Revit deliverables. And in doing so, they are at the mercy of Autodesk’s next action or inaction, as it were.“

You can technically bring a Revit file into AutoCAD or some other third party application, but doing so requires a conversion process where key values are often lost. Like you might see a 3D image of the Revit design but not the data specifying the thickness of the doors, for instance. One might use competing tools around the edges but everyone refers back to Revit as the master software.

Among the legacy CAD players, Autodesk was the first to convert its base of perpetual licensees to subscriptions.  After discontinuing perpetual license sales in 2016, Autodesk began pushing maintenance plan customers (that is, customers who were still paying for maintenance plans on perpetual licenses they purchased in prior years) to subscriptions through carrot and stick, simultaneously jacking up prices on maintenance plans while offering big discounts on subscriptions.  The migration was somewhat tricky because 90% of Autodesk’s revenue came through channel partners, many of whom had to be helped through cash flow challenges that the subscription model imposed on them.  Also, some smaller, lower revenue customers, who balked at the incremental cost of a subscription vs. the annual maintenance fee on already purchased perpetual licenses defected to competing solutions. 

Today, everyone appreciates the benefits of a subscription model and is willing to look past the near-term profit hit required to pull it off.  But pre-Valeant, activist investors were all the rage and the subscription model wasn’t so universally understood and adored.  And as these things often go, Sachem Head and Eminence, two activist hedge funds who took a combined ~12% equity stake in the company in fall 2015, seized on Autodesk’s declining revenue, depressed margins, and tepid share performance⁴ to call for cost cuts. 

They won 3 board seats (out of 11) and CEO Carl Bass, who fought the activists, resigned in February 2017 (he claims that succession planning had begun before the activists got involved).  Autodesk’s stock has appreciated ~3.5x since November 2015, when Sachem and Eminence filed their 13D’s, though I doubt they deserve much credit as the move to subscriptions, a revenue model that investors came to adore, was already in progress when the activists got involved and I don’t think cost efficiencies had much to do with Autodesk’s success.  Ironically, one could argue that Autodesk’s pursuit of subscriptions was a key contributing factor to the deflated margins that led these activists to push cost-cutting prescriptions in the first place.  By cutting revenue and margin guidance that it laid out during its October 2014 Investor Day just months later, management certainly didn’t help matters…but whatever!  Shit was going to look ugly for a few years no matter how diligently expenses were managed.  It’s funny to think that in 2015, analysts were all steamed that Autodesk was guiding to mid-twenties margins for 2019 when what they expected was upper-30s⁵…well, in fact, Autodesk’s margins came in at just 12% and stock quadrupled anyways.  Good job everyone.