Appian Corporation (APPN) Earnings Call Transcript & Summary

Third session of the Future of Application Development Conference. This session is going to be with Appian, where we have both CFO, Mark Lynch; as well as VP of Product Strategy and Deputy CTO, Malcolm Ross. We're going to be talking all things low-code. Appian has been the pillar company in driving low-code and automation in the enterprise, so looking forward to a fruitful discussion there. Let me quickly get through the disclosures. Please note that this webcast is for Morgans Stanley clients and appropriate Morgan Stanley employees only. This webcast is not for members of the press. If you are member of the press, please disconnect and reach out separately. For important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to the Morgan Stanley sales representatives. And we're also joined by Keith Weiss, who heads our software group. Keith is going to ask some questions around the team as well. Maybe to start off with you. First off, welcome Mark and Malcolm. Thank you for joining us at this conference.

Thanks for the invitation. We look forward to it.

So Mark, maybe to start off with you, last quarter was another really strong quarter, and I think it like surprised a lot of people just because it's tough to get out there in front of customers, just given the lockdown, and yet subscription revenue growth accelerated to 40%. And I think it's fair to say that the company has been executing this year probably better than feared. So maybe 2 questions. How would you characterize the demand environment overall? And then if you could sort of break it out between federal and enterprise, maybe that would be a good place to start. And just any other sort of general context about what drove the acceleration in Q3?

Yes. So I think, overall, when the pandemic hit and everybody scampered, and we're all talking -- I'm actually in the office sitting, but we're all pretty much remote still, businesses realized they had to change, and they had to change immediately. And so one of the things that Appian does really well is we allow businesses to build applications quickly and allow them to change and be nimble. And so that basically put a spotlight on us, and we stayed relevant during the early parts of the pandemic by rolling out a COVID application, right? And it was really an ability for both companies as well as universities, higher-ed universities to have an application to bring back their -- either their students or their employees back to the workforce safely. And so we offered that initially free, and we developed it really quickly, took 2 to 3 weeks to build, and then we ultimately started selling it. And people -- and I think really that we got a couple of dozen logos from that. It kept us relevant during the pandemic. And then basically, people realized they needed to get these applications. They've been talking about digital transformation for a long time. Our sales cycles are historically relatively long, so the deals we ended up closing in Q2 and Q3 were being just dated prior to the pandemic, and yet they closed -- and we saw instances where deals closed earlier because they realized they needed the application now versus waiting. We also saw renewals come up during Q2 and Q3 that closed earlier as well, just because they're afraid that budget dollars may disappear in the future. So there was definitely some tailwinds from the COVID that we -- candidly going into it, we didn't -- we weren't sure what was going to happen. I don't think anybody really did, and that was good. We also saw -- in last quarter, we saw a lot of strength, you alluded to it, in the federal space. Software bookings were twice as much during Q3 of 2020 than they were in 2019, and a lot of that was -- had to do with the prebuild solutions or procurement solution that we built. We sold that in several agencies within the federal government. But also the federal government -- and the other thing a lot of people ask is was the federal business on-prem or was it cloud, and the vast majority of it was cloud. And one of the things we have is a key differentiator is that we have the -- one of the highest certifications, security certifications for the DoD for cloud deployments, and it's an IL4 with a partnership with Smartronix. And we leaned on that and closed a few deals in the federal agencies that we normally wouldn't have gotten because of that certification. And there's not that many companies out there with that certification. So that was another positive thing for us going into it.

That's a fantastic overview of Q3 and some really interesting insights there as well. Malcolm, to bring you into the conversation, I wanted to maybe start just by starting with maybe 1 or 2 of your sort of favorite customer examples that highlights the development velocity on that Appian can enable. And particularly when you sort of bake off with your competitors, anything that you would sort of highlight that just shows what Appian is able to do for customers as they start to think about building these low-code capabilities?

Yes. And probably to build off Mark's first reference, the workforce safety solution we actually built ourselves and brought to market is probably a great example of that, where we just had a very small internal team. When we faced the same issue that every other organization had with the COVID-19 lockdowns, unable to come to the office, requiring those regular health surveys of our employees, we needed to come up with a digital solution fast as well. So we build that ourselves in less than a week and then said, "This is a great idea. Let's get this to market," and have dozens of customers now ramping up on it. So it's a great testament to the speed at which a low-code platform can adapt very quickly to the needs. When I think, though, of the broader like a customer use case, it's not also just the speed that you get but also the accessibility of the design environment. Another great example is a customer of ours, who is one of the largest trucking companies here in the United States. And you imagine like a trucking company, put yourself in their shoes. Can they hire the best IT talent coming from MIT or Harvard, competing against Google, competing against these leading kind of cloud brands as far as getting the top IT talent? Yet these organizations have extreme pressure on them for digital transformation. They see what's going on with Uber and Lyft and the whole transformation that's occurring in transportation with digital engagement with our customers, and they recognize that, that can be applied to the trucking industry as well, and they have to respond. Now what they challenged with is they had a large cadre of traditional .NET and mainframe developers who would typically take a very long time to deliver software solutions, and that's how a low-code platform allowed them to really transform their business operations from a legacy mainframe system to a modern digital solutions, Android tablets, check-in, checkout experience doing a full cycle of truck management. Now their challenge wasn't just speed, though. It was talent acquisition. It was having a tool set that is able to take nontraditionally skilled people, not people who do not have computer science degrees and have them communicate their requirements effectively to the software and have it expressed very quickly to become new digital solutions that work on these devices. So to say it's only a speed is part of the value proposition, but it's also just that -- expands the talent pool that you get by using low-code solutions to enable faster development as well.

That's great context. And when we think what are the -- the debates that Keith and I are looking to address in this conference is low-code as a self-contained sort of stand-alone development platform versus low-code just being an inherent capability or feature of all kind of modern software applications. I think we're talking to cloud player when they're talking about their serverless edge platform, they're really focused on trying to make this low-code in the sense that they want to -- they want more than just your classic developers to use it. How do we think about how the market is going to evolve around those 2 axis your -- the Appians of the world versus just a broader low-code movement within software? Any thoughts there?

Yes. Low-code has a lot of analogies to the cloud movement over a decade ago, where cloud was this -- if you asked what cloud was in 2005, right, it was Salesforce. It was companies who were like exclusively operating software of the cloud. Netflix was still selling CDs in the mail, right, if you remember that time. And -- but now cloud becomes a general-purpose strategy for bringing products to market. Whether it is a streaming service from a major entertainment company or whether it is enterprise software, it has applicability across all aspects of how we deliver commerce to deliver our solutions. Low-code started as a design construct, as a platform for building applications typically in the enterprise software space fast, but now it's becoming a ubiquitous term that can be applied to any aspect of how I deliver my product. It's the idea that I'm going to try to break down that barrier of not forcing people to learn how my systems work but have systems learn how to talk better with people and have people express what they want out of their computer systems more directly in a natural language, which they talk and have them be able to consume those services more quickly. So we track over 240 vendors who advertise themselves into the moniker low-code now. Anywhere from video game software companies to artificial intelligence companies use that low-code term. So low-code is not necessarily evolving, but it's becoming a product management ethos and a go-to-market ethos which has to be the underlying of any go-to-market strategy that says I'm going to make my product as easy to use as possible by customers coming from new solutions in the software realm.

Makes a ton of sense. And when we kicked off this conference, Keith and I walked investors through a presentation where we sort of highlighted kind of the -- kind of big trends that are going on in the space. And we talked about 25 million professional developers going to 40 million over time, going at a nice split, but it's pretty clear that demand for development talent outstrip supply. And so I guess the question here is where do you think the population of low-code or so developers stand today? And over time, is there like a rule of thumb to think about what percentage of enterprise's labor force can be converted to development talent over time?

Yes. Well, as we all know, there's huge movement towards more people programming software. If you have a few kids as well -- I have 3 kids, ages 12, 9, 8. They're already doing software development. I already have a 12 year old who's gone through a drone programming class. The idea of controlling computers more directly through expressing what we want to do through programming a low-code design is becoming a norm in both our -- the youngest generations and how we connect business enterprises. Now to cite maybe more public source, I was actually in a meeting with Paul Vincent, who's the lead at Gartner for low-code as far as low-code announced in the market. And there was a report that they published called the Gartner 2020 Digital Friction Survey, where they did an analysis of how many people are actually participating in the active software development. So they found out about -- the total population of the enterprise, about 49% were pure technology end users, so just consuming software, not actually building things. 5% were technologists and central IT department, so your classic developer, I work in central IT, I have a computer science degree, and I'm building software. Another equal 5% were technologists, again, computer scientists, but aligned with business unit constructs. So about 10% of our organization was traditional software development, and then the remaining 41% were business technologists, people who were building software solutions. Oftentimes, these were maybe very complex excel spreadsheets. They were robotic process automations. They were low-code solutions. They were Microsoft Access databases, but they identify their own role is I was not just a consumer, but I was a creator of new technologies which this organization was using. So you look at that overall market landscape, 51% of that entire market landscape of organization is in the act of building software, building solutions to their organization, but the untapped nature is that 41% outside of traditional IT, how could I better allow them to build increasingly complex and robust and powerful scalable applications than, say, just their spreadsheets.

That's -- sounds like a very compelling opportunity in terms of that untapped talent pool, which sort of leads me to my next question. Just how far-reaching do you think this vision is going to extend to? When we look at some of the other segments of software, we have a data integration, application integration space that's been pretty code-heavy, to some extent, business intelligence, the machine learning, advanced analytics sort of workflow as well as sort of custom software development where you guys play. Do you -- is there an opportunity to extend into some of those other segments? Like where should we think about Appian extending this vision? Are you going to start to stay in that sort of custom Appian development world? Or can we see a day where you are training a machine learning model within the Appian platform?

Yes. Well, we already are. And maybe to answer that with another Gartner quote or related to this is that they predict 65% of enterprise software would be built on low-code by 2024, another prediction that Gartner put out. Where Appian is headed, though, is firmly focused on enterprise software, how do I enable business organizations to express their digital operations more directly in low-code software? Now we've had a lot of success in that in the workflow space, process automation, user experience design, mobile application design. And where we're really innovating right now is in data architecture. So one of the biggest challenges we've seen in there is when you choose to build a new software application, an initial choice often is designed -- is based on the data architecture I'm using behind the scenes. Whether I'm going to use a traditional relational database that maybe is very good for transactionality and for complex queries and logic, or do I want to use a document data store like a MongoDB? Do I want to use a vector column or data store? Do I want to use a time series data store or are there no SQL environments or DynamoDB? We're using 2 poles and other types of architectures. There's a lot of complexity in how you set up data initially and then make sure that you're optimizing that data over time for all the different ways people want to interact with it. They want to do fast aggregations, real-time transactions, reporting, analytics, things like this. So we're really tackling that head-on. And in our latest release, we actually announced a new Synced data store, which allows us to take data from any customer data stores. So whether it's any number of relational databases, Salesforce system or other web service and cache that in a local Synced data store, so the Appian product, for real-time transactionality and fast aggregation. So we're trying to make it really easy for a low-code developer to connect any database they want, suck that data in and then work with it in a very low-code sense without having to worry about creating a view or indices, just kind of get what they want out of that data store. So that has a lot of legs in the future as far as breaking down data architecture, simplifying it, making it very approachable and having the system itself automatically adjust based on the way that we know the users want to use that data store. If it needs to go from a transactional system to a more analytical-based system, it can adapt organically. The other thing with artificial intelligence, there's a lot of companies specializing in artificial intelligence, right, so specializing in auto ML and machine learning algorithms. We're really focused on specializing in AI in the case of where we're mainly seeing a lot of experience or use case scenarios, specifically in document processing and process optimization. So Appian has a lot of history and process workflow management across organizations and dealing with a lot of documents as people ingest documents. So we've been doing a lot of investments and using pure AI approaches to artificial -- to document understanding. The traditional OCR business was mainly based on upfront template design. I would take in like, say, an invoice, for example, and then I would lay out the bounding box area, say, like, okay, in this area, I'm going to have the invoice number. And this area is going to be the table of invoice details and required prescriptive design upfront, which was sometimes very onerous and also fragile because if you had a change in the invoice layout, it would break the entire model. I couldn't extract information anymore. So that was kind of 10 years ago OCR. Today's OCR systems, which Appian is investing in, is pure AI-based, where I just ingest invoices. I use a lot of document understanding features to recognize first this is an invoice and then do automatic key value pair analysis to extract that data and put it into the system. When I have failures of the AI system, we're using a workflow engine to basically route failures of the AI to a human to do reconciliation and then automatically goes back to the machine learning engine and retrains it and retrains it. So as humans work with the AI system, we can automatically have the positive feedback loop and create the AI to be smarter and smarter and smarter. So I think we have a really good advantage in this marketplace because we have such mind share in the human workflow side that allows us to really tie in those bots and AI systems to humans very tightly and have that kind of positive feedback loop ongoing. I'll pause there.

I'll kind of jump in with a question. This is Keith Weiss with Sanjit. I'm going to go out on a limb, and I say -- I think Gartner gets this wrong in how they characterize this marketplace, right? Because when I hear you speak and some other -- of the sort of vendors here, it seems like there's technologies in use of an end goal of workflow automation, right? We're trying to automate people's workflows, put it into an application, utilize data that's often trapped in sort of legacy systems to sort of a more efficient, more optimized workflow on a going-forward basis. And low-code platforms is one of those mechanisms, right? And you brought it up yourself. It seems like RPA is another one of those mechanisms, these various mechanisms in the service of automating those workflows. Can -- so one, do you kind of agree with what I'm saying? And two, how should we think about like the dividing line? Like what is more so on the low-code side of the equation and what's not in the RPA side of the equation? Or is that just like -- is that dividing line going away?

Yes. So low-code versus an adjective that applies to both workflow and RPA. So Appian directly offers robotic cross-automation as part of our product offering. So that's part of our inherent capability, and we apply our low-code methodologies to it. So low-code simply means I want to make it easy for business technologists, non-technically trained users to build what they need inside the Appian tool without having to learn complex syntax and scripting languages. That said, RPA is often confused in the market with other technologies, specifically workflow process automation. So RPA, the biggest misnomer the word RPA is the P word, robotic process automation. It should be called robotic task automation because what robotic automation does is emulates a specific human task and does that in an automated sense. Often times, you have like a call center agent who needs to move content from one screen to another screen. And that's error-prone because that's human-based, so they want to do an RPA bot instead to make it -- make that more efficient. But that's not really integrating that task into the larger ecosystem of the customer journey. I have an inbound customer request, then it gets to the customer service agent, then it goes to 5 other people. RPA has no context on the higher-level workflow. So that market is really called business process management workflow capabilities of how I lay out the master process. Now I think Gartner has it right in a specific term, which they actually call hyperautomation. Hyperautomation is what -- how they describe the confluence of robotic cross-automation technologies, artificial intelligence and business rules capabilities, so prescriptive business rules, as well as predictive AI. The iBPMS, BPM workflow technologies, integration, API integration, event processing and overall business operations, they encourage our customers to evolve beyond robotic cross-automation. Think about the more holistic workflow operations and areas for automation across the entire area. One of the biggest things that bothers me is robotic cross-automation is often misapplied in areas, even things like automating Microsoft Office interactions. If I want to extract data from Microsoft Excel and put it into PowerPoint, that can be done entirely through APIs. You don't actually need to do that through the desktop scripting. But robotic cross-automation has had a huge success over the past 3 years by penetrating the system development organization by showing that a nontechnical person can come in and build a script on their desktop environment, not contact IT and implement an automation for their specific tasks, which gives them a lot of benefit. I think where that's evolving is we had a lot of shadow IT and RPA over the past 3 years. We see a lot of CIOs evolving that decision to be, okay, I can't have 3, 4 different RPA products and business units doing this all over the place. How do I consolidate that into more hyperautomation practice where it can then have a more centralized control over all the automations and the environments is kind of where the market is evolving on automation.

Got it. If I could sneak one more in. So Sanjit and I have worked on a couple of like recent IPOs and indirect listings, guys like Palantir and C3.ai‚ right? And we sit down and listen to their stories, and they're like, "Listen, we go into these legacy systems. We take a bunch of data out. We put it into a model, normalize it, enable people to build applications on top of it." And I'm sitting back and thinking, "That sounds a lot like what like Appian does." Like am I wrong in that? Like is that a broadening of your competitive set into sort of what those guys are doing? Is there some fundamental difference in that?

I think you just described like almost every enterprise software application. Take data, move it into an app, build something around it. That describes legacy MicroStrategy tools like business intelligence environments as well. That's the fundamental of what enterprise software does. I guess the question is what can you do with it and how quickly can you do with it? So that's why, I would say, going back to the low-code moniker is, yes, so many companies are doing that, shuffling data around, building apps and new experiences around it. I think the buyers are going to really trend towards who has low-code capabilities, who is actually able to deliver that promise as quickly as possible and able to do that in a repeatable and long-term, sustainable fashion as well. So we have yet to compete against Palantir, I think, but the markets are always kind of a little bit emerging or confused in the enterprise software space.

To pick up on Keith's question on the competitive environment more broadly. I just did a call on Appian this morning. I think what investors really want to understand is how does a customer decide what gets built on what are now multiple available platforms, right? And you guys often see, I think, IBM and Pega, ServiceNow with its Now Platform, Salesforce with Lightning. There's some pure plays like Mendix, OutSystems, K2. Microsoft is going out with its Power Apps story, AWS coming with Honeycode. So in some ways, what applications sort of -- what are the types of applications are to self-select in what -- in like, say, an Appian versus a Salesforce or Microsoft Power Apps? Any way to sort of think about that? Or do you think that it really is like all these applications are up for grabs, and this can be a winner take most or all market?

Yes. Let me take that.

Go ahead, Mark.

You can tag on, Malcolm, as well. What Appian does, we allow companies to build very complex, mission-critical applications very, very quickly. And so all -- a lot of those vendors you just named can't build complex, sophisticated applications, and I'll let Malcolm describe what that means exactly. The only other real vendor that can actually build complex applications would be Pegasystems, right, but they have an older technology. It's brutal. It's not really cloud, even though they say it's cloud. And it takes a lot longer, and the total cost of ownership is much higher. So that's our defensible position. And the beauty of Appian is if a CIO wanted to standardize, if I got all these applications, I get a backlog of 3 years of projects that have to get done. And I don't have the software engineers to do it all, then I'm going to need something to do. And so if you were to select Appian, you can build some very complex applications that you need to do within the enterprise, but you can also build a whole bunch of the simple, easy, lightweight workflow stuff as well, and you can have all of your application development standardize on one platform. If you're not going to be able to do complex things with Power Apps, for example, or definitely not with Honeycode. But I'll let Malcolm add on to that. But that's kind of our -- that's our key differentiator.

Yes. And I think we have a very different go-to-market model as well. If you look at Salesforce and Microsoft and ServiceNow's go-to-market model, they're perceiving low-code as a reason to keep people baked into their ecosystem, right? So I have never seen a opportunity come up where our customer does not already have Salesforce, but they're actually looking at Salesforce's low-code platform. Typically, they're looking at Salesforce because they've already been a Salesforce customer for a number of years, and then they're looking -- while they've been extending Salesforce and looking for easier mechanisms. Salesforce is just not appropriate if you want to say, if I'm a Microsoft Dynamics shop, right, I don't use Salesforce. I use another CRM tool. I also maybe use Workday, and I use some SAP. It totally doesn't make sense to go use Salesforce's low-code platform to orchestrate Dynamics. What makes sense is where Appian's approaches is being a best-of-breed technology, where we want to be an independent low-code platform that allows you to bridge across Dynamics, Salesforce, Workday and combine these different data sources from any location into a single application. So we have a little benefit there of not being tightly coupled to these other SaaS architectures and the ecosystem, but we're allowing you to be kind of independent platform that produce all these other areas to build new application experiences around.

You hit on another theme that Keith and I have been looking to address, which is the cloud and the implications of broader cloud adoption, particularly as it relates to the app dev tools, right? And when we kicked off this conference in that presentation I was referring to earlier, I think we had a slide that said about 38% of the software development tools are in the cloud today, but that's expected to quickly cross the chasm over the next couple of years. And I know your own SaaS business is growing pretty -- at a pretty fast rate. But what are the implications as these software development tools move to the market overall but also for Appian? Does that -- is there -- did that drive more consolidation with fewer vendors? Does that accelerate your product velocity? What are sort of the next sort of implications as the tools market moves to the cloud themselves?

Sure. Do you want to take that, Mark? Or I'll take it?

Oh, you take it. I'd like to hear your thoughts.

Yes. I mean it's the natural movement. If a low-code is becoming a predominant platform as opposed like Java or .NET, things like this, it's going to be the natural evolution that you're going to see more package applications, patches SaaS applications built on low-code platforms. Appian has been really focused on this year as for exceeding that market with our workforce safety solution, with our institutional onboarding solution. We also have a really healthy ecosystem of partners who are also contributing to prebuilt solutions on the Appian product. So we fully expect to continue that business and continue to focus on prebuilt solutions in 2021 built on the Appian platform and seeing that market grow as far as not just as a development platform but also as prebuilt SaaS applications built on low-code. They'll also allow for a high degree of ease of configurability to adapt to customer-specific needs.

And I think to the extent -- I think it's -- all of us deal with multiple systems, like we have Concur and you've got NetSuite and all these -- you have the different log-ins, and you have different pass codes, and it's a huge pain in the ass. And if you can basically have one interface and access a ton of different applications, it's kind of like where everybody needs to go to. The beauty of Appian is the ability to be in the cloud. I'd say about in the past 2 years, about 80% of our software bookings were cloud, but you also had the ability to be on-prem, if you wanted to do as well, and you can toggle back and forth. We have had a lot of customers where they started on-prem and then they migrated to the cloud once they felt comfortable with the platform and the security parameters around.

Makes sense. And in the last couple of minutes, I kind of wanted to go back to -- this is kind of more of a question for you, Mark, around the IT budget. And where do you think low-code sort of stands with in terms of IT budget priority today versus a year or 2 ago? That's the first question. The follow-up to that is funding for things like Appian, does that come at the expense of kind of your traditional custom application development platforms or application development stack? Or does it come from some other part of the budget?

The first one is, as you guys know, you guys took us public. Low-code wasn't even a thing. When we -- IPO, we were going on, we talked about low-code, and we felt like we were evangelists out there talking about it. And now low-code is everywhere, right? SAP just announced their new workflow, low-code platform yesterday, right? So it's becoming ubiquitous. The beauty of that is the fact that it's a pull now. We're getting pulled into opportunities. There's RFPs for low-code. We're a recognized leader within low-code for application development. So that's actually a positive. And so to your point on the -- where we are on the IT budget stack, we're at the higher end, right? So if you think about it, there are going to be more applications written 5 years from now than they were in the past 40 years. And I have the same issue with the software engineers and software developers, right? Malcolm said 65% -- Gartner said 65% of all application developments can be done on a low-code platform by 2024. Forrester said it's a similar thing with 75%. So that's -- we're -- if we basically -- our focus right now is we've got a great platform. We've got a great mousetrap. We all know the best mousetrap doesn't always win, so we're focused on execution. And so this year, during the pandemic, we took the opportunity to beef up significantly our senior executive team. We have a new CRO, a new CMO, new VP of Sales Ops and a new Senior Vice President of Customer Success. Because what we look at it, in front of us, it's really an execution, go to market, crystallize, get the brand awareness out there. So when you think of low-code, you think of Appian. And it could be a tremendous opportunity. So we're focused on execution going into 2021. We got through the pandemic pretty well. Things aren't -- nobody is out of the woods quite yet. But once they do, I think -- and as you know, a lot of companies, their IT budget shrank. Ours internally shrank. We tightened our strings as well. And yet, we were still able to get business during that time period. So when those purse strings open up again, we're hopeful that we'll even get more business opportunity.

Just great insight. Maybe just to wrap up the conversation. We're about 2 minutes over, but I want to -- did want to hit on this topic, and it's a jump ball between you and Malcolm, but it's related to partners. And you have some key ones, whether it's KPMG, the Accentures and Deloittes. Can you give us a sense of what their -- what the nature of that partnership has been historically and where that can move going forward, meaning that -- I used to work at Accenture. So is Appian being used as part of like the solution-building process that an Accenture would deliver? Or are they becoming more of a reseller, if you will, of the Appian platform to its clients. Can you just talk about like what that relationship has looked like historically? And where do you want it to go.

I'll give -- are you comfortable with this question, Malcolm?

No. Go ahead, Mark. Yes.

I mean, historically, I'd say prior to the IPO, the partners looked at us almost as competitors, right? We're doing a lot of professional services. We have a lot of professional services business. And they were -- partners were afraid to build practices around Appian because we're a private company. We're in Northern Virginia. They weren't sure what was going to happen to us, et cetera. And I think the IPO legitimized the business and legitimized Appian in the eyes of these strategic partners. And the fact that we went out there with Morgan Stanley as the lead. It was one of the more successful technology IPOs in 2017. Basically, we had several individuals within each of these organizations that were willing to bet their careers on Appian and build practices within these organizations. KPMG is the biggest partner, both internationally and domestically. But Accenture is a fast riser. To your point, Sanjit, 2 years ago at their 2018 user conference, Accenture wasn't anywhere. They weren't even a partner of our. Then we signed an international agreement with Accenture. In 2019, they were Diamond Sponsors. So they went from nothing to a Diamond Sponsor which was the highest level. They're aggressively building out of practice. They're aggressively bringing us new opportunities, and so -- and we've talked about it on the call -- at our last earnings call, the last 4 quarters of the new logos that we closed, 70% or more were influenced by a partner. And those 4 partners are in order of importance today, and I think it's going to change: KPMG, Accenture, PwC and Deloitte. So the partner strategy, I think, is incredibly important because you can only hire as many sales reps. It's a linear process. Hiring sales reps, get them trained, you get them out there, you get marketing, and it's very expensive and time consuming. The partner ecosystem, if we do it right, gives us an opportunity to potentially grow exponentially. So that's what we're really excited about.

Well, with that, Mark, we're about 5 minutes over but a fascinating conversation. Malcolm, thank you so much for your insights on helping us better understand where low-code is headed and where Appian is headed. I learned a lot. And Keith, thank you for joining us as well. In terms of programming note, we got our last session with PagerDuty at the top of the hour with CEO, Jennifer Tejada; and the CFO, Howard Wilson. So stay tuned for that. And Mark and Malcolm, thank you for making -- taking some time out of your busy week to join us at the Future of App Dev Conference.

It's been great. Thank you.