Appian Corp (APPN) 2018 Q4 法說會逐字稿

Greetings, and welcome to Appian Fourth Quarter and Fiscal Year 2018 Earnings Conference Call. (Operator Instructions) As a reminder, this conference is being recorded.

I would now like to turn the conference over to your host, Staci Mortenson, Investor Relations.

Thank you. Good afternoon, and thank you for joining us today to review Appian's Fourth Quarter and Full Year 2018 Financial Results. With me on the call today are Matt Calkins, Chairman and Chief Executive Officer; and Mark Lynch, Chief Financial Officer. After prepared remarks, we will open up the call to a question-and-answer session.

During this call, we may make statements related to our business that are forward-looking statements under Federal Securities Laws and are made pursuant to the safe harbor provisions of the Private Securities Litigation Reform Act of 1995, including statements related to our financial results, trends and guidance for the first quarter and full year 2019; the benefit of our platform, industry and market trends; our go-to-market and growth strategy, our market opportunity and ability to expand our leadership position, our ability to maintain and upsell existing customers and our ability to acquire new customers. The words anticipate, continue, estimate, expect, intend, will and similar expressions are intended to identify forward-looking statements or similar indications of future expectations. These statements reflects our views only as of today and should not be reflected upon as representing our views as of any subsequent date. These statements are subject to a variety of risk and uncertainties that could cause actual results to differ materially from expectations.

For a discussion on the material risks and other important factors that could affect our actual results, please refer to those contained in our 2018 10-K filing and other periodic filings with the SEC. These documents and the earning call presentation are available on the Investor Relations section of our website at www.appian.com.

Additionally, non-GAAP financial measures will be discussed on this conference call. Please refer to the tables in our earnings release in the Investor Relations portion of our website for a reconciliation of these measures to most directly comparable GAAP financial measures.

With that, I'd like to turn the call over to our CEO, Matt Calkins. Matt?

Thanks, Stacy, and thank you all for joining us today. In the fourth quarter of 2018, Appian's subscription revenue grew 44% year-over-year to 38 -- $33.8 million. Our subscription revenue retention remained high at 117% as of December 31, 2018. For the full year, subscription revenue grew by 40% to $115.7 million. These results exceeded our guidance.

Our growth in 2018 demonstrates the increasing value we obtained on a per customer basis. Of 87 net new subscription customers last year, 10 purchased $1 million or more of software on a TCV basis. While adding more new logos than ever, we also increased total annual revenue per client to a high of $520,000. We ended the year with 37 -- 38 7-figure ARR customers. I'll say that again, 38 7-figure ARR customers, up 58% from 2017.

Appian is the first and only company to go public as a low-code vendor so far. A low-code platform enables users to quickly and intuitively draw their idea for a new application, and then the platform translates their drawings into software. There are 2 natural consequences of the low-code idea. First, it's fast, everyone understands that. Second, the applications are more powerful because they are effectively written by the low-code product, and here's what I mean. If you write an application on the Appian platform, basically you tell us what the application should do, and then we give you an application that does that thing. We are translating. We're writing the application, really. And if we're going to write it automatically, we might as well write it with the degree of advanced functionality that you wouldn't have had the time to implement yourself.

For example, the application is going to run natively on every mobile device and every cloud. It will be extremely scalable and meet the highest security standards. It will have outstanding integration capabilities, automated testing, hot failover, if you want it, and a lot more. You won't have to specify that you want these features in order to get them. They won't take any more time, and for the most part, they don't cost any more money. They are what I call free power. The natural result of having a platform write your application for you is that the application can be arbitrarily powerful.

So while in the past, the quality of an application would be a function of the developer, now the quality is a function of the platform. This is a profound shift, and it augurs well for Appian, because Appian has proven for years that it could automate the most mission-critical processes and applications.

We signed a major manufacturing company as a new customer in the fourth quarter. This Fortune 100 firm picked us over a major competitor because the customer specifically wanted a recognized leader in low-code. They needed to replace the legacy system for managing premium break requests, which are triggered when products or components need to be shipped outside their standard operating procedures.

An Appian partner successfully built this application in 6 weeks to support a 4,000-person division, including executives who will approve requests using their mobile devices. The customer expects the application to save them $9 million to $10 million in the first year.

We also added a top 5 global credit card company this quarter. They needed to automate the process used to collaborate with businesses to create branded credit cards from account onboarding to card management.

Before Appian, their business partners would have to complete a series of paper forms, which will then be manually processed. With Appian, the company digitized these forms for roughly 100,000 submissions per year. One of the customer's employees learned how to use Appian after watching our sales consultants build a custom demo on-site. He extended that demo into a free cloud trial. And by the way, our free trial experience has much improved lately. This firm finished their branded credit card onboarding application in just 8 weeks.

A top 5 global insurance broker also became a new customer this quarter. They needed to replace their legacy system for tracking commissions within their health and benefits division. Their legacy system has been expensive and complicated to implement and did not keep pace with the company's growth. The company estimated that an internal build would take up to a year, but in only 2 weeks, a single Appian sales consultant built a custom demo that included the primary features needed in their application. Our speed was critical to them because they have an executive mandate to move off of their legacy system by the end of the first quarter.

We also won at the Federal Motors Carrier Safety Administration, which is responsible for regulating and providing safety oversight for commercial motor vehicles and their drivers in the United States. The agency became a new customer because they needed to replace a hard-to-maintain legacy system that was used to track medical examiners' certifications and commercial driver medical exam results. In the new application, based on Appian, tens of thousands of medical examiners will be able to submit the results of their medical exams for agency review. We were selected over a major competitor because the agency can deploy on to our secured FedRAMP-approved cloud platform while retaining the flexibility to migrate to a different cloud, if desired.

Our European business produced solid results in 2018. The team increased subscription revenue by 74%, grew total revenue by 52% and close deals with 38 new customers. For example, a Big 4 accounting firm expanded their use of Appian in the U.K. by purchasing millions of dollars of additional licenses in Q4. They needed to deliver a managed due diligence service for their customers. In the past, they provided these services through a tool built on one of our major competitors. They switched to Appian, and they'll use our platform now for up to 300,000 know-your-customer verifications per year.

We also won a 7-figure deal with one of the largest companies in Italy. This new customer provides postal, logistics, communications and financial services. Their IT team needed a platform to modernize their businesses, including 35 back-office processes. They'll first use Appian within their financial services division to manage fraud investigations. Before Appian, thousands of agents had to log in to several systems to investigative a fraud report. But when they complete their implementation, agents will be able to fully investigate and act on fraud reports without leaving Appian.

Our U.K. team won a 7-figure deal with a top 10 global investment bank this quarter. The new customer wanted to manage deal life cycles within their global transaction banking business. Before Appian, managing a deal required users to access up to 7 legacy systems. Now all knowledge gathering and client servicing for 80,000 clients will happen in a single Appian interface.

We also won a 7-figure expansion with the U.K. division of a top 5 European bank. They turned Appian to enhance their compliance and business banking operations. The bank will use Appian to unify data about their customers to improve customer service, track and manage 50,000 suspicious activity reports per year, and improve underwriting operations for a small to medium businesses. We won due to our focus on the customer outcomes and our platform's quick time to value.

Partners are increasingly influencing more of our business. They helped us obtain about half of our new customers last year. They also influenced about 1/3 of the customers who ordered more than $1 million of software in 2018. As we do more business with our partners, we have introduced new programs to ensure customer and partner success.

First, in 2018, we improved and relaunched our A-Scores program. We grade every Appian practitioner skill with an A-Score, that's what we call it, an A-Score from 1 to 3. This helps customers hire the right team and rewards partners who staff projects with real experts. The point is to ensure good outcomes on the Appian deployments.

Second, we launched a new program called Appian Architect Services. It's an advisory service to provide additional support to our customers and partners. We successfully tested this program in Q4, so this year, we'll roll it out globally.

A-Score's and Architect Services will improve our already strong customer outcomes, which have been the engine of Appian's growth since we started the company.

Now I'll turn the call over to Mark for a deeper discussion of our financials.

Thanks, Matt. I'm pleased with our Q4 and 2018 results. I'll review the financial highlights for the quarter and full year, and then provide details on our Q1 and full year 2019 guidance.

For the fourth quarter, subscription revenue was $33.8 million, an increase of 44% year-over-year and above our expectations. This includes a $1 million onetime subscription acceleration, without which, subs revenue would have been 40%. We believe 40% growth is indicative of the performance of the business during the quarter.

Our total subscription software and support revenue was $35.1 million, an increase of 38% year-over-year. Professional services revenue is $25.1 million compared with $25.2 million in the prior year period and up slightly from $24 million in the prior quarter. While we can have some variability on a quarterly basis due to new business wins and project starts, we continue to expect our services growth to moderate as our partners become a larger part of our ecosystem. Total revenue in the fourth quarter was $60.2 million, up 19% year-over-year, and ahead of our expectations.

Our subscription revenue retention rate was 117% as of December 31, which was at the high end of the 110% to 120% range that we target on a quarterly basis. We continue to be pleased with our customers' expanded use of our platform. We ended the year with 378 subscription customers, adding 87 net new customers during the year. We ended 2018 with 436 total customers compared to 356 at the end of 2017.

During 2018, the number of customers with ARR growing -- greater than $1 million increased by 58% over 2017, demonstrating the deep value our customers get from Appian.

Our international operations contributed 27% of the total revenue for Q4 compared with 30% in the prior year period. For the full year, international operations contributed 29% of total revenue compared with 27% in 2017. This is reflective of the strong growth we are experiencing, both domestically and internationally.

Now I'll turn to our profitability metrics. For the fourth quarter, our non-GAAP gross profit margin was 65% compared to 64% in the same period last year, and a slight increase from 64% in the prior quarter. Subscription software and support non-GAAP gross profit margin was 91% in the fourth quarter, consistent with 91% in the fourth quarter of 2017. Our non-GAAP professional services gross profit margin was 29% in the fourth quarter compared to 36% in the fourth quarter of 2017 and 31% in the prior quarter.

Total non-GAAP operating expenses were $47.7 million, an increase of 29% from $37.1 million in the year-ago period. Non-GAAP loss from operations was $8.5 million in the fourth quarter, ahead of our guidance, and compared to non-GAAP loss from operations of $4.9 million in the year-ago period.

As you know, foreign exchange gains and losses can fluctuate. During the quarter, we had $0.9 million of foreign exchange losses compared to $0.3 million of foreign exchange gains in Q4 2017. Our guidance does not consider any additional potential impact of financial and other income and expenses associated with foreign exchange gains or losses as we do not estimate movements in foreign currency exchange rates.

Non-GAAP net loss was $9.1 million for the fourth quarter of 2018 or a loss of $0.14 per basic and diluted share compared to non-GAAP net loss of $4.8 million or a loss of $0.08 per basic and diluted share for the fourth quarter of 2017. This is based on 63.8 million and 60.4 million basic and diluted shares outstanding for the fourth quarter of 2018 and the fourth quarter of 2017, respectively.

Turning to our balance sheet. As of December 31, we had cash and cash equivalents of $94.9 million compared with $107.3 million as of September 30, 2018. Total deferred revenue was $111.7 million, up 25% year-over-year. With respect to our billing terms, the majority of our customers are invoiced on an annual upfront basis. However, we also have some large customers that are billed quarterly and others that are billed monthly. We will continue to remind investors that changes in our deferred revenue are generally not indicative of the momentum in the business.

Backlog as of December 31, 2018 was $230 million compared with $240 million as of December 31, 2017. The average contract length in the 2018 backlog was approximately 15% shorter than in 2017. Most of our on-premise contracts are now effectively 1 year in duration because of our focus on reducing long-term commitments to on-premise deployments.

During the fourth quarter, we used $7.4 million in cash flows from operations due to the timing of cash collections, primarily related to our international customers. Global cash collections have been strong in 2019 year-to-date.

Now I will quickly recap full year 2018 results. Subscription revenue was $115.7 million, representing growth of 40% year-over-year. Our total subscription software and support revenue for the year was $126 million. Professional services revenue for 2018 was $100.7 million, up 18% compared to 2017. Total revenue for 2018 was $226.7 million, up 28% compared to 2017.

Non-GAAP loss from operations for the year was $30.7 million compared with a loss of $18.8 million in 2017. This is in-line with our stated strategy to invest for growth to capture the long-term opportunity. We will continue to build on our momentum by supporting our go-to-market initiatives and the continued development of our platform.

Non-GAAP net loss was $33.4 million in 2018 or a loss of $0.54 per basic and diluted share compared to non-GAAP net loss of $17.3 million or a loss of $0.30 per basic and diluted share for 2017. This is based on 62.1 million and 57 million basic and diluted shares outstanding for 2018 and 2017, respectively. For the full year 2018, cash flow used from operations was $31.3 million.

Before turning to guidance, I'd like to give our investors an update on our adoption of ASC 606. As we've previously -- as we disclosed previously, we plan to adopt a new standard using the modified retrospective method. Since we are an emerging growth company, we've elected to delay the adoption of ASC 606, which means that we won't have to adopt 606 until we publish our 2019 10-K. As a result, our financial statements for the first 3 quarters of 2019 will continue to be issued under ASC 605 accounting standard. During 2019, our plan is to provide investors with some disclosures that will provide visibility into the financial impact to our results had we early adopted 606.

We don't expect the new standard to have a material impact on the timing of revenue recognition related to our cloud-based subscriptions and standalone professional services. However, we expect it to have an impact on the timing of revenue recognition related to our on-premise term license contracts because the new standard requires us to recognize the majority of revenue from these contracts upon delivery of the software. It is important to note that the vast majority of our on-premise contracts are 1 year in duration.

In addition, we expect to defer more commissions other than that the standard, which would reduce the amount of commission expense during the reporting period. To put things in perspective, approximately 65% of our total subscription revenue was from cloud subscriptions in 2018 versus 55% in 2017. In addition, for the past 3 years, approximately 65% to 75% of our software bookings were for the cloud. The more business in the cloud, the less of an impact ASC 606 will have on our revenue. We ask that you keep your 2019 and out-year models under the ASC 605 accounting standards until we officially adopt 606.

With that, let's turn to guidance. For the first quarter of 2019, subscription revenues is expected to be in the range of $33.3 million and $33.6 million, representing year-over-year growth of 31% to 32%. Total revenue is expected to be in the range of $59.5 million and $59.8 million.

Non-GAAP loss from operations is expected to be in the range of $10.5 million and $10 million, with a non-GAAP net loss per share of $0.17 and $0.16. This assumes 64.3 million basic and diluted common shares outstanding.

For the full year 2019, subscription revenues is expected to be in the range of $148 million and $150 million, representing year-over-year growth of between 28% and 30%. Total revenue is expected to be in the range of $258.5 million and $262.5 million.

Non-GAAP loss from operations is expected to be in the range of $29.5 million and $27.5 million, with a non-GAAP net loss per share of $0.46 and $0.42. This assumes 65.1 million basic and diluted common shares outstanding. We also expect to have approximately $20 million out-of-pocket CapEx during the first half of 2019 related to our new headquarters.

Our guidance is in keeping with our goal to drive 30% subscription growth and to have subscription be the primary driver of our business. We will continue to make investments in R&D to improve the speed, power and usability of the platform. In addition, we are making sales and marketing investments to deliver on our subscription growth goals and create an organization that can scale.

We'll now turn it over to questions.

(Operator Instructions) Our first question comes from the line of Bhavan Suri with William Blair.

Can you hear me okay?

Yes.

Great. I guess, I just wanted to first touch on this sort of offloading of services engagements. We think of Appian sort of a subscription mix shift story. And we think the greatest capacity of sales capacity should drive new logos. We've seen that acceleration in '17. I mean, greater reach of the partners, obviously, driving that, should help. But I guess, just if you could talk specifically about how that offloading of services works to the partners. And then are you seeing a level of interest you're expected to see a year ago from partners in taking over those professional services aspect of the engagements? Maybe we just start there.

Yes. I feel partners are extremely interested. So yes, I do see that. The way we do it generally is that we approach a customer with the partner. The partners sources the customer for us, recommends us into the space where we're engaged in a joint campaign. That leads us to approach the customer together. In cases like that, it's obvious which partner will get the business, and so it's not a difficult handoff at all. The real check for Appian is to be sure that we're able to convey to partners a greater share of our professional services behavior without having any impact on the reputational quality, the quality of delivery, the quality of customer experience. And so as we do this shift, part of our focus goes on to making the shift happen, and the other part goes on to making sure that we don't have any quality dilutions. So things like the A-Scores or the Architect services is an active focus on maintaining quality during the course of the shift.

Got it. Got it. And then, I guess, a follow-up question to that. As I think about that, is that shift to partners potentially delaying the move to subscription? Meaning, I've got to go reevaluate the application, rethink about the architecture. They're going to spend time during their maybe workflow or change management, and that's a time and billings, whatever for them, so they are going to extend that out. Obviously, the contrary to that is lots more partners, lot more reach, lot more CXO-level relationships that should accelerate growth. And so if I think about '19 guidance, and maybe even rather than '19 sort of a 3-year outlook, it feels to me the overall impact should be acceleration, whereas in the near-term it might be slower, but is that the right way to think about it or how should we think about sort of what that means? Because there's 2 sides, obviously. And I'd love to just get your -- Matt, your color on sort of that mix.

There are 2 sides. That was an interesting take on it. I found generally that leads that come in from partners, which is to say selling opportunities that are sourced jointly or exclusively through a partner are a little bit more advanced the moment they reach us, a little more developed and sharpened from the customer perspective. They know what they want. They're in discussions already. Those deals are likely to -- more likely to close and more likely to close faster than a deal that we had to cultivate from standing start as it were. So I think that the partner channel is a secure channel, a high-reliability channel for us right now.

Our next question comes from the line of Raimo Lenschow with Barclays.

This is Mohit Gogia on for Raimo. My first question was around, so you mentioned the 87 net new subscription customer adds this year, and are pretty much in line with the adds last year, if I remember correctly. And I mean I was just wondering if you can drill down a bit into as to the make up of those adds as to, if there a significant change into some thing like maybe the verticals or maybe how sort of like those, that the sales cycles have changed for those net new customers? That will be very appreciated. And I have a follow-up question.

Yes, all right. I do want to add some color on that. We're getting substantial saturation, like a lot of penetration into these clients. And the fact that we've got the highest revenue per client since we've been public, it's up to $520,000 per customers, shows that we're adding more value on a per customer basis. And I want to emphasize that to show that we can grow in multiple dimensions. Count of customers and then depth of value provided and revenue on a per customer basis.

Understood. But I guess, if you could think about the core verticals, so they are pretty much -- if you remember, financial services, I know you mentioned the health care and pharma and garment, obviously. But is that where you're seeing the most action or you mentioned about manufacturing geek, manufacturing customer win. Just wondering as to the horizontal sales motion as to if you can provide an update there?

I think that you are correct in your supposition that we are getting the wins that are right in front of us. The sectors that we're most familiar with, the customers who already know us. I think you're definitely seeing that behavior. And it's good business, but yes, that's what we're getting.

Yes, I think the composition this year versus years past is very similar with those 5 verticals, but we are also, obviously, making inroads into other verticals like manufacturing, et cetera.

Understood. And I have a follow-up question. So just in terms, I think, you mentioned around the contract backlog grew -- sort of like length coming in a bit shorter this year. Should I read into that as to that sort of like meant headwind to deferred revenues and billings this quarter or am I misinterpreting that? I think that's all my questions.

Yes, I don't think so. Because the deferred revenue is kind of just the billings periodicity. So it's either annual upfront or quarterly or monthly. The backlog is a duration, instead of, on average, you spend 3 years, historically, and it's shorter than 3 years and predominantly because the on-promise contracts are now about 1 year in duration.

Our next question comes from the line of Sanjit Singh with Morgan Stanley.

I kind of want to follow up on Bhavan's question, sort of around the sources of growth. If I think about your net expansion rate this year at 117%, that seems pretty durable and relatively consistent what we saw last year. And you've actually accelerated subscription revenue growth. And so I wanted to get a sense of, is your ability to onboard customers and drive revenue from those new customers, is that improving or accelerating versus last year? It seems that, I think you talked about 87 -- 10 of the 87 are already $1 million plus ARR customers, and I just wanted to get your view on the ability to get customers to drive material -- get new customers to drive material revenue?

Okay. Well, I think, this is very important topic for us to be able to move a customer from a lead to a buyer to a repeat buyer. So we focus on the things that could have an impact there. And I don't want to say that our impact is complete. There's a lot more that we could be doing, but familiarity is one critical thing. We want to achieve familiarity by approaching the customer with a known partner or approaching the customer in the context of a need for applications or processes which they are already aware of and which they are looking for a solution in. A developed opportunity is a much better, faster thing for us. And the more we can approach a customer, knowing exactly what they need, bringing references from similar customers, telling stories, maybe bringing some intellectual property, maybe bringing a partner who already is familiar with this issue, the faster we're going to be able to move customers through that cycle. So it is a topic that we think about a lot.

Perfect. And then, I guess, my follow-up question is sort of around the competitive environment. I think Amazon is making some noise around the low-code opportunity and potentially building a solution in this space. How do you sort of view competition from Amazon? And do you view it more as sort of evangelizing the market, sort of reinforcing your message to customers or how do you sort of gauge the potential competitive dynamic for someone like Amazon AWS?

That's right. Well, I think that we've got competitors of different types, for sure. We have some that are distant in terms of functionality, but loud in terms of marketing. And then we have some that are closer in terms of functionality, but not say that the juggernaut that Amazon would be. I believe that Amazon as a distant entrant, which is to say, with a very different skill set and offering profile from what we have is not a competitive threat, but could benefit us in terms of raising the awareness and popularizing the low-code concept. So I am pleased to see interest brewing like this. And I believe that the inevitable consequence of interest like that is that everyone, at some point, going to declare themselves a low-code provider, and this shouldn't concern Appian too much because our offering is very much different from the kind of perfunctory offering that would probably be a large distant competitor's entry into this market. We've differentiated ourselves. And what I was saying there in the initial statements about how these -- in the past, an application's quality will be a function of the developer, but we are now making it in low-code, so it's a function of the platform. That's a real kind of sophistication division. So the platforms that are the most sophisticated, the most proven, the most reliable, are going to have an enduring barrier. I think the high end of this space is a protectable space. There is a substantial barrier of reputational and functional entry, which is to say it will be hard to enter it based on reputational and functional factors, both. So I feel that the position is relatively secure and that entrance by well-known entities are going to do more to help us through popularization than they hurt us through competition.

Our next question comes from the line of Richard Davis with Canaccord.

This is Luke on for Richard. Could you, for us, drill down a bit on your free cash flow this quarter, specifically what were the drivers there? And when you guys see the firm sustainably going into the black on free cash flow? And then also, where do you guys see services going as a percent of revenue in, say, a year or 2?

I'll talk about the free cash flow in the quarter. It's predominantly based on timing of some international receivables. And as I mentioned in my prepared remarks, we've had year-to-date strong collections from that. So it's really just timing. I'll let Matt talk to kind of the forward-thinking cash.

Yes, that's right. I'll address when we might be in black and when -- where services are going. Okay. So the balance between our inflow and outflow is something we look at all the time. Very careful consideration goes into what the profile of the firm should be and whether we should make investments. We are genetically a -- and in the black operating organism. That's how we got here. We, more or less, broke even from the day we are founded to the day we went public. We are comfortable operating in a situation where we cover all of our expenses. The fact that we are not doing that as this moment -- at this moment is a deliberate and carefully considered decision. We are making that decision because of 2 factors: the assessment we have on the market and its potential; and on Appian's ability to grow and to occupy space in that growing market. Because of those factors, I believe, it makes sense for us to be investing as we are at a relatively careful, but notable rate. This decision is constantly being reconsidered and revisited. And it will be in the light of all future changes, we keep a close eye on this one. Next, I want to answer your question about professional services, where that's likely to go over the next year or 2. As you've seen, in 2018, our license, our subscription revenue is growing a good deal faster than our professional services revenue. And I would expect that pattern to continue. I think that we're going to see more professional services on the Appian platform moving to partners and that the growth will be modest on the Appian -- on Appian's professional services revenues, the growth will be modest. At the same time, I hope to see substantial license growth and our plans for 2019 illustrate that.

Great. That's helpful. And then a follow-up, if I may. When you guys grow an existing account, help me understand the sales model a bit? Are those generally pull versus push opportunities? And if so, is the commission structure different? And then what are the implications on margins over time as growth from within the installed base contributes more to the bookings mix?

Yes. There is a commission difference. We would prefer -- I think it's more difficult to get a new customer than it is to get an existing customer, so there's naturally some degree of incentives, are encouraging us to seek new logos. I think that -- one of my favorite things about this business is that we do not reach a saturation point easily. In one of our customers, we can sell again and again. There is a great frontier of new opportunities, and we start looking for that frontier the day we arrive. Before we make the first sale, we are already looking for the second sale. We are training our reps to do this because we know that there are multiple targets within each opportunity, multiple threads to pursue, multiple places where our software could make a quick and meaningful impact. And we want to be hunting those right away. That's how we do it.

Our next question comes from the line of Alex Kurtz with KeyBanc Capital Markets.

Just back to the partner discussion and how they drove half of the new business. How should we think about this factoring into operating leverage longer term? I assume you want that number to be higher over time. And then the second question around the partners and their ability to drive new business, what's the competitive dynamics and win rates when you look at a partner-led transaction versus something that maybe one of your internal salespeople executed in the field? That would be great.

Yes, I'll speak to the second part. Everything gets better with partners involved. It's like the cloud, right. There's a few factors in our business, where if we have that involved, everything is going to be better. So yes, the partners mean a higher win rate, a better deal, a cleaner value proposition. So we love that. We definitely want to encourage that. With regards to the leverage, do you want to talk to that?

Yes, I'll talk to that. So the relationship we have with the partners is they eat what they kill, which means that they basically do the services for the engagements that they bring us in and we sell the software. So from a leverage perspective, that's a wonderful model. We've been doing that a lot internationally, and we're starting to do that domestically as well. So as we gain more and more traction through the partners, obviously we should start to see leverage there.

Now you're factoring this into how you think about profitability, is that changing the discussion about the timing, as you see the mix shift to more partner involved? I mean, I'm just trying to square back to that...

Not really. Because if you look at it, direct sales is still an important sales motion for us. And so we're going to continue to invest in that. We're going to continue to invest in marketing. In fact, I would argue that we've under invested in marketing forever. And so we're going to continue to invest in those areas while we leverage the partner. But we look at the partner as a potential catalyst of a growth acceleration, but we're still in the early innings right now.

Our next question comes from the line of Chris Merwin with Goldman Sachs.

Just, I guess, a follow up a little bit on the last question. Can you talk a little bit about sales productivity and the fact that it's improving steadily? It sounds like certainly the partner network maybe has been contributing to that, but I guess, if we just think about the productivity of your direct sales force, can you just talk about the improvement you've seen and the runway for further improvement?

Yes. I'd say that our improving sales productivity comes down to 3 factors. Partners is one of them. Also, we're doing internal training, we're really focused on that, especially towards the middle of 2018. I think that, that showed a difference. And then third, we are prepping our salespeople with more targeted content that allows us to accelerate and expedite the opportunities once they discover them. So I think all 3 of these factors are going to help us in productivity.

Okay, great. And maybe just one follow-up on international. Can you talk a bit about where you're investing right now? What market look promising as you continue to expand internationally in 2019?

That's right. We would prefer to make a lot of investment in a few places to make a few investments in a lot of places. So that's the overarching approach here. We've got a few offices in Europe that are very strong. And we will see -- you'll see increasing focus on those. I think that's how you build a reputation. Instead of being present everywhere, you want to be great somewhere. And then that's a far more portable thing. I can take that reputation for being, say, a great in Switzerland, and move that to being a great in Sweden. But if you dilute too much and you don't create the management focus and maintain the culture and the reputation, then the growth is going to sputter. So we intentionally focus our growth into rather fewer regions. But we make sure that it's good growth.

Our next question comes from the line of Eric Lemus with SunTrust.

You gave the statistics around the ARR growth from customers with over $1 million in ARR. But can you talk about the mix of those deals? Are those mostly -- the new deals, are those expansion deals or net new customers?

All right. Most of those 38 who are $1 million ARR customers were existing deals, so they're expansions. However, they're not all. And either way, I think, the 58% growth in that category says some really good things about the degree of value that we can deliver and customers' willingness to pay for it.

Got it. And then another question on partners, somewhat similar to what you're talking about earlier, Matt, about going deeper versus broader. But when you look at your overall strategy with partners coming into 2019, would your strategy be more focused on going deeper with current partners or would you see more opportunity with adding to your partner network?

Yes, I want to be careful not to express a preference for deeper over broader because, I think, broader is also extremely important and it's core to our strategy. So I think our goal is to be able to provide multimillion-dollar ARR value on a per customer basis for some of our larger customers, which, I think, we're fully capable of doing. And at the same time, broaden our appeal to gather many new customers at the bottom of the pyramid and give them time to work their way up. A successful strategy will have both components going forward.

(Operator Instructions) Our next question comes from the line of Derrick Wood with Cowen and Company.

With the change in your sales leadership at the beginning of the year, are there any notable changes in the sales structure to be aware of? And can you give us a sense of how aggressively you plan to add that sales capacity in 2019?

Okay. First of all, as they change in sales leadership, this was exceedingly harmonious and incremental. We -- our current leader of the sales organization has been with the firm for 15 months and was transitioned gradually into this role. Had a lot of oversight. We had a lot of mutuality between the 2 leaders. They've gotten along great, they've collaborated. So this is the most gradual sales handover that I have seen. And so very little, actually, has changed. The sales operation feels like we didn't change it at all other than just the usual annual shift, which is about the same kind of annual shift that we've done in previous years. So it's been very continuous. Was there another part to your question?

Just to get a sense for capacity build in 2019, maybe just linearity of hiring or how aggressive you want to be adding sales capacity versus maybe last year?

We are going to continue to add sales capacity, and we feel that there is plenty of opportunity to sell into, and we are not at risk of saturating the market as we increase the size of the sales force.

Okay. And if I could just squeeze one more in. I've got a high-level question. AI and robotic process automation, obviously a couple of major technology trends taking place today. How are you guys positioned around these opportunities? And what kind of new initiatives could we see from you over the next year or 2?

Okay. You're going to keep seeing more AI from us for one thing. I think this is a big issue. We've got some cool AI functionality. We've done some fantastic integration work to enable our customers to make easy use of AI. For those who don't recall, I've said this in previous calls, so just briefly, our approach to AI is that we're going to be the most practical, simple, straightforward way to make use of AI. We're not going to try to write better AI than Google or Amazon can write, we're just going to let you make easy use of those high-powered AI algorithms in a kind of a snap-in capacity. We're also going to ship the product with some internal AI, but we understand that the world's best AI is going to be external, and we're going to make it simple for you to access that. So we're going to develop this concept further. We're going to make it even easier. We're going to -- we're going to develop it in ways that I can't announce right now. But we are very excited about AI, and we'll continue to announce practical innovations and new use cases as we go forward. We're also excited about robotic process automation, which we've got -- we have partnerships here. Well, there's a lot of excitement around -- amongst our customers for it, and we've shown a good capability to integrate with bots and govern them, be an orchestration layer that fits nicely with the RPA capabilities and organizes them to work at scale, which is a really critical piece of the puzzle for many of the larger customers who are running not a few dozen, but maybe a few hundred bots, and need an orchestration layer. I think Appian fits that bill extremely well. So we focus there. You'll continue to see more of that. I think there's a naturally good partnership between our functionality and RPA's functionality. So a strong affirmative to both categories in your question.

Ladies and gentlemen, we have reached the end of the question-and-answer session. And I would like to turn the call back to Matt Calkins for closing remarks.

Well, I want to thank everyone for their time this evening and their interest in Appian. And with that, we will close the call.

This concludes today's conference. You may disconnect your lines at this time. Thank you for your participation.