Alteryx Inc (AYX) 2017 Q3 法說會逐字稿

Greetings, and welcome to the Alteryx Third Quarter 2017 Earnings Conference Call. (Operator Instructions) As a reminder, this conference is being recorded.

I'd now like to turn the conference over to your host, Mr. Chris Lal. Thank you. You may begin.

Thank you, operator. Good afternoon and thank you for joining us today to review Alteryx's Third Quarter 2017 Financial Results. With me on the call today are Dean Stoecker, Chairman and Chief Executive Officer; and Kevin Rubin, 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. These statements are not guarantees of future performance but rather, are subject to a variety of risks and uncertainty. Our actual results could differ materially from expectations reflected in any forward-looking statements.

For a discussion of the material risks and other important factors that could affect our actual results, please refer to our SEC filings available on the SEC's website and our website as well as the risks and other important factors discussed in today's earnings release.

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 their most directly comparable GAAP financial measure.

With that, I'd like to turn the call over to our Chief Executive Officer, Dean Stoecker. Dean?

Thanks, Chris, and welcome to everyone joining us today for our Q3 2017 earnings call. We are very pleased with our third quarter financial results. We achieved strong revenue growth, positive non-GAAP operating income and generated positive cash flow from operations.

During the third quarter, we reported revenue of $34.2 million, representing an increase of 52% over the prior year period. Our non-GAAP operating income was $885,000 in the third quarter, resulting in non-GAAP net income of $0.02 per diluted share. This quarter further demonstrated the strength of our land and expand subscription model, fueled by the power of our complete end-to-end analytic platform for the enterprise.

During the quarter, we landed a large number of customers new to the Alteryx platform experience and continued to expand our existing customers' Alteryx footprint, evidenced by our strong dollar-based net revenue retention rate that remained above 130% for the fourth consecutive quarter.

We have designed and built our end-to-end analytics platform to serve only -- not only the millions of analysts and statisticians around the world, but the Chief Data Officers and analytic leaders who are driving digital transformation in large enterprises. These leaders are focused on weaving a data and analytical culture into the DNA of their company. The citizen data scientists, trained statisticians and CDOs alike are seeking an end-to-end business-friendly platform capable of delivering simplicity and sophistication to win with analytics.

We've continued to enhance and extend the capabilities of our platform to further solidify our category leadership position. In mid-August, Alteryx Connect became generally available as an add-on module to Alteryx Server. Alteryx Connect is our collaborative data exploration platform for the enterprise that empowers users to discover, share and collaborate on data assets, visualizations, reports and workflows that are often siloed across large organizations. With Alteryx Connect, users can easily find and leverage curated data assets, otherwise locked down in Teradata, Salesforce, Tableau and many other persistence layers, ultimately driving efficiencies for data workers and decreasing their time to insight, all while being done in a governed, secured and collaborative manner.

Post-launch, we've seen meaningful demand for Connect from both new and existing customers. For example, Unitymedia in Germany landed with Alteryx Connect, because data is at the center of their business. To gain control over their data, they started using the data cataloging capabilities of Connect to standardize and give meaning to the massive number of business terms the company utilizes. They are taking full advantage of the capability to link metadata, data assets and visualizations and most importantly, the people who have curated those assets.

Additionally, at Inspire Europe in September, we officially announced Alteryx Promote, based on our acquisition of Yhat. Alteryx Promote, another add-on module to Alteryx Server, extends the advanced analytics capabilities of the Alteryx platform, empowering data science teams with an end-to-end solution to build and, most importantly, deploy and manage predictive models and machine-learning algorithms without requiring IT support. Promote requires no coding and exposes machine-learning model as REST APIs in scalable and flexible environments, either on-premise or in the cloud, allowing data science teams to execute the models in mobile apps or websites with ease.

With Alteryx Promote, models built by both data scientists and business analysts can be quickly deployed, managed and monitored to help the CDO maximize the value from their analytic investments. Alteryx Promote is currently being offered to a limited number of customers and will be generally available in early 2018.

We believe the demand for the Alteryx Platform will increase globally, as evidenced by an IDC study and our Inspire Europe conference held in the U.K. in September. According to research commissioned by us and undertaken by IDC, there are now 30 million advanced spreadsheet users worldwide, with 5.5 million residing in Europe alone. These citizen data scientists face the same challenges, spending way too much time and money in advanced spreadsheet efforts. Inefficient and risky spreadsheet usage cost businesses EUR 55 billion each year, largely because of repetitive and error-prone tasks like copy and paste.

We also see the demand for our platform growing as our Inspire Europe conference this year attracted nearly 800 attendees from 30 countries, including a meaningful number of partners, all records for us.

As with Inspire U.S., key topic areas for attendees included predictive and spatial analytics as well as Alteryx Connect and Alteryx Promote. Inspire Europe drove meaningful customer and prospect dialogue as well as a 62% growth in attendance and a 50% increase in free trial downloads. We believe that international markets remain a largely untapped opportunity for us and one that we will continue to invest in to drive growth.

In Q3, our international revenue grew 83% over the same period last year, while accounting for 23% of overall revenue. We believe our investments are paying off in significant ways. For example, one of the largest companies in Japan, SoftBank, a telecommunications and Internet provider, has recently expanded its use of Alteryx across all business operations. SoftBank initially implemented our platform within its mobile division, where they we wanted to replace a combination of complex difficult-to-use tools and highly manual processes for aggregating data from multiple sources to perform analytics.

With Alteryx, the ease-of-use and processing speed enabled them to realize a significant reduction of person hours through analytic automation with more timely, accurate and insightful reporting for management. This success led to a broader rollout to business users across Japan who now share common workflows to increase the volume and precision of data that can be analyzed, but ultimately, better serve customers.

Adoption continues at a rapid pace, as use cases are spreading across a wide [array] of SoftBank departments.

During third quarter, we added 231 net new customers and did business in 68 countries around the world. We exited Q3 with 3,054 active customers across a variety of industries and use cases. Many of these new customers are very recognizable brands landing with Alteryx for the first time. For example, in healthcare, we landed industry leaders, including both Sharp and Bayer HealthCare; in business services, we landed Grant Thornton and FedEx Corporate Services; in consumer goods, we have companies like The Hershey Company, TreeHouse Foods and Master Lock; in travel and hospitality, we added Travelzoo; and in academia, we added John Hopkins University to our growing list of valued customers.

The expand side of go-to-market model continued to perform quite nicely in Q3. Customers continue to realize value in the Alteryx platform, giving them the confidence to further expand with us. In the third quarter, our dollar-based net revenue retention was 133%, a 4-point improvement in net retention achieved in Q3 of 2016. For example, Caisse d'Epargne is a division of BPCE, the second-largest banking group in France. Their adoption of the Alteryx platform was an initiative of the newly appointed Chief Data Officer and Head of Analytics, intending to bring advanced analytics to the forefront to better serve customers and operate the bank more efficiently. The CDO wanted to move away from repetitive manual processes in Excel and enable self-service analytic functions across the enterprise. Alteryx was first implemented in Q2 of this year and they expanded in Q3 with Alteryx Server and additional designer licenses, demonstrating the vital role Alteryx is now playing for the division. Users are now able to work with a localized French language version of Alteryx to analyze data from the group's many data warehouses.

Initially, use cases cover analytics for location intelligence, branch optimization, customer segmentation as well as internally facing efforts to assess employee productivity, variables impacting the workforce and even trade union relations.

Lennox International is a global provider of climate control solutions, both for residential and commercial markets. The corporate IT team overseas BI, analytics and the company's data warehouses, with the objective of ensuring decision-makers across the enterprise are powered with accurate data and key insights. Alteryx loaded the strategic platform chosen to increase the team's agility to support the growth of the organization. One of the deciding factors in making Alteryx the enterprise standard was the ability to standardize, integrate and enhance disparate data sets, required for automating the analytic process. Lennox has seen steady adoption of Alteryx, including continued expansion with Server, additional designers as well as our package data set options for users across different business units. The Alteryx expansion has been part of IT's initiative to push more self-service for business users who needed more advanced analytics and statistical analysis capabilities.

And one final customer example I'd like to mention is a true testament to Alteryx's end-to-end analytics platform story. Royal Dutch Shell PLC, one of the world's leading energy companies, has embraced a new culture around data analytics to continually reimagine business processes. As Shell's Alteryx user base continues to expand, including 88% of lab users across multiple business units. Shell recently added Alteryx Connect to enable their analysts and data scientists to quickly find the relevant data assets within their complex enterprise infrastructure.

The variety of applications of our platform across Shell is quite impressive. Here are a few examples. Shell's upstream operations team uses predictive analytics capabilities in Alteryx to optimize the ordering, storage and utilization of spare parts inventory for onshore, offshore oil rigs, ranging from well-head to pipeline parts. Shell's carbon capture and storage facility in Canada, called Quest, worked with the analytics team to develop atmospheric monitoring algorithms in an effort to minimize environmental impact. The Shell Downstream Lubricants Supply Chain used Alteryx to develop an award-winning suite of tools that provide critical information on inventory, margin, forecast accuracy and blending options. These tools have taken manual processes built in Excel and are into a fully automated end-to-end workloads, making quality data available significantly faster than was previously possible.

Shell's Downstream Trading Compliance team uses Alteryx to help monitor operations across multiple markets, achieving compliance with current regulations and creating transferability as markets and regulations develop.

Shell Exploration had a highly manual process for analyzing information coming back from drilling campaigns. Using Alteryx, they have constructed a new well portal where various subject matter experts can visit to analyze future extraction opportunities. And finally, Shell Downstream Retail units is just starting on their Alteryx journey, using Alteryx to digest multiple data sources and transform manual monthly processes into accurate, error-free and automated reporting around performance of their marketing campaigns.

These are all great examples of the value of our extended platform. The improvements we are making in our land and expand model, and the role the emergence of the CDO is having on our business. We are realizing a demonstrable increase in the percent of new customers that expand, while the timeframe for the expansion has shortened. This has led to improved sales and marketing efficiency, a lower customer acquisition cost payback period, and improvements in operating income and gross margins, all of which Kevin will touch on in a moment.

In summary, we are very pleased with our results for the third quarter. Not only did we achieve strong financial performance, our thrill of solving method is resonating with customers and we are getting recognized around the world for our analytic capabilities We were recently announced the gold winner of the first-ever Gartner Peer Insights Customer Choice Award, driven exclusively by customer feedback. Out of the 59 vendors reviewed, Alteryx ranked #1 in business intelligence and analytics capabilities. We accept these acknowledgments humbly and we expect to continue to expand our platform to address the needs of business analysts, data scientists as well as Chief Data Officers around the globe. We believe our opportunity is large and growing and we are well positioned for long-term success.

With that, for more color and commentary on our Q3 financials, let me turn the call over to Kevin Rubin, our CFO. Kevin?

Thank you, Dean. I am also very pleased to be speaking with you this afternoon. I will begin with the third quarter results, followed by reviewing our fourth quarter and full year 2017 guidance.

Let's start with our third quarter results, which were highlighted by 52% revenue growth, positive non-GAAP operating income and positive cash flow from operations. Revenue was $34.2 million, an increase of 52% year-over-year. International revenue was nearly $8 million for the quarter, an increase of 83% year-over-year and 23% of our Q3 revenue.

In the quarter, we added 231 net new customers and ended with 3,054 total customers. We delivered a strong Q3 dollar-based net revenue retention rate of 133%. As Dean mentioned, this is the fourth consecutive quarter above 130%. Please keep in mind that our dollar-based net revenue retention can fluctuate a bit, depending on mix of business and seasonality.

Before moving forward, I would like to remind everybody that unless otherwise stated, I will be discussing non-GAAP results. Non-GAAP measures, including our guidance for the fourth quarter and full year 2017, excludes stock-based compensation, acquisition-related adjustments, including amortization of intangibles, changes in fair value of contingent consideration and related income tax adjustments, offering costs related to our follow-on public offering and impairment of long-lived assets.

In addition, non-GAAP net income loss per share, basic and diluted, excludes the accretion of our Series A redeemable convertible preferred stock outstanding prior to our IPO and non-GAAP weighted average shares used to compute non-GAAP net income loss per share reflects the conversion of preferred stock into common stock as if such conversion had occurred on January 1, 2017 and 2016, respectively.

In periods of non-GAAP income, we adjust non-GAAP weighted average diluted shares outstanding to include the effect of diluted shares. A reconciliation of GAAP to non-GAAP financial measures has been provided in the financial statement tables included in the earnings release we issued earlier today.

Our gross margin was 86% in the third quarter, an improvement compared to 82% gross margin in the third quarter of 2016. Gross margin was positively impacted by a larger portion of subscription revenue. We have also seen more and more users leveraging our community site for self-service support and related inquiries, engaging with other users, partners and Alteryx employees, resulting in lower overall support costs. I would, however, like to note that with respect to the recent launch of Connect and the upcoming launch of Promote, and the sophistication of these offerings, we expect that we will require some additional support services. Therefore, we believe we may see slightly higher service and support costs going forward.

Operating income was $885,000, which equates to an operating margin of 3%. This was ahead of our guidance as it -- and is an improvement compared to an operating loss of $2.9 million or a negative 13% margin in the third quarter of 2016. We continue to focus our sales and marketing investments in Q3 on programs to drive awareness and adoption and through the addition of quota carrying salespeople.

During the quarter, sales and marketing headcount increased 17% year-over-year, the majority of which was in revenue-generating functions. We also accelerated some investments in countries within Europe and Asia, as we discussed with you last quarter, and expect to see the full run rate of these investments in the fourth quarter. In addition, the business, overall, benefited from stronger performing customers, driving a higher dollar-based net revenue retention rate and more efficient revenue in the quarter.

In R&D, as we focus on innovation and delivering more value through our platform, we have increased the pace of investment in 2017 versus 2016. Net income was $1.2 million and net income per share was $0.02 based on 62.3 million non-GAAP weighted average diluted shares outstanding. While we were slightly profitable this quarter, we will continue to focus on investing for longer-term growth, including international expansion and building out the sales and marketing and product development teams. As a result, we do not expect to achieve sustained profitability in the near term.

Turning now to our GAAP balance sheet. As of September 30, we had cash, cash equivalents, short-term and long-term investments of $182.6 million compared with $182.7 million as of the end of June. We generated $748,000 of positive cash flow from operations in the quarter and $6.4 million through the first 9 months of 2017.

Finally, we ended the quarter with 515 employees, up from 413 employees at the end of the third quarter 2016.

Now let's review our guidance. For the fourth quarter of 2017, we expect GAAP revenue in the range of $35.5 million to $36 million. We expect our non-GAAP operating loss to be in the range of $1 million to $2 million and non-GAAP net loss per share, basic and diluted, up $0.02 to $0.03. This assumes 60 million non-GAAP weighted average shares outstanding, basic and diluted. For the full year 2017, we are raising guidance and now expect GAAP revenue in the range of $128.5 million to $129 million, representing year-over-year growth of 50%. We expect our non-GAAP operating loss to be in the range of $9.2 million to $10.2 million, and non-GAAP net loss per share, basic and diluted, up $0.15 to $0.17. This assumes 56.4 million non-GAAP weighted average shares outstanding, basic and diluted. We continue to expect to generate positive operating cash flow from operations for the full year 2017.

This was another strong quarter for Alteryx. We are pleased with the traction our solutions are gaining in the market and the value we deliver to our customers.

With that, we will open up the call for questions. Operator?

(Operator Instructions) And our first question is from Brent Bracelin from KeyBanc Capital Markets.

I've got a couple here. Starting with you, Dean. I wanted to dive into the seed expansion opportunity. Obviously, the net retention rates here continue to be very high, I think over 130% for the last 4 quarters. So as we think about your ability to kind of expand in the existing customer base, how much low-hanging fruit is there? What's your best guess relative to footprint penetration today in installed base versus opportunity?

Great question, Brent. Let's not forget, first and foremost, that this is a $28 billion addressable market. There's 30 million disenfranchised data analysts that are living in Excel. And so the audience for the Alteryx platform is enormous, with a little over 3,000 customers and a relatively small footprint, overall, I think we probably are in a low single digit in the very most penetrated accounts. So if you think about the market in general, there's about 8% of all employment is data workers. We're nowhere close to being penetrated anywhere, North America or any other part of the world, certainly, not in any other verticals we service today.

Got it. Very helpful context there. And then shifting to the land opportunity. You've had very steady new customer adds, above 200 a quarter for quite a while here now. Obviously, you're turning profitable a little earlier than we had thought. What's the plan from a sales investment standpoint? Are you kind of behind plan there to drive new land opportunities? Are you comfortable with the range you're in now? How should we be thinking about the sales investments, particularly as we kind of think about your thinking around 2018 and additional sales investments to drive growth?

So we continue to invest pretty heavily in the organization, not just in sales and marketing headcount, but in our development headcount to continue to evolve the platform based on what customers are looking for. We continue to invest internationally. You saw that we had 83% growth in international business. We took business in 68 countries around the world. So we are actually hiring as aggressively as we can. Some of that, you don't necessarily see in the quarter, summer months were kind of tough to do some hiring. And then, you wouldn't get the full effect at the end of Q3. But we are investing pretty heavily, even back to the 2 acquisitions we made this year. So I don't think we're behind at all. I think we're careful about what we're doing. I think that part of the net retention expansion that you see, which went up 4 full points from the same period last year, is that we actually use our own platform to prosecute our own analytics to better understand the kind of cohorts we want to bring on as net new customers. And I think you're seeing better execution across the board by the sellers we have and our enablement for the sellers that we're bringing on board. And it's illustrated in the calculated reduction in [GAK], the improvement in margin in and, yes, that little surprise on operating income.

Fair enough. Last question is more on the technology side. I know it's early days with the predictive analytics tools, but anything -- any trend-wise pop-up this quarter relative to the mix of customers, downloading those predictive analytical tools more so then you saw in prior quarters? Or is it still very early?

I think it's somewhat anecdotal evidence, but we continue to see predictive capabilities and spatial capabilities become a big part of the analyst's day. As we've reported before, only about 15% of the outbound activity from Alteryx goes into a visualization products and an increasing percentage continues to evolve into the higher-order advanced functions. And I think what we're also beginning to see, as we've done into this co-innovation program with our Alteryx Promote product that will be released in Q1, we're starting to see people do some really interesting things in co-innovation around building our models and importing Python scripts to build machine-learning algorithms that can instantly be deployed on the Promote platform, really revealing the true benefits of data science. And so over time, I think you'll start to see Alteryx users community become more and more data science-y as evidenced by just the kinds of conversations that happened at our Inspire Europe conference.

Our next question is from Greg McDowell from JMP Securities.

Dean, one for you first. I wanted to ask specifically about Alteryx Connect. You mentioned it was GAed in mid-August and it sounds like you have really positive traction so far. I think you used the words meaningful demand. I was just hoping you can maybe walk through some typical use cases and how we should think about Alteryx Connect landing as a standalone product without the core Alteryx versus it being up-sold with an existing Alteryx customer? And then I have one follow up for Kevin?

Good question, Greg. So I think that what we've been hearing from customers for a long time is that as we begun to allow them to liberate analytics in a code-free, in a code-friendly environment in their enterprises, for both citizen data scientists and the trained statistician, organizations are realizing that there's data assets all over the place and new data assets are being curated each and every time somebody pushes the go button inside of Alteryx, or every time they touch a Tableau dashboard, or write an Excel spreadsheet or a CSV. And so the proliferation of data and the myriad of persistence layers that exist today has created a -- somewhat of a complexity for organizations to make sure that there are single versions of the truth, not in the persistence layer, but in the analytic processing. And so for the last couple of years, we've been hearing Chief Data Officers and analytics leaders tell us that they wanted to have a social collaborative metadata management layer, a data cataloging service that would basically tee up all of the relevant assets an analyst might need to begin their analytic journey. You probably heard me say before that the hardest problem in analytics is knowing what question to ask. The second hardest issue in analytics is knowing where to begin. And our Connect platform actually brings chaos to order when it comes to finding the right assets, whether it's dashboards, reports, even Alteryx workflows or models, and it allows the analyst to begin their analytic journey, and we're seeing early adopters become far more efficient at finding and beginning -- finding the data and then beginning their analytic journey. But -- and I will say, it's still early in the process. We just rolled it out in mid-August, good traction for both new and existing customers. We don't intend to sell it alone. I suppose there'll be some candidates who start their buildout of the Alteryx platform with either the Connect or the Promote component. But we see analytics as an end-to-end process, and that was our intention with both end caps to the platform.

That's helpful. And Kevin, just a quick 2-parter for you. Metrics look great across the board. One that really stood out is gross margin of 86%, and that's absolutely best-in-class for subscription companies. And acknowledging that might dip down a little bit to support Connect and Promote. I guess, in order -- in terms of impact, are we just talking about maybe 1 percentage point or 2, so back to the 83% to 85%, something deeper than that? And then a quick second part of my question is just -- I know billings isn't necessarily the best way to track the business. But as I look at the change in deferred from Q3 to Q4 last year, it was a very meaningful increase and I was just wondering, as we think through change in deferred for Q4 this year, is there anything we should be aware of in terms of what transpired from Q3 to Q4 from a deferred perspective?

Sure. So let me hit the first one. So in terms of margins, we haven't changed our view in terms of our long-term target margins, which were pretty consistent with what you mentioned. So I think the order of magnitude that we're, I guess, describing with respect to a slight increase in support costs around Connect and Promote, it's just that. In terms of the calculated billings, I mean, yet, to your point, we don't guide to it specifically. I think that I would look to our Q4 guidance as a directional view with respect to how we think about Q4 deferred.

Our next question is from Mark Murphy from JPMorgan.

I wanted to ask you, Dean, when you think about the output targets -- if you had to take a stab at trying to stack rank, all of those environments, I guess, you would call them the consumption layers, to which Alteryx is sending the data. So if you have spreadsheets and documents and databases and dashboards and visualization products, et cetera, in what order would you place those? And are you seeing -- I think you alluded to this a bit, but are you seeing anything where you could comment deeper in terms of the mix shift?

So it's a very long tail, I will say that. It basically illustrates the need for a platform like ours to be agnostic to both the persistence layer and the consumption layer. So we see fragmentation on both ends in a very significant way. I don't think we've seen any meaningful shift in the last few months. I would say that we're beginning to see more people go to things like Redshift. But we see a tons of customers who write out the PDFs, many are going into Power BI. We're seeing everything from spreadsheets to PowerPoint to -- back to relational databases as IT comes into the mix and Chief Data Officers try to leverage Alteryx to do a lot of the heavy lifting in the data centers. But it's a very long tail. We support pretty much everything out, and I think it illustrates the need for analytics to be consumed in a pane of glass that's appropriate for the user or the consumer of those analytics. And I think, over time, as people begin to embrace our Promote product beginning next quarter, we'll start to see people consuming analytics in web and mobile apps, and that will be the destination. So our thesis has been, you should be able to visualize all of your data from the beginning of your asset find in Connect, all the way through your analytic buildout. And then ultimately, by the time you're done, you're actually deploying machine-learning algorithms or complex analytic processes in meaningful panes of glass. So just as an illustration within Promote, while we haven't GAed it yet, we have a tool that allows you to instantly deploy machine-learning algorithms to your Salesforce instance. And we believe that, that is a logical place where a lot of these algorithmic processes will occur. So it's a long tail. It hasn't shifted much, and the fragmentation is to our advantage.

And just to clarify, Dean, on the last point. You think salesforce instances would be a very common target for machine-learning algorithms? Is that what you're trying to describe?

I think it's one of many. I think that if you look at every website, every mobile app, some of the early adopters of Promote in our co-innovation program are actually using it to redirect traffic for next-best action models, for websites, picking -- helping determine what channel of distribution should be -- the customer should be going to. So I think we'll see a long tail of consumption from Promote as well.

And then a final question for Kevin. The sales and marketing line seemed to show some stronger efficiencies, maybe than what we might have expected in the quarter. Can you walk us through the dynamic there? Basically -- it does sound like that you're growing the headcount. So to what extent did that occur, maybe more on the discretionary marketing side versus to what extent, or if at all, was that linked to sales commissions or anything else that might have just correlated on the -- to the booking kick side at all.

Thanks, Mark. So the vast majority of the increases were directly related to sales and revenue-generating capacity. So we are heavily focusing the investment in being able to grow our sales organization. Marketing, in the quarter, we did have our Inspire Europe conference, but that's not a meaningful percentage of the increase.

Our next question is from Michael Turits from Raymond James.

I guess, I'll start with Kevin, and then move on to Dean. So maybe just on Mark Murphy's question. I think it's -- we see where some of the growth in OpEx came, but it came in lighter on OpEx, better efficiencies. So what really drove the ability to get efficiencies this quarter? It looks like a decent guide into next quarter, not quite as strong. And then anything you can tell us directionally about how we should start to think about OpEx into next year and margins, that would be helpful?

Sure. So I mean, maybe let me elaborate a little bit then on the answer. The investments that we are making, largely in people, I think, to the prepared remarks, we did see those come in later in the quarter, so you don't see that there'd be expense associated with those investments on a full run-rate basis. You'll start to see those into Q4. I think in terms of just the strength of the business, I mean, as evidenced by continually -- a continuing strong net revenue retention of north of 130%, we have been able to drive more efficient revenue through the model. I mean, we have better recognition, we've got more products to be able to talk to, it's a bigger enterprise platform story. So I think we're seeing all of those dynamics play out in the results that you're seeing from a revenue growth versus investment perspective.

Okay. And then Dean, you entered into this partnership with Plotly to do data visualization at the high end for analytics. Is this, in any way, a shift in your strategy around visualization towards wanting to engage more there?

No. Actually, we have views on where visualization has been and where it's going. Our customers need both static and interactive visualizations, so we have actually put a big effort around improving the functionality and flexibility of our own static visualizations, easy-to-design charts and reports and dashboard-like environment. But we also see a change around how visualizations can impact analytic outcomes. And so the thesis there is that you shouldn't have to wait until the end of your analytic workflow to see and understand your data, that you should be able to instantly connect it to a data source that you've found, let's say, through Connect, or for any other use case just directly to a database, right now, expect and see what's going on with the data. Null cells, completed fields, just information that will give you relevancy as to how useful that asset is going to be in the analytic process. So we believe that getting that knowledge improves the efficiency of your analytic process, so that by the time you're done, you're actually ready to deploy the analytics in metaphors that matter across that fragmented consumption layer that we talked about earlier. And Plotly is a great platform to do this in. It's probably the broadest, most widely used interactive dash boarding capabilities, and we've had great success with it so far.

Our next question is from Brad Sills from Bank of America Merrill Lynch.

I wanted to ask about comments that you -- were made earlier on just sales efficiency. Obviously, margin came in quite a bit better than where we were modeled and primarily from sales and marketing. And I think you alluded to the customer acquisition cost coming down, and then the cost of maybe the expansion business, if I'm also reading into it correctly, also coming down. So if you could just provide a little color on what's driving efficiencies in both ends of the business, that would be helpful, please.

Thanks, Brad. So I think, again, we have done a better job, as Dean described, using Alteryx, frankly, to understand our cohorts and our sales motions, and we're seeing that play out in the results. We are seeing that the net new logos, the new customers that we're landing, are better-quality logos, which, just by the nature, provides for a more efficient expansion cycle thereafter. We're seeing improvements in net revenue retention that's a result of some of the dynamics that I had previously mentioned. And so while we certainly are investing in the backdrop of growth, we have seen our incremental revenue come at a more efficient pace than we've seen previously.

That's great. And if I may, one more, on international, you mentioned Japan was strong. Any other countries where you maybe perhaps (inaudible) to the upside in terms of progress since entering those markets or ones where you see potential upside from here?

So we're investing in quite a few earliness. As I said, 68 countries, we did business in this last quarter. We're seeing trail downloads around the world and lots of other countries that we have no partners in, for that matter. So we continue to invest in EMEA today, both in London, Munich and our effort around France and in opening an office in Paris. We'll look at other markets next year as well. In APAC, we have opened up Singapore and have begun to staff that office. We are currently in search for a country leader in Japan. And we have meaningful customers. Softbank being just one of the many in Japan. So -- Australia is a big market for us as well, and we have no feet on the ground there. So we'll continue to invest heavily in these markets where we see early traction on land, where we can then put those customers into our [cohort files].

The next question is from Derrick Wood from Cowen and Company.

This is Jim Fitzgerald sitting in for Derrick. So how are you guys feeling about current go-to-market efforts? And are there any new initiatives that you guys are working on, especially with respect to how you're positioning yourself for showing as a data science platform?

That's a great question. So we've continued to see the evolution of the needs for an Indent platform coming from CDOs, IT, analytic leaders. So we continue to invest in that. It hasn't changed our go-to-market motions at all. It does require a bit of a change in our sales enablement, which we're going through. We'll have a sales kickoff that will enable our team, for example, on the sales playbooks for Promote. But it hasn't changed our approach at all. We haven't had to hire a different persona to sell. We still have a relatively short and inexpensive land cycle. That said, we're beginning to see expansions happen earlier and participation and expansions happen at a higher pace. So it hasn't resulted -- there's a shift in the platform needs and the personas that we're selling to, bottoms up to analyst, or top down from CDOs and analytic leaders. Hasn't had -- it hasn't required a change in our sales playbooks at all. Mostly, just sales enablement.

Okay, great. And then one more, if I may. As CDOs are getting increasingly involved, have you noticed any change in the sales cycles or maybe the deal sizes?

No. I think, even to my earlier point, we -- even with CDOs, we still have -- we don't let the sales people go off-playbook. We still land with typically 1 or 2 seats in a 45-day cycle. But what we do -- what the CDO brings is a need to harness the networking effects of people, data and technology that isn't always evident when you're selling to a line of business leader or certainly, a line of business analyst. And so what happens is, we tend to get time to value very, very quickly in a simple use case that causes great consternation for the customer. That use case then drives continued expansion, and relatively quickly because the CDOs have a mission to liberate analytics across the enterprise. And we're beginning to see more of that today. I think Gartner has said that by 2020, 80% of Global 2000 companies will have CDOs in seat. I don't know if that's actually progressing as quickly as they think, but we are seeing -- and perhaps, this is some of the impact we see on our network retention numbers continuing to improve, is that we do start to see more CDOs coming in to the picture earlier, or people who are proxies for CDOs, even if they're IT that are spending line of business in IT requirements.

Our next question is from Jesse Hulsing from Goldman Sachs.

I wanted to follow up on the question around Connect, and then, I guess, also probably can include Promote in this as well. I guess, in an average customer how much of an uplift, percentage wise, do you think Connect could provide? And also, kind of the same question for Promote, how much of a lift could this be?

Well, we really haven't factored in any of those into our own models. I think we went into the extension of the platform based on the needs that customers had. I guess we're so early in the cycle, we don't really know what the impact is necessarily going to be. We would hope that it would add to stickiness in the accounts. It might be that it could lead to additional designer sales, as we've now made it easier for people to find the data assets that matter. It could lead possibly to more Server sales if scientists are now able to instantly deploy models inside of Promote. So there's lots of things that could happen. We don't have an update at this point. We're still early in the game, we haven't even launched Promote. We have a bunch of people in co-innovation efforts. But this is a great question that we should get from you again Q2, Q3, Q4 of next year.

That's fair. And I thought the Shell customer example was pretty interesting, in that it kind of showed your evolution from a tool into more of an end-to-end analytic solution. I'm wondering, there's been a lot of questions around sales efficiency on the call. How much involvement do sales people have in that type of expansion? I mean, is the sales rep holding the hands of the buyers all the way through and helping them go to procurement and do that sort of thing? Or is this kind of a natural organic expansion where sales involvement maybe isn't as heavy as you would see in a typical enterprise software sale?

Yes, I think the latter is what's going on. We -- obviously have sellers who are tied to customers, and they do what's necessary to move the sales process through the system. There is some enablement that goes on to get customers to that first use case, although, that's relatively small, too, as evidenced by the services or the lion's share of our business being subscription-based as opposed to services-based. So I think the key is finding the alpha users who have a demonstrated use case that can spend the time to get to an outcome. Most of our customers will report magnitude improvements and operational efficiencies, processes that went from 100 hours to 5 minutes, they get all their time back. They then want to ask bigger questions, harder questions with more data and it particularly, when CDOs are in the mix, it's easy for -- it's an easy decision when they see the kind of paybacks that they're getting. So I won't say it's, hands off organic growth, but it's not the heavy enterprise selling that you used to see with stack vendors deep in IT.

Yes. And Kevin, one quick one for you. 606 or ASC 606, any update there, particularly on the rev rec side?

Yes, there's not -- Jesse, there's nothing beyond what I think we've mentioned previously. We elected to maintain the private company timeline of adoption, which puts us out to the end of '19. So we do have an opportunity to see how the standard continues to be interpretive -- interpreted and involved and can address it as we get closer to 2019.

(Operator Instructions) And our next question's from Bhavan Suri from William Blair.

Can you hear me, okay?

We certainly can.

I guess, Dean, you and I have chatted about this before. But I'd love to understand a little bit on how the trajectory of the use case split went in terms of advanced analytics versus data prep. But more importantly, when you look at sort of whatever the 600-odd partners, are they sort of tied to one use case more than the other, or are they sort of pretty broad? How should we think about that? So how are use cases and sort of how are partners playing into that split between the 2 use cases of data preps and then sort of the advanced piece?

Fair question. So I think it's probably mixed. I think that there are some partners that are tied more to a single use case or a vertical, and some of them probably understand the land -- the data prep and blending component to a point solution downstream. In many cases, those are people who might have a [click] reseller arrangement or a Power BI or a Tableau arrangement. But we're seeing a growing number of analytic consulting firms around the world embrace Alteryx for all of their analytic work. And we've probably indicated this before, but if you think about all the major consulting firms from KPMG to BCG, to Accenture and Deloitte and on and on and on, they're all using Alteryx for their analytic processes. So I think people are beginning to realize that data prep and blending is the on-ramp to higher-order analytic processes, not just descriptive and diagnostic use cases that you would typically end up seeing or visualizing in a BIS product. And so that's being to change. I personally have been involved in a number of partner meetings recently, helping define the growth path for partners around becoming more data science-y, and people are now acquiring PhD statisticians to help build out their own team to take advantage of the platform that Alteryx has built.

That's very helpful. Then, you touched on a little bit, when you're talking about sort of visualization and static, I've mean, obviously, given the real-time world we live in, there's this whole concept of data in motion. And I was wondering sort of the -- how Connect plays into data-in-motion. And sort of how -- I was trying to do real-time pieces, data prep. And, obviously, advanced statistical analysis is critical, but how are you seeing that play out? Or is it still early, and sort of trying to integrate data-in-motion with some of the existing sort of ways people look at some of the advanced statistical analysis?

Well, I think data-in-motion is alive and well. I think that if you take a look at Promote, for example, because of the scalable, flexible architecture of Promote, you're able to run machine-learning processes in a fraction of the time of machine-learning algorithms that were previously repositioned or reconstructed by IT. That's been determined in data science as the quant builds a model. And according to Rexer Analytics, only 13% of PhD's say their models always gets deployed. So without Promote, data science could be dead. And with Promote, you're able to instantly run your models, and we've seen use cases, even in our co-innovation effort, where models used to take 8 minutes to run in previously architected solutions typically done by IT, and they're running in milliseconds now inside of Alteryx Promote.

Yes, that's cool. I'm going to squeeze one quick one in here just because of us talking about some of the legacy guys and some of this modeling PhD space. But have you seen SAS, or even IBM with the acquisition of SPSS and Intelligent Miner and other things, start to think through this, the same way you guys have from a let's expand the usage, and sort of the deployment of the visualization and ease-of-use capabilities? I know everyone's talks about Tableau competitively and other things. But I'm more focused on some of those guys, are they sort of innovating that space? How do you view them competitively?

We have a fairly detailed orientation around the strategic landscape. And we watch what people are doing, we have the backdrop of knowing what customers want in a general-purpose analytics platform. I don't see them necessarily being the innovators in the space. Clearly, they are still the leaders, but typically in an IT environment, and it's clear that we've gone from these systems of record deep in IT to systems of engagement out in the line of business. And so we don't necessarily -- we worry about all the people who could potentially compete with us. We watch what people do. But I'm not sure the innovation coming from the stack vendors of yesteryear worry us as much as new players who might be well funded to have a strategic roadmap that kind of looks like ours.

This concludes the question-and-answer session. I'd like to turn the floor back over to management for any closing comments.

Thank you, operator. Before adjourning the call, I'd want to thank all of you for being with us. I want to thank all of our employees for their contribution to our continued success this past quarter and a big shout out to all of our customers and partners around the globe for their continuous support of Alteryx, the company, and our end-to-end analytics platform for the enterprise. Thank you for joining us today, and we look forward to speaking with you all again very soon.

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