Alteryx (AYX) Q1 2020 Earnings Results Review

Alteryx released their Q1 2020 earnings report on May 6, 2020. Results were mixed – they beat Q1 estimates for revenue and were almost inline on EPS. However, the headline surprise came with a substantial reduction in Q2 revenue and EPS estimates as a result of a COVID-19 driven slowdown. The stock initially dropped by as much as 10% the next day, but gradually recovered to close down 2.9%. Following the results, most analysts lowered their price targets. Alteryx also made a major product announcement on May 11th and has recently posted some very interesting new job openings. There is a lot to unpack here and I will try to cover all the new activity, in addition to the earnings report. Then, we can look at what this means for the investment thesis going forward.

Dissecting Alteryx Product Architecture

Before I jump into reviewing the financial results and other announcements, I wanted to give my perspective on a couple of items that could be sources of confusion for investors, relative to the Alteryx product offering and its pricing structure. Granted, perception can be reality, but some explanation might help investors understand the landscape and draw their own conclusions. More importantly, we have some hints of future technology architecture direction, which further mitigate possible concerns and could open up new audiences to the Alteryx product offering.

On-Prem versus Cloud Solution

I have heard some commentary relative to the Alteryx offering that it might be hindered because the software is deployed “on-premise” and is not cloud-based. At a high level, investors know that cloud migration generally involves replacing legacy software packages that are deployed on local corporate networks (on-prem) with software that is run on servers hosted remotely with cloud vendors (cloud). Additionally, we have the concept of SaaS, in which the vendor runs the software for the customer from a central location. Alteryx’s offering falls somewhere in between these currently, with indications that they are evolving quickly.

Putting aside terms like cloud-based or SaaS, I think the best way for investors to understand Alteryx’s software architecture and distribution model is by considering these elements:

• Client. The software that runs on the user’s local machine (desktop, laptop, tablet, mobile phone, etc.)
• Server. The centralized software system to which the clients connect to exchange data, kick off background jobs and interface with other systems.
• Software Distribution Model. The way that software updates are distributed and applied to both the client and server.
• Pricing Structure. The way that a customer is charged for licenses to use the product.

Client

Most software clients can either run as a native application that is installed on the user’s operating system directly or runs within a web browser. The user’s machine in this context can either be their desktop/laptop, a tablet or a mobile phone. Applications running natively have inherent advantages over browser-based clients. The code is compiled, so it runs faster. Native apps can control the UI more precisely, can get more resources (CPU, memory) from the machine and aren’t confined to the browser’s runtime sandbox (for security reasons). The downside to running a native application is that software updates have to be triggered by the user. In a browser app, the user just hits “refresh” to get a new version. In a native app, they have to kick off an installation process. However, native app installs can be pretty seamless these days. Assuming users can connect to a central app distribution store, like the way the Apple App Store works or the company’s own web site, then updates are pretty easy to manage.

Some software companies make their clients available as both a web client and a native client. Slack is a good example. I can run Slack in a web browser, or I can download the Slack app onto my Mac and run it natively. As an active user of Slack, the native client is much better. The UI is tighter and much more responsive than the browser version. Also, the app allows me to toggle between multiple workspaces at once, get desktop notifications and launch the experience quickly from my dock. Accessing documents from my local machine to share is easier.

Another client application model to consider is how we run software on our mobile devices. I can either use a software service through a browser on my iPhone (mobile web) or I can download that company’s app from the App Store (native app). Generally, the native app performs much better for the same reasons as on a desktop. It can access system resources directly (camera, touchscreen, location, etc.), can paint the screen on a per pixel level and will run faster because the app is compiled code. Of course, accessing a company’s web application through a browser is usually the initial touch point, as the user doesn’t have to go download an app on the first encounter. So, it is critical that a software provider also have a web experience that supports basic functionality. Eventually, heavy users do go download the app, as it’s easier for repeat use (think eTrade or Twitter).

In the case of Alteryx, the client software is the Designer product. It runs as a native application on the user’s machine. The software can be downloaded from the Alteryx web site, much like how other companies distribute a native client. The Designer application is currently only available for the Windows OS, specifically MS Windows 7 or later. The user’s machine needs to be fairly powerful to run the software, with at least a quad core CPU, 500GB-1TB of disk space and 8-16 GB RAM. For this, the user would need a decent laptop or desktop machine. You couldn’t run this on a mobile device, but I don’t think a data analyst is going to perform heavy data prep, blending and machine learning tasks on their phone.

On the surface, the limitation to only running on Windows OS might be considered a knock. Slack runs on both Windows and Mac natively. However, I think this is okay. Alteryx users, at least for now, are going to be power users. This would presumably be the primary tool they use daily. So, needing to provision a Windows machine to do their job is a reasonable assumption. Also, it is fairly easy to run the Windows OS on a Mac through a dual-boot tool, like Boot Camp.

Given these circumstances, I actually prefer that Alteryx focus its developer resources on the Windows-only client. If they tried to roll a native Mac client, they would have to effectively double the dev resources for the Designer team. Mac apps are built in another language framework with different OS APIs.

Alteryx Designer is necessarily a fat, native client application. This has a couple of benefits, in my view. First, due to the type of workload, I think a native client provides much better performance. The Designer client can support active data connections to various data sources, both local to the machine (load a csv) or remote (connect to another SaaS or database). Large datasets can make full use of the machine’s local memory and disk. This would not be feasible in the memory space of a browser and sophisticated data manipulation would much more difficult using interpreted Javascript data structures. Second, because Designer is a fat client, it can be run independent of a server. In fact, many enterprises start with just Designer licenses. By running independent of the server, it also doesn’t require a network connection back to the server to be useful.

Think about it like Excel. Many data analysts today still run Excel natively on their desktop because they have large datasets and create complex logic in macros. They can import data from external sources through connectors. As data is updated, it can be processed locally into different views, pivot tables, etc. This would be much slower in a browser.

With all that said, Alteryx management has discussed the idea of creating a browser-based version of Designer. The CEO just mentioned this at a recent analyst conference and the CFO discussed this further in an interview with TheStreet. It sounds like large enterprise customers are asking about the feasibility of a browser-based version of Designer in order to facilitate distribution of clients to a large number of users.

Thick client on a desktop is much harder to manage than just simply entitling a user and having them go to a browser. We did show a prototype version of that at our conference last year. That’s kind of the direction we’re going.

But really, the focus is on, how do we help our largest clients massively deploy Designers through a browser rather than a thick client install.

Alteryx CFO, INterview with theStreet, May 12, 2020

So, it appears that Alteryx is already moving in the direction of making a browser-run version of their Designer application. My assumption is that this would be stripped of some features, in order to be performant, and offered on a different licensing arrangement than full-featured Designer. Call it Designer Light (Update – It seems this will be called Designer Express, based on a product manager job posting from yesterday. I discuss below.)

What really caught my eye, though, is the comment around “massively” deploying Designers. Given Alteryx’s new motion to empower every employee at an enterprise to be a data worker, this might represent a way to dramatically expand the organizational reach of the Alteryx platform, beyond just the power user data analysts, and generate more client license revenue.

Server

The server side of the software model provides something for the client to connect to. For thin clients, like those that are browser-based, the server handles most of the heavy processing work. The browser really just renders the UI, while the server processes every user action, controls business logic and all data access. Without the server, the browser is pretty useless. For a fat, native client, the server tends to handle functions around sharing data with other clients and running jobs remotely to free up resources on the user’s machine.

The server software can be run on a single, more powerful computer, or a cluster of machines that all run the same software and are stateless (SaaS architecture generally). In the stateless configuration, there is usually a database (or set of databases) that the servers share. For the case where the server software is installed onto a machine (the server), it is generally managed by the enterprise. This server machine can be located on the enterprise’s local network (on-prem), in a data center they manage (hybrid) or onto a machine provisioned by a cloud provider (self-managed cloud). In these cases, the enterprise controls the server machine, applies software updates, ensures security requirements meet standards, etc.

For most SaaS offerings, like Slack, the vendor completely manages the server side of the model. There is no option for an enterprise to run the Slack server themselves (that I know of). In this case, the software provider handles all software updates, monitors performance and manages security controls. They also go to great lengths to isolate each customer’s data through federation, but at the end of the day, resources are shared. This might not work for the most security conscious organizations.

To make matters more complicated, some software vendors offer support for both server management models. They allow the customer to either download the server software and host it themselves, or the customer can utilize the vendor’s server cloud and not have direct control over the servers. Some examples are:

• Microsoft – Active Directory or Azure AD
• Elastic – ECE or Elastic Cloud
• MongoDB – Enterprise server or Atlas

In Alteryx’s case, the Server product represents a downloadable software package that is installed onto a server machine that the enterprise manages. This can either be a server machine that sits on premise or on a rented machine in the cloud. In either case, it runs on Windows Server 2008R2 or later. The machine specs are heavier than for Designer, which we would expect. Virtual environment support is available for Nutanix, Azure, AWS, VMWare and GCP (basically the main cloud vendors and hybrid solutions). So, it can be deployed in the cloud.

The requirement to run Windows Server on the server side isn’t really a concern. It is just as easy to provision a Windows host as a Linux one on most cloud providers. And this shouldn’t be viewed as a limitation, as SaaS vendors use both OS versions to run their centralized server software (although mostly Linux). It wouldn’t make sense for Alteryx to rebuild Server to run on Linux, at least for the self-managed version.

Since the Designer client can operate autonomously, the Alteryx Server focuses more on collaboration features, offloading work and handling external requests. Here is a short list of functions:

• Saves workflows and data models for other users to access.
• Runs data processing jobs remotely, so that the user’s desktop resources aren’t consumed (although Designer can run the jobs too).
• Handles scheduling of workflows as jobs that run with some periodicity.  Admins can also manage the active queue of jobs and control priority. 
• Provides user access controls around workflows and shared data.
• Supports some shared dashboard functionality for basic data visualization.
• Hosts data connectors that can be shared among Designer users (keeps credentials stored centrally).
• Can expose data sets as APIs that other applications can consume.
• Can export data sets to other data visualization tools, like Tableau server.

The big question is whether there would be advantages in having Alteryx run the Server product for enterprises, like what a SaaS company might do. Perhaps. The benefits would be to eliminate server management overhead for customers and speed up software updates. I don’t think there would be additional network effects, by being able to centrally view all customer activity (like what CRWD or OKTA does), but observing common behaviors might offer some insight for future product development. Regardless, the effort to centralize the Server offering into a parallel “Cloud” product, like what MongoDB, Microsoft and Elastic offer wouldn’t be very difficult. I would estimate that much of the existing Server software code could be re-purposed.

Software Distribution Model

This addresses how software updates are distributed and applied. I think we covered this in the discussion above about Client and Server. In Alteryx’s model, the customer is responsible for applying software updates. In a typical SaaS model, the vendor handles all software updates on the server side. If the client runs as a web application in a browser, the updates are automatic. If the client runs as a native app (like Slack app on desktop or mobile), the user has to update it. For Alteryx’s current distribution model, the customer handles the application of software updates to both the client and the server.

Pricing

Distributed software applications are usually licensed in a couple of ways. First, there is the traditional software license that applies a single fee to purchase a pinned version of the software. This is how big vendors like Microsoft and Adobe previously sold their software. The customer could purchase a single license for the software for a one-time fee and never pay again. If they wanted an update, they usually paid a smaller incremental fee for the newer version. With SaaS, the pricing model evolved to per user per month. This is how popular SaaS offerings, like Slack work. As SaaS became popular, the big packaged software vendors had to evolve quickly. They moved to a subscription model, in which the user pays a monthly fee. This fee gives them access to any version of the downloadable software, along with some sharing features on the server side. Microsoft did this with their Office 365 offering, as did Adobe with their Creative Cloud. For example, a user of Photoshop, which is a resource intensive native application for creating graphics, could download the Photoshop application and then get free updates as long as they paid the monthly fee. As part of the enterprise plan, they can also use some of the content sharing and collaboration features.

The Alteryx pricing model is similar to Adobe’s enterprise offering, except that licenses are purchased on a per user/year basis. Currently, each Designer license is listed at $5,195 per user/year and Server costs $78,975 per year.

On the Q4 and Q1 earnings calls, several sell-side analysts asked about the pricing of Alteryx products and the recent 35% increase in server costs. They seemed concerned about the list price of Alteryx products and whether that would impact sales going forward. I suppose it could on the surface, but product list price usually doesn’t coming into play in IT budgeting, at least in my experience. As a CTO and VP Engineering at several past organizations, I would generally review the IT budget with the CFO as a roll-up by line item on a quarterly or annual basis. Each software service, whether pure SaaS or per user license, would just be listed as a line in a spreadsheet with a total cost for that time period. We would then go through each line item and decide if the total cost was worth the value to the organization. In Alteryx’s case, a medium deployment of 20 Designer licenses and a Server license might cost $200k a year. If that was the primary tool used by the analytics team every day, the total cost might be just fine. Broadly deployed SaaS applications can quickly rack up costs as well for an organization with thousands of users. Okta, Office 365 or even Datadog could generate annual costs in excess of Alteryx for a large installation. Those might be justifiable as well. It is all evaluated through the lens of value created for the business.

On the earnings call, the Alteryx leadership team mentioned closing a 7-figure deal with a European financial company in March. In this case, I imagine the global CIO or CDO simply looked at the total cost of the Alteryx solution in the context of the value it provided for the organization. That person probably wasn’t concerned what each license cost. Granted every organization will be managing costs in this COVID-19 impacted environment, but I don’t think the licensing model is the issue. Cost reductions are generally applied across the board, like to cut an organization’s IT budget by 20%. In that case, the result might be to give up or not renew a couple of Designer licenses, but I would apply the same approach to any other SaaS tool licensed on a per user per month basis.

Also, the perceived flexibility of SaaS or subscriptions billing on a monthly basis versus annual license fees isn’t a benefit in reality. Most SaaS vendors will offer a discount for multi-year deals or minimum annual spend commitments, so enterprises get locked into a longer contract regardless.

With all that said, subsequent comments at the Piper Sandler and J.P. Morgan analyst conferences lead me to believe that Alteryx is pursuing an expansion of their product portfolio and possibly different pricing models. The CEO teased a major platform release in June, an upcoming intelligent suite and a browser-based version of Designer. These could bring alternate pricing options for Server and Designer “light” that align with conventional SaaS models.

Job Postings Provide Insight

Another source of clues about what Alteryx is planning from a product development point of view is to look at their recent job postings. I have noticed some significant changes in the last two weeks in the types of jobs posted, specifically for the technology team, that give clear insight into how Alteryx is evolving their tech stack and product architecture. A few of these jobs were just added yesterday, so this situation is evolving quickly. Some of the wording below is borrowed from the job descriptions directly.

• Technical Product Manager, Designer Express. Broomfield, CO. Will drive the tactical execution efforts for Alteryx Designer Next, the next generation data analysis application.
• Engineering Team Lead – Cloud. Broomfield, CO. Will build the next generation of the flagship Alteryx product, Alteryx Designer. Mentions that Alteryx is moving a legacy desktop application, Alteryx Designer, to the cloud. They are looking for a team lead with knowledge of micro-service design patterns. Also seeking understanding of automation and deployment techniques, including how best to leverage Docker containers and deployment pipelines. Node.js / Express.js will be the standard technology used for developing new micro-services. Other tech requirements include Docker/Kubernetes, cloud provider tooling experience (AWS, Azure, GCP) and environment scripting (Ansible, Chef).
• Software Engineer. Broomfield, CO. Basically, similar description to Team Lead role in skill set, but for multiple individual contributors. Will design and implement low-latency, high-availability and performant micro-services, using Node.js and Express.js design patterns. Experience with other server-side languages like .NET, Java / Spring, Ruby, Python, etc.
• Senior Software Engineer – OS Tools. Boston, MA. Focus on building out the open source libraries for automated modeling that Alteryx owns as part of the Feature Labs acquisition. The role will help the Open Source Tools team create and maintain innovative, open source tools for machine learning.
• Senior Front End Engineer. Boston, MA. Will be working on a new web-based data science project with a rapidly growing team. Will use modern Javascript (React, Redux) frameworks with HTML5 and CSS3 to render rich browser-based UIs. Will consume machine learning and data science APIs.  Experience working with AWS (EC2, S2, Lambda) and using automated testing tools like CircleCI.
• Team Lead, Engineering. Redwood City, CA. Will lead a team of 3-5 developers (might be multiple of these teams). Will help build a next generation platform SDK and API suite.

To interpret this for non-technical investors, these jobs all utilize modern software engineering practices and development frameworks. The architecture is cloud-focused, making use of browser-based UI on the front-end with scalable APIs and micro-services on the back-end. If anything, this is a clear sign that Alteryx is investing heavily in modernizing their tech stack with an eye towards addressing some of the architectural limitations to bring their platform more inline with typical SaaS offerings. Additionally, the reference to Designer Express aligns with comments about a browser client. The Express label makes me think this will be offered in addition to Designer “heavy”, which implies distribution to a broader audience of users and perhaps new revenue streams.

Okay, that was a bit involved, but hopefully provided some useful context for investors. Let’s move on to review this quarter’s performance.

Headline Financial Results (EPS is Non-GAAP)

• Q1 2020 Revenue was $108.8M, up 43% year/year. Consensus estimates were for $105.6M, about a 3% beat. Original company guidance from Q4 was for $105-108M. Q1 2019 Revenue growth was 51%, and Q4 2019 Revenue growth was 75%. Investors should note Alteryx has demonstrated a pattern of increasing revenue growth from Q1 through Q4 for the past two years, with a drop between Q4 and Q1 being normal due to seasonality.
• Q1 EPS was ($0.10) vs. ($0.09) expected. Original company guidance was for ($0.11) to ($0.07). Q1 included about $6M of one-time expense associated with the Annual Global Kickoff meeting, and costs associated with rescheduling user conferences.
• Q1 Non-GAAP operating loss was $3.2M, compared to Non-GAAP operating income of $1.4M in Q1 2019. This translates to an operating margin of -2.9% for Q1. However, we should consider the $6M of one-time expenses, which would have resulted in operating income if removed.
• Q2 Revenue estimate of $91-95M, representing growth of 10-15% year/year. This compares to the analyst consensus of $108M for 32% growth and represents a significant reduction.
• Q2 EPS estimate of ($0.12) to ($0.18) vs. ($0.05) expected. The miss is due to the lowered revenue forecast, as most Alteryx expenses are fixed (headcount related and they didn’t furlough any employees).
• Q2 Non-GAAP operating loss is expected to be in the range of $(9.0) to $(13.0)M. This would represent an operating margin of -11.8% at the midpoint. Also driven by lowered revenue expectations.
• Due to the uncertainly of the length of the COVID-19 situation, Alteryx withdrew its 2020 guidance. Given that the situation could worsen, this is reasonable.
• Ended the quarter with cash, cash equivalents and marketable securities of $992M.

Other Performance Indicators

• Q1 revenue generation was progressing above plan for January and February, but slowed down substantially in March, as companies transitioned to shelter-in-place. This started in Asia and Europe and then moved to the U.S. Alteryx typically closes the majority of its revenue for a quarter in the 3rd month (March). So, it’s fair to assume that the outperformance in Q1 would have been larger without the disruption to March.
• Starting in March, Alteryx experienced lower expansion rates from companies in the heavily impacted sectors of travel and hospitality, manufacturing and retail. Also, they experienced a moderate increase in churn rate most notably in Europe. European deals are primarily driven by reseller relationships, which leadership said had trouble adjusting to the new social distancing environment.
• Had a record number of adoption licenses in Q1. This means customers were allowed to use products for free initially and then would start getting charged later.
• For April, CFO commented that they “saw new business activity resume and was consistent with activity levels in April 2019”.  This is tricky to interpret for investors. On one hand, it sounds like they went from practically no new business in March to a nice uptick in April. On the other hand, equivalent business levels to a year ago would be a step back. My interpretation, colored further by the CEO’s positive interjections throughout the call, is that that they are pleased with the upward trajectory from March to April and assume that will continue. Also, it’s worth remembering that the majority of Alteryx’s quarterly business is closed in the last month of the quarter. The CEO said “I think we were pretty excited that we could continue to drive the number of logos we did post-work from home, post shock from COVID.”
• Non-GAAP gross margin was 91% in Q1. This compares to 90% in Q1 2019.
• Ended Q1 with 1,478 employees, up from 1,291 associates at the end of Q4 and 936 associates at the end of Q1 of 2019. This represents a 58% annual increase, with the majority in R&D and S&M.
• Alteryx exhibited some significant variances in Research & Development and Sales & Marketing spend year/year. I assume that the $6M one time expense for Q1 falls into S&M and may explain the spike there. For R&D, this could be a result of the Feature Labs acqui-hire closed in Q4 2019 and the nearly 200 employees added in Q1. Given Alteryx’s large opportunity, I am happy to see more R&D spending. Twilio is taking a similar approach, with 21% of revenue allocated to R&D in Q1.

• Overall DBNER was 128% in Q1. This compares to 130% in Q4 and 134% in Q1 2019. However, on the earnings call, management mentioned that DBNER for Global 2000 customers was 140%. Since G2K customers would make up a large percentage of total revenue, a DBNER of this magnitude portends favorably for future revenue growth.
• RPO increased to $400.4M at the end of Q1, and was up 87% year/year.
• Leadership provided a new milestone metric in this earnings report of reaching $400M in ARR in Q1 of 2020. Alteryx doesn’t regularly report this metric, but did note that in Q2 of 2018, they crossed the $200 million milestone. Effectively they doubled the size of the business in 1.75 years. 
• Booked more than $107M in total contract value in Q1, up 53% year-over-year.
• CFO mentioned that Q1 experienced a “favorable product mix resulting in the upfront portion of revenue recognition being at the high-end of the range and an average contract duration of approximately two years, consistent with prior periods.” This also continued through April. Alteryx recognizes more revenue upfront from the sale of its Server product, as opposed to Designer. Also, Server contract lifetime tends to be 3 years on average, versus around 1 year for a typical starter Designer contract. The fact that product mix and contract duration in Q1 remained constant with prior periods indicates that the 35% price increase for Server introduced on February 1st hasn’t generated pushback from customers or impeded new sales at this point. We will need to monitor this going forward, but speaks to the value of this product.
• U.S. revenue was $80.5 million, an increase of 52% year-over-year while international revenue was $28.3 million, an increase of 22% year-over-year. Strong U.S. performance was driven by continued robust customer demand, specifically within the global strategic customer segment.
• Ended Q1 with 6,443 customers, an increase of 356 or 5.8% from Q4. This represents a 30% increase over Q1 2019. Of these, 12 customers are in the Global 2000, bringing Alteryx’s penetration of the G2K to 37%. The fact that customer growth is remaining on a straight line, coupled with their high DBNER, is another encouraging signal.

• Highlighted customer activity in several distressed categories in Q1. While we can assume most of this was before the COVID-19 lock-down, the breadth of companies is impressive. Leadership mentioned doing business with the following:
  • Travel and Hospitality: Australia Leisure and Hospitality, Caesars Entertainment, Choice Hotels, Copa Airlines, Graham Circle, Kurtz Corporation, Royal Caribbean Cruise Lines, Six Continents Hotels and Sun Country Airlines.
  • Energy: Chevron USA, Dominion Energy Services, Mid-American Energy Company, Pacific Gas and Electric and Thai Oil
  • Financial Services: BNP Paribas, Royal Bank of Canada and Standard Charter Bank.
  • Health Care: Allergan, Pfizer, Eli Lilly Hospital Corporation of America and United Health Group.
  • Telecomm: Saudi Telecom Company, SoftBank, Telstra and Vodafone. Have the top 10 telecomm companies in the U.S. as customers.
  • Closed a 7-figure transaction with a European financial institution in the last week of the quarter.

• Gave a small update on the strategic alliance program. This is expected to generated $1B in revenue over the next 5 years. PwC signed on as the first global strategic partner in early February. The CEO mentioned that they are starting to see a lot of inbound activity from this, particularly around digital transformation and that they are building sales pipeline.
• Named a Leader in Gartner’s 2020 Magic Quadrant for data science and machine learning platforms.
• Launched the Advancing Data and Analytics Potential Together (ADAPT) program recently. ADAPT is a free resource to help up-skill workers and formally train them in analytics. It provides free software licenses, access to Alteryx learning resources, software certifications and even a nano-degree in business analytics in collaboration with Udacity. The CEO mentioned at a subsequent analyst conference that 5,000 people have signed up for this in the last week.
• Saw a record number of unique registrations for their Community offering in the last 30 days, which allows customers to share best practices and get training. Customers who have registered for Community expand three times more than non-Community customers.
• Announced that Alteryx leadership will be attending 5 analyst conferences in next month. They are hosted by JP Morgan, Piper Sandler, Cowen, BAML and William Blair. Of these, Piper Sandler is new from 2019.

Analyst Reactions

After the earnings announcement, 4 analysts provided updates. They all lowered price targets, with 3 out of 4 maintaining a Buy equivalent rating. Given the marked deceleration projected for Q2, the analyst price target changes were reasonable. Interestingly, the analysts with Buy ratings only lowered their price targets by $10 on average. This may reflect an expectation for outperformance in Q2 and the remainder of the year. The average price target for these updates is a little over $130, representing a 9.2% gain from the closing price of nearly $119 on May 7th.

Date	Analyst	Rating	Price Target
5/7	Wedbush	Outperform	Lowered from $146 to $138
5/7	DA Davidson	Neutral	Lowered from $145 to $110
5/7	Needham	Buy	Lowered from $152 to $140
5/7	Cowen	Outperform	Lowered from $145 to $135

After the earnings results, Wedbush lowered their price target the least amount and maintained an outperform rating. Analyst Steve Koenig provided this commentary.

Wedbush analyst Steve Koenig lowered the firm’s price target on Alteryx to $138 from $146 and keeps an Outperform rating on the shares. The analyst argues that although the company beat revenue targets slightly in Q1, forward visibility is murky, with management issuing “disappointing” Q2 guidance and pulling full-year guidance.

TheFly, May 7, 2020

The largest price reduction came from DA Davidson. Analyst Rishi Jaluria maintained his neutral rating and provided this commentary.

DA Davidson analyst Rishi Jaluria lowered the firm’s price target on Alteryx to $110 from $145 and keeps a Neutral rating on the shares after its Q2 guidance came in “significantly below consensus”. The analyst notes that the company is likely being overly conservative, but also points to slower activity levels and higher churn reported in March.

TheFly, May 7, 2020

Alteryx APA

Near the end of the earnings call, Dean Stoecker (CEO) teased an upcoming product announcement representing a new category of software that Alteryx intends to own.

If you’ve looked at category creation in the past whether it was ACM or ITSM or SFA, the winner in that category tends to get 85% of the market cap of the entire category. And we intend that to be Alteryx. So we believe we’re particularly well positioned for, not just the short-term but the medium-term and the long-term.

Dean Stoecker, Alteryx CEO, Q1 2020 Earnings Call

This left investors wondering what he was referring to. On Monday, May 11, Alteryx announced its enhanced Analytic Process Automation platform (APA). The intent is to unify analytics, data science and business process automation into one platform. Alteryx contends that current analytics and data science work is addressed through a disparate set of point solutions. They want to provide a platform that allows data workers to prosecute the full spectrum of tasks in one system.

Beyond the technology consolidation, their approach is also very people centric. It is positioned to make analytics and automation available to a broad audience, essentially any employee with a need to better understand their company’s data. Alteryx primarily sells into the data analyst teams within business units versus centralized IT. Part of Alteryx’s intention is to expand the user set, by making data processing tasks easier through their “code free” approach. This is bolstered by the theme of empowering citizen data scientists.

The Alteryx APA platform provides an integrated solution that addresses all steps in the analytic process. It includes 260 automation building blocks that can be combined through a visual interface to create an analytic pipeline. These range from facilitating data import to cleansing, generating insights and exporting to external systems. APA addresses five phases of the analytic process:

• Automate data asset inputs. This allows data from many sources to be pulled into the pipeline for processing. The platform includes connectors for 80+ common data sources, like Salesforce, Oracle or MongoDB. These can be in the cloud, on-prem or even represent files on the user’s local machine like CSV, PDF, etc.
• Data quality and preparation. Provides tools to cleanse, prepare, and blend the data. Both structured and unstructured data can be handled. The goal is to normalize the data across the different sources, adjust for errors or missing fields and combine it all into a single data set.
• Data enrichment and insights. This layers in data from outside sources to complete the view. Can add geospatial context or pull in data from third parties, like demographics. With this, the user can manipulate their view of the data to generate business insights.
• Data Science and Decisions. To facilitate processing the data set to support predictive outcomes and repetitive querying, the systems allows the user to build complex analytical models. For data workers without coding skills in Python or R, assisted modeling tools and step-by-step guides are available. Data scientists can still load their models or tweak the code generated by the systems automated tools. Machine learning and AI are injected into the process as well.
• Automating Outcomes. This involves sharing the output and insights gained from the process. The sharing can be accomplished by writing back to a permanent data store, generating a spreadsheet or publishing a report. Once an automated process is built, it can be repeated to update the insight output on a scheduled basis. Also, APA supports lightweight dashboards for quick views, but is meant to export to more robust data visualization tools, like Tableau.

ML/AI capabilities will be layered into the data processing flows and leverages Alteryx’s October 2019 acquisition of Feature Labs, a data science company launched out of MIT. Feature Labs brought capabilities around automating the creation of machine learning models with a focus on feature engineering. These tasks are normally handled through manual, error-prone processes that rely heavily on skilled data scientists. That acquisition also brought Alteryx some street-cred, as Feature Labs maintains a number of open source libraries for data scientists to automate feature engineering tasks. At time of acquisition, these open source libraries had been downloaded over 350,000 times and Alteryx plans to continue support for this work.

The product offerings associated with APA appear to remain the same. They are still oriented around Designer, Server, Connect and Promote, along with optional add-ons for custom datasets. The pricing model hasn’t changed either. However, given the comments around product development pipeline and new job postings, I think we can expect this landscape to change and see new product offerings in the future.

The APA launch is being supported by a media blitz, including coverage in Forbes, Gartner, Fast Company and Wired. Additionally Alteryx is hosting a kick-off event that will be live streamed on May 20th. Participants beyond Alteryx personnel will be the Chief Product Officer at PwC (Alteryx’s global elite partner), the Research Director of Analytics at IDC and the VP Decision Science at Coca-Cola. The implication is that Coca-Cola is already a heavy user of Alteryx APA. Also mentioned in the release is a new customer, the Al-Futtaim Group, an operator of over 160 franchise brands across retail, financial services, automotive and health across the Middle East and Asia.

While most of this announcement appears to be around product positioning and marketing, Dean Stoecker (CEO) talked about a robust product roadmap this year at the JP Morgan Conference on May 12th. He said there would be more product innovation for Alteryx in the next 4 quarters than there has been in the last 10 years. He mentioned a few examples like a browser-based version of Designer and an Intelligent Suite. Another focus will be around ease of use towards applying advanced data science capabilities for users. These assertions are supported by the recent job postings mentioned above. In a recent interview of the CFO by TheStreet, the CFO also reinforced that M&A will continue to be part of their strategy to selectively pick up strong development teams and unique IP.

I think APA represents a bold move for Alteryx. As Dean fairly references, past software companies have built large ecosystems around a new category. Some software industry veterans claim that the company that defines a new category can expect 75% of the share. Examples include ITSM (ServiceNow), SFA/CRM (Salesforce) and HRM (Workday). Alteryx has wisely pulled IDC into the mix, as industry analysts (IDC, Gartner, Forrester) are the ones who come up with the different categories of software in order to structure the landscape for their reports (Magic Quadrant, Wave, etc.). By being first to a category (because you defined it), will generally mean that you will also be in the top right position in the competitive landscape. Launching with PwC will further reinforce the APA category, as we can imagine PowerPoints to enterprise c-suite members including new references to “APA”, giving the impression that they will be missing out if they don’t allocate budget to this area.

My Take-aways

• I think Q1 was tracking towards more outperformance before COVID-19 hit. Leadership mentioned that they saw a significant slow down in March, which is also the third month of the quarter. I think Q2 estimates are likely conservative as activity in April would be low and not provide much insight into quarter close. With that said, some peers are still putting up great results (FSLY, TWLO, DDOG), so I don’t want to make too many assumptions.
• My interpretation is that Alteryx’s traditional sales motion has been very tied to in-person interaction and the licensing model requires more overhead to get contracts completed. Reference the 7-figure European deal mentioned on the call that required many customer touchpoints to close. My sense is that the Alteryx sales team can adjust to this new motion. I don’t think it is a limitation in the product or value.
• The challenge for Alteryx in projecting Q2 revenue, as compared to other SaaS businesses is that only 60-65% of their sales are ratable, versus 100% for other companies. For an average SaaS company with a contract duration of 2 years, they would have visibility of 85-90% of revenue for the subsequent quarter. Going into Q2, Alteryx can count on only 60-65%. Of the remainder, 15% is from contract renewals and the rest is generated by new business. Because almost 40% of Q2 revenue is in flux and Alteryx typically closes most of its business in June for Q2, it is reasonable that they took a conservative view for Q2 estimates. They did mention the word conservative several times on the call.

And I think that the April performance is actually fairly strong. We’ve always been conservative even though we’ve seen some glimmers of hope in April. It turns out of the logos we brought-in in April 35% of them were actually in the highly impacted verticals, that’s also some silver lining in this. But as you know, we’ve been fairly conservative.

Dean Stoecker, CEO Alteryx, Q1 2020 Earnings call

• I can appreciate the quandary that customer CIO/CTOs were in during the height of COVID-19 disruption in March/April. With the world shutting down, spend would rightfully focus towards “keep the lights on” IT areas and necessary support for transitioning to work from home. What is obvious is that this has favored hosting infrastructure (DDOG, FSLY, NET), identity/security (OKTA, CRWD), communications (TWLO) and remote collaboration (ZM, DOCU). For an analytics package, I can see that spend being put on hold in context. Not because it isn’t valuable to the analytics team, but because in the heat of the moment, I wouldn’t categorize it as mission critical. However, investors shouldn’t conflate mission critical with valuable. As enterprises have more visibility into the future and we slowly reverse the shut-down posture, then value-add solutions (that generate efficiencies, cost savings or revenue opportunities) like Alteryx should see spend return.
• Putting aside COVID-19 impact, I was encouraged by other indicators of business performance that continued in Q1. Specifically, increasing customer count and G2K DBNER.
• APA is very interesting. It is mostly a marketing move for now, but presumably will be followed by platform re-architecture, feature improvement and new product offerings. Establishing a category around APA should provide the market fit for the product. The bevy of new job postings around modern cloud architecture, APIs, lightweight clients and open source provide confidence for me that Alteryx won’t fall into the trap of becoming a legacy software offering.
• Along these lines, Dean said something interesting towards the end of the call about renaming their annual user conference “Analyticon”. Alteryx is recognizing the convergence happening in the analytics and data science space. This aligns well with APA. They are going to open up their user conference to a much broader ecosystem, adding prospective customers and adjacent technology partners, such as H20 and Snowflake (which some analysts have categorized as competitors). They plan to bring all market participants into a single conference.

Risks and Items to Watch

• I think the primary risk for Alteryx going forward is around execution in Q2 and prolonged impact from COVID-19 on enterprise spend. Also, while Alteryx had great momentum through Q1, if buying behaviors somehow change or analytics investment is permanently de-prioritized, this would be a real problem for the investment thesis.
• The pivot to APA is exciting, but will represent a lot of work. The new category definition may not take hold. The modern software development motion may take longer to execute or engineering culture may not shift quickly. This is a big bet that might take several quarters to play out.
• The competitive landscape seems clear for now, but is always something to watch. Potential competitors might out-maneuver the APA initiative with categories of their own. A bigger player might roll something similar.

Investment Plan

It’s hard to encourage patience for investors, particularly when other COVID-19 beneficiaries (TWLO, FSLY, OKTA, DDOG, DOCU) are pushing to new highs. However, I feel that Alteryx has the potential to surprise investors later in the year and generate strong momentum going into 2021. This is based on positive execution in Q4 and through the first part of Q1. The new initiatives with APA and modernization of architecture should also broaden the customer landscape and reach within enterprises. These might drive new revenue streams layering on top of strong expansion rates for existing products.

My general investment strategy skews towards looking forward 6-12 months and trying to anticipate where a company will be at that point, relative to valuation, product positioning and dynamics in the addressable market. For me, this means holding my position in AYX. Unless the COVID-19 situation gets much worse, I expect Alteryx to return to higher levels of growth going into 2021. This would provide a steep valuation ramp this time next year. The introduction of APA, with associated industry interest from PwC and IDC, should provide further tailwinds. In the earnings call, the CFO reiterated their long term margin targets, which should support high valuation assuming reasonable revenue growth returns.

I understand if some investors will move over to other COVID-19 plays in the interim, while the thesis plays out. In the meantime, I am maintaining my long term price target for AYX of $325 by 2024 and will keep most of my existing position. I will shift a portion of my large AYX allocation in the near term to other stocks that recently outperformed (TWLO, FSLY) in order to balance out my positions. However, I will monitor AYX’s progress through this year and may add back as the situation becomes clearer going into 2021. In any case, I think AYX represents a solid choice for investors with a long-term horizon.