c15d441a-5b32-4147-8bb9-d78d991e3263_UiPath+Robotics+Process+Automation+Premium+Analysis.pdf
I’m genuinely curious as to whether UiPath will be the first in the tech universe to hold its opening IPO price. There’s a chance it does and I will attempt to communicate why this company could be one of the strongest performers the public markets have seen from my industry.
UiPath is becoming a darling within tech circles. That’s maybe the most critical thing to understand as the Partner Network for UiPath could potentially form a moat. The high switching costs most certainly help the company to become defensible. UiPath makes the choice very clear on what company to choose for RPA, no matter what your organization’s needs are, and from there they give enterprises no reason to switch by continually iterating a cutting-edge approach to RPA.
The last time I was this excited about an IPO was two years ago with Zoom Video. We held off to buy the stock post-lockup. That decision will be yours to make. What distinguishes UiPath for me is that AI stocks have a way of sneaking upward in price. We saw this with Nvidia during the tech rout as Nvidia (the AI bellwether) has performed well this year compared to other leading tech stocks with strong earnings.
Pictured above: Nvidia, the AI bellwether, has outperformed other leading tech companies this year. Little does the market know, that AMD is also becoming a force in AI with the Xilinx acquisition.
What is RPA – Macro Overview
Robotics process automation has many supporting macro statistics because its essentially machines replacing humans. The ROI is astounding when you have an error-free employee who works 24/7 and does not tire or need bathroom breaks. To illustrate, a few automations can save 20 minutes of work per person daily and enabling 10K employees with a software robot could save more than $30 million a year (based on an average salary of $35/hour).
There are many fears that RPA will eliminate jobs to the detriment of society. Proponents say this isn’t exactly true, rather RPA will eliminate menial and low satisfactory tasks. According to analysts like Forrester, 14.9 million jobs will be created by 2027 to work alongside robots. It’s not clear though how many jobs robots will replace and if the 15 million is actually a deficit.
According to McKinsey, $3.6 trillion of work can be automated. The piece of the pie that UiPath is after is the automation of applications for enterprises. The number of applications deployed by enterprises has increased by “approximately 70% over the past four years,” according to Wall Street Journal.
The 10,000-foot view of what RPA solves is that interoperability of applications is cumbersome with “a compounding effect on the complexity of business processes” and work done by IT departments.
According to McAfee, the average enterprise has deployed 464 custom applications and deploys an additional 37 new applications in a 12-month time span. Companies with fewer than 1,000 employees run 22 custom applications while companies with over 500,000 run 788 custom applications, on average. The majority of these applications (58%) are used internally while 36.2% are used by customers, partners and suppliers. These larger enterprises – with the 788 applications on average — are the companies that UiPath is targeting.
Source: McAfee, 2017
In the United States, real output per hour grew 31% during the decade ending December 2009 while it grew 13% in the subsequent decade through 2019. UiPath believes this decline in output is due to the overwhelming number of applications and software within enterprises.
This is partly because applications are specialized and are not able to address the end-to-end processes that enterprises require. The concept that UiPath proposes is to automate those steps and have a human review the exceptions rather than every detail of every order.
Product Overview:
It’s important to start with product for a company like UiPath – and most companies in tech, really – because without knowing what the special sauce is, companies can get lost in the noise. UiPath is a platform that allows companies to run software bots that process automations. What separates UiPath is the use of AI computer vision to read information, hence having the acronym “UI” in the name, which stands for user interface. The company also leverages machine learning to think and process the information and robotics process automation (RPA) to interact with applications.
The combination of computer vision and machine learning is UiPath’s special sauce. The AI-based computer vision increases the reliability of automation. The AI-based computer vision is able to adapt and interpret varied document types and user interfaces. This is the missing piece in automation that other forms of orchestration or choreography do not have.
The key sentence in the S-1 filing is this: “Our platform enables the reusability and reliability of UI elements by capturing them as objects in a repository.” This means that the AI computer vision is able to dynamically recognize and interact with variables and dynamic objects or applications. In plain terms, it means UiPath can emulate a human by responding to variables the way that humans can. (Anyone working on autonomous vehicles will tell you, the issue with full autonomy is the variables, not the mechanics of driving).
UiPath’s architecture is UI-based orchestration. This increases the reliability of the automation as it’s able to adapt and interpret varied document types and UIs. As stated, the platform captures UI elements as objects in a repository.
To compare, here are other ways automation and/or integrations are handled:
· Integration orchestration: When on-prem wants to integrate with SaaS platforms, companies like MuleSoft and Webmethods offer third-party connectors. This is called Integration Platform as a Service (iPaaS)
· Business process orchestration: Offers business processes a central process, yet requires human intervention.
· API based orchestration: Lacks a central component and is event driven
· Event driven architecture: Event driven to where the events are autonomous through choreography rather than orchestration (the difference being that orchestration requires a composer while choreography establishes a pattern that does not require supervision).
UiPath recently acquired Cloud Elements to add API-based automation to its core offering of UI-based automation. This is the first time the combination will be offered in a single platform. This places UiPath on the same playing field as the orchestration methods listed above, yet with the combination of computer vision/UI-based orchestration. This acquisition takes aim at that market share by providing the best of both worlds. The acquisition brings 200 new native integrations to UiPath
Why is UiPath Better?
The next question to answer is why is computer vision/UI method better? The first is that UI-based automation is not confined to specific APIs. The result of using computer vision (and the other components that I review below – but let’s keep the focus on computer vision for now) is that UiPath is an end-to-end solution rather than a point solution. The goal is to automate the process, not the API, and other orchestrations lack the ability to automate across many applications and link AI capabilities to execute. Without computer vision, the end result will not be human emulation.
The company points out in the S-1 filing that the typical AI/ML environments are developed by data scientists yet need to be used by other departments that carry the processes out (billing or customer service, for example). My takeaway on this is that the other methods for integration and automation do not necessarily cut down on the number of people required and/or does not reduce the technical abilities required to work with the automations. By requiring data scientists to be the central and only hub, end-to-end automation is not possible.
The modular setup is also an advantage. Solutions can be integrated into new, third-party technologies for future development.
When we talk about robots, we are talking about software robots that are on a desktop computer, can work across programs in the background, are able to build applications, send emails, and interact with chatbots. This is achieved with this build:
· AI computer vision can dynamically recognize and interact with variables and dynamic objects and applications
· AI-enabled platform helps identify which processes should be automated including interoperability with 75 AI technology partners
· Document Understanding leverages optical character recognition (OCR) and natural language processing (NLP) and ML to handle processes with humans handling only the exceptions
· Low-code Development drag-and-drop tools to serve a range of technical skills
· Governance and Security ensures compliance
Product Specifics:
UiPath is an expensive product and this is reflected in its customer concentration at the enterprise level. There is a Community Edition that is free, which is a smart way to onboard more developers at the student level.
Studio is UiPath’s integrated development environment (IDE) that allows access to the Automation Cloud. There are three variations: Studio, Studio Pro and StudioX. The difference is what technical level the user has with StudioX requiring very little skills (i.e. “low code”) with drag-and-drop while StudioPro requires Advanced skills.
Automation Hub:
Automation hub allows for central management of the automation pipeline. It’s a command center to see and control the end-to-end system. It also allows the administrator to visualize automation complexity and understand the impact and ROI.
Process Mining:
Process mining taps into a data source from enterprise applications and makes use of this event data. The goal is to streamline processes to become more efficient. For instance, if your goal is to improve customer retention rate, then you can track how long customer service responses take, delivery rates, and what is causing delivery problems so you can address the situation. Process mining also helps you identify bottlenecks that can benefit from automation.
Process mining changes all this by tapping into a data source that already exists. This is done through the ETL “extract, transform, load” process. Most of your enterprise applications (like SAP and Salesforce) record every activity and transaction that happens within each stage of a process. This is called event data.
Business processes suitable for process mining include accounts receivable, claims and accounts payable. In financial services, it can be used for loan approval, risk and investment management or fraud. In health care, process mining can be used to reduce paperwork and streamline processes like the spike in demand for testing we saw during Covid.
Task Mining and Task Capture:
Task capture allows for the mapping of business workflows. Employees can record the process they want to automate and Task Capture will gather data for each step. The software generates a process map into a file for the development team to use to create automations.
Task mining will have its public launch in 2021 and will allow enterprises to record work performed by users across a list of applications.
Business Model and Automation Flywheel:
UiPath benefits from a flywheel effect. The reason that a flywheel effect occurs is because when companies use UiPath to add robots, they see a substantial return on investment, and then deploy more robots.
The company is built and ready to scale with flywheel effects as UiPath can be customized for every need of the enterprise. The robots are designed to work in any environment (cloud, hybrid, on-premise), for any level of technical ability (low code to advanced code), is licensed through subscriptions annually or multi-year, and can work alongside a human or be fully automated, is additive, and can be used as a unified solution or individually (that’s a mouthful).
The point is that UiPath is prepared to offer a solution for any customer need and to scale as the needs of the enterprise changes.
If a picture is worth a thousand words, then perhaps this helps illustrate the flywheel effect:
The graph above shows that the 2016 cohort of customers have increased their ARR from $395,368 to $22.7 million in a five-year time span. This is a multiple of 57X. The company’s top 50 customers have grown ARR by 81X. This is measured by the ARR generated in each customer’s first month as a customer.
There are some examples in the S-1 filing that show up to 69X increase in customer ARR within one year. That’s an outlier with others increasing 32X, 6X and 4X in one year. Even the lowest number here is impressive, and is driven by cross-department sales, increase of robots per employee, increased adoption across products, and expanded use cases.
Partner Program and Developer Ecosystem:
Partner programs are especially important for a company like UiPath as it can help to extend business models and also help to scale a product very quickly. It’s also important for global growth across various regions. UiPath is the automation back bone to many other products that you’re familiar with: Microsoft, Google, Amazon Web Services are integrated with UiPath so their cloud computing customers can utilize cloud-based AI capabilities. Other integrations include Adobe, Alteryx, Oracle, Salesforce, SAP, ServiceNow and Workday.
UiPath is unique in that it’s tailored to citizen developers who prefer low code and also advanced developers. Right now, the company counts 750,000 registered developer accounts. There is a marketplace with 1,200 vetted and pre-built automation activities that can be used for enterprise workflows. UiPath says there are 10,000 downloads per month in the Marketplace.
Market Size and Valuation:
I expect that UiPath will trade at a high valuation into the foreseeable future. This is because of the specific trend the company is capturing. I think individual investors can get lost in the noise of popular stocks, but I don’t think institutions fall prey to this as easily. You’ll notice the companies that are favorites among retailers are not favorites among institutions. That’s one reason we run this site, is to bring to your attention to companies that you won’t want to miss out on.
What about sector rotations, like we just had? Even still, companies like Snowflake retained their position of highest valuation, comparatively speaking. It went from a jaw-dropping 80 forward P/S in February to about 50 forward P/S – yet this company still led the pack of cloud software valuations.
We typically don’t buy above 40 forward P/S (you can reference when I discussed this during heightened exuberance in this Motley Fool video). We don’t know UiPath’s forward valuation until we see more analyst consensus numbers come out. The earnings report on June 8th will help quite a bit. However, if we adjust the revenue growth to a reasonable 65% forward, then we see UiPath trading at 41X.
Please note, this is calculated on 65% revenue growth next year, which is an educated guess. We will know more about UiPath’s forward estimates in the coming weeks.
If we take the LTM valuations, we see UiPath stretching the upper limits again. LTM is important here since UiPath’s forward is not available yet.
The I/O Fund does not give financial advice and highly valued tech companies require an appetite for risk. We simply tell you what we are doing with our own money. We bought UiPath starting last week after the blog notification went out and we will be watching it closely near the lock-up expiration as to whether we need to exit and enter again. This extra work is worth the long-term trajectory that robotics offers us as tech investors.
The reason UiPath is unique from the others on this list is because the opportunity for RPA is fully in front of the company, whereas many of the others are centered in a trend that is in motion.
I also think UiPath will continue to take more market share relative to the RPA market, and the key metrics around increased ARR help prove this as they are a bit mind-blowing. Other than ARR, I can’t quantify exactly why I think UiPath will take more market share other than it has a solid reputation and there is a buzz around the company that is hard to communicate. People are pleased with UiPath, they’re upgrading and buying more, and it’s becoming the company that everyone (who is anyone) needs to partner with.
Market Size:
There are a range of estimates on the size of the Robotics Process Automation market.
According to the S-1 filing, IDC places a $17 billion value for 2020 to reach $30 billion by 2024. Meanwhile, UiPath states in the S-1 that the “fully automated” enterprise is a market of $60 billion. There is also a reference in the S-1 to an estimate from Bain & Company placing the market for automation software at $65 billion.
Forrester states there are 1.69 billion knowledge workers globally. So, that’s helpful to picture TAM.
Third-party analysts are more conservative – no surprise there as S-1 filings usually publish the highest numbers available. According to Statista, the market will be worth $10 billion by 2023.
Global Market Insights places the market at $23 billion by 2026. Grand View Research states the market will reach $13.74 billion by 2028 at a CAGR of 32.8%.
That’s quite the range and is tough to extract a real market size from these numbers. However, there is agreement across the third-party sources that the robotics process automation market (specifically, RPA only) was $2 billion in 2020. This means UiPath owned about 30% of this specific market at $600 million in revenue. If we take the $10 billion Statista is putting out there, then UiPath can see a path to $2 to $3 billion in revenue by 2023 in RPA specifically. I always give room in these market growth estimates, by the way, so 2023 should be considered 2024 by Statista. It’s hard to nail down market growth with adoption in tech products.
Financials:
Please note, UiPath’s fiscal year ends January 31st.
We covered the financials here in a short write-up.
The good news about UiPath is that as the top line improves, the bottom line is also improving. Revenue growth from $336 million to $608 million, represents 81% growth for fiscal 2021. As mentioned in the valuation section, we won’t have a complete picture on forward growth until the upcoming earnings report on June 8th. Right now, I’m assuming we will see growth this year in the 60-percentile range. This is a guess so I will update you when we get real numbers from the earnings and analyst consensus reports.
In the past, operating margins have been an issue for the company with an operating margin that was triple digits in the red (154%) in 2019. The company’s current operating margin is (18%). Adjusted operating margins were (113%) and (4%), respectively.
The net losses have also improved from ($520) million in 2020 to ($92) million in the current year. The fiscal year 2020 losses were at (155%) of revenue compared to (15%) of revenue in the current fiscal year.
The bottom-line losses were partly driven by sales and marketing costs which were at 144% of revenue in fiscal year 2020. Surprisingly, R&D is low at 39% of revenue in 2020 and 18% of revenue in 2021.
The current gross margins of UiPath are at 89% which is among the highest in the software industry. This has been consistent with 82% margins in fiscal 2020. Free cash flow is at 4% of revenue, or $30 million in fiscal 2021.
The company is an enterprise-focused company and is well utilized. As of last year, the company counted 80% of the Fortune 10 and 61% of the Fortune Global 500 as customers. This grew to 63% of the Fortune Global 500.
As stated, the main business model is to increase spend per customer as enterprises will deploy more robots across more departments, increase number of products used, and expand use cases for automation. That’s important to repeat because typically this high of penetration could actually be a headwind for growth.
The dollar-based net retention rate for the company was 153% in 2020 and 145% in 2021. After subtracting churn, the gross retention rate is 96% and 97%. With that said, this metric is becoming less meaningful the more cloud and subscription-based companies come on the market. We have many in the 130 to 160 range and it’s similar to checking vitals – we want to know the company is healthy but it doesn’t tell us much about the nuances of longevity.
The company has a NPS rating of 71, helping to illustrate a high level of customer satisfaction. Despite serving large enterprises, including 8 of the Fortune 10, UiPath does not have any customer making up more than 10% of revenue.
Conclusion:
We finally have the AI pureplay we’ve been waiting for in the software category. This company is likely to be valued high into the foreseeable future due to the attractiveness of the trend (robotics). We feel institutions will want to participate in this trend with this risk/reward ratio that UiPath offers, which to reiterate, is lower risk on execution due to the excellent end-to-end platform, strong partner program, leading market share and ability to scale quickly across enterprises, with the differentiation of computer vision.
Please look for Knox’s trade notifications as he patiently builds this to become a core position of ours. If you made me choose, (and we do have to choose as we disclose allocations to you), I would place UiPath above Snowflake on conviction. Post-IPO lockup, and pending Knox’s excellent skills in finding the right entry, you can expect to see UiPath in our top 10 by year-end or shortly thereafter. Note that we aren’t rushing into a position at this valuation, rather keeping our toe in the water, and navigating as we see what the market does.
