Thoughts of SaaS services

I have been working at a US company for almost a month recently. An interesting fact is that I have counted the SaaS services I have used for specific scenarios since joining, and surprisingly, there are as many as 23 of them. All of these SaaS services are provided by third-party companies. In other words, Databricks is a customer of all these companies. This aroused my curiosity and made me want to learn about the situation of these companies.

Okta

As a new employee get my computer, the first thing I need to do is to set up an Okta account. Okta is a third-party SaaS company that provides identity authentication and permission control. Since US companies usually purchase a large number of third-party company software services, these software services must have their own account system. Okta is designed to solve this problem, that is, any employee can access the software services of all other third-party companies through a unified Okta account. At the same time, access permissions can be set for Okta accounts to fine-grained manage different employees’ access to different company resources.

Yes, Okta is such a “simple” service. Generally speaking, when we talk about scenarios such as single sign-on (SSO), multi-factor authentication, and permission control, these are the core scenarios of Okta. Okta is a third-party company, which is not surprising at all. What surprised me was that Okta is a Listed Company with a market value of $20 billion (financial report [1]), with a total revenue of $2.6 billion in 2024, an increase of 16%, and a signed order amount of $4.20 billion. What kind of god company is this? Countless question marks are buzzing in my mind.

What is Okta’s competitive advantage? Compared to Microsoft Entra ID and Google Cloud identity, Okta is a vendor-neutral diversified identity authentication solution that connects to 7000 + Saas applications on the market and covers almost all Saas service authentication scenarios on the market. In order to provide a leading development platform for developers in the fields of authentication and security, Okta spent $6.5 billion to acquire Auth0 company (see announcement [2] Okta thinks it is worth it because it believes that authentication is a $55 billion market). Biometric identification , hardware authentication , dynamic authentication , and AI risk control constitute high-level technical barriers. Application ecology , developer ecology , and technical advantages can be said to be the core factors that make Okta stand out among many competitors.

Why is Okta’s revenue so high? After winning the first place in competitiveness, customers of all sizes have to choose one in the identify field. Choosing Okta, the top 1, is probably correct. Okta adopts a subscription charging model, with each user paying monthly. For example, our company has nearly 10,000 people, guessing that each person pays 25 dollars per month (can only rely on guessing, because Okta’s official website does not provide the price of enterprise [3]), which contributes 3 million dollars in revenue to Okta in a year. Okta claims on its official website that 19,300 + organizations subscribe to their services, with hundreds of millions of MAU users [4]. These data generally support their revenue figures.

Workday

After I joined the company, the HR would send me a message to fill in some information on Workday to ensure that I could receive my salary normally. In the process of using Workday, I felt that this system was a system with strong customization ability and simple style, and there seemed to be nothing more special. Workday’s San Francisco work building is just next to the Databricks work building. I searched for information about Workday and found that it is a Listed Company with a market value of $69 billion. Its revenue in 2024 reached $8.40 billion, with a growth rate of 16%. Workday’s core business is Enterprise Human Resource Planning (HCM) and financial management software , mainly serving scenarios such as compensation, attendance, benefits, and performance. Workday is a leader in this segment, with Oracle HCM and SAP Successfactors as its core competitors.

Glean

When I was still in the former company, the team had researched the company’s products. At that time, Glean raised $260 million at a valuation of $4.60 billion [5], and announced that its annual revenue (ARR) exceeded $100 million [6] in Q4 2024. It is very rare for growping ARR to exceed $100 million in this field, only in three years. We also know that Glean is a leader in the AI RAG field, but the public information does not explain any technical principles, only some product descriptions. At the same time, we did not experience the product environment. In addition, the team unanimously felt that Feishu’s information automatic closed loop and search were still useful, but we did not understand the value of Glean. In the end, we were left with questions, but we could not get more inspiration.

After I joined the US company, I found that the information here was completely scattered in dozens of systems, such as Google Doc, Google Slides, Google Drive, JIRA, Slack, company blog, workday, Github, email, and so on. It is very difficult to achieve efficient global retrieval within the enterprise. Glean uses AI to solve this problem, and Databricks is a customer of Glean (Glean’s official website has announced [7]). Interestingly, Glean also announced the use of Databricks’ AI/BI Genie to analyze structured data and use UnityCatalog to implement fine grain permission management for search results [8]. Databricks leads in intelligent BI, while Glean leads in data source integration, embedding models, and ranking models. They purchase from each other and form a strong alliance.

The way to use Glean is to add a Glean browser extension on the Chrome, so that the browser opens a new tab by default, which is similar to google search. You can search for any information in the search box, and the whole experience is almost the same as Google. Only Google is external, and Glean is internal.

So what are the thresholds and challenges in this direction?

First, a company’s data sources are complex, possibly hundreds of them, and it takes a considerable amount of resources to complete the docking.

Secondly, efficient retrieval and ranking of Structured Data, Semi-Structured Data, and Non-Structured Data is also a threshold.

Thirdly, enterprise customers have high requirements for fine grain data permissions and security. For example, employees at different levels have different permissions for documents, and search results cannot disclose permissions. Deleting a Google doc requires triggering synchronous deletion by the enterprise search engine, otherwise there will be data leakage.

Glean solves these problems in a simple way, forming competitiveness in the three dimensions of “accuracy”, “safety”, and “simplicity”, making it a popular choice for enterprise search.

Freshservice

As a new employee, there are many questions that need to be consulted with the IT team, HR team, and administrative team. Databricks purchased Freshservice service (refer to Freshservice’s official announcement [10]), which can be simply regarded as a ticket system. At ByteDance, people usually find people directly on Feishu or directly call oncall, and the other team will respond and solve problems in real time. However, US companies respond asynchronously. First, a ticket is requested through Freshservice, and then wait for a day. After the other team replies the next day, the ticket is closed by oneself, making it easier for everyone to focus on doing things. Freshservice is such a product. To my surprise, the company behind FreshService, FreshWorks, is a $4.30 billion SAAS company with a total revenue of $720 million in 2024 and an annual growth rate of 21% [9]. I don’t know much more about it.

Other

There are also some niche Saas scenarios, to name a few.

On the first day of work, I needed to apply for a work badge and take a photo on it. I received an email from an external website with a mobile phone access link, which asked me to take a photo with my phone. Finally, after taking the photo, it was automatically uploaded. The website will use AI to beautify it and automatically submit it to the company’s work badge process. Yes, this website solves the problem of a 1-minute scene, taking photos with a mobile phone and AI to beautify the picture. The company behind this will have a revenue of about 1.80 million US dollars in 2024. The small and beautiful team of about 10 people has been working for many years.

On the first day, I didn’t have lunch at mealtime because I followed everyone to the table and found that everyone was looking for takeout on the cabinet. I originally wanted to grab some takeout to eat, but found that everyone’s name was written on the takeout. So I found some snacks for lunch. Later, a colleague told me that I needed to register and order on a website, and the company account would automatically reimburse the meal expenses. This is the business of this website. Its revenue was about 4.50 million US dollars last year, and it was established in 2013. There are now a total of 40 people. The product page is clean and clear, with no other functions except for the meal selection button.

After a few days, I chose medical insurance. For beginners, choosing US medical insurance is very complicated, no less than reading a bunch of insurance papers. For example, medical suppliers are divided into several types: HMO, PPO, and EPO, each with different medical restrictions and cost structures. The terminology is even more cumbersome, such as Deductible, Copay, Coinsurance, Out-of-pocket max, In-network/out-of-network, etc. Each concept is daunting. Supplier type, work location, family population structure, age, medical needs, out-of-pocket budget, etc. Each one is a variable, and when multiple variables are combined, people are confused. I used to enjoy medical insurance in China, but I never suffered from this kind of suffering. But if there are pain points in the US, there is a solution. The company provides a third-party company’s SaaS website. I answered a survey on the website and it automatically recommended a medical insurance plan to me. Later, I confirmed with a colleague and it was quite reliable.

Summary

In fact, in addition to the above, there are many Saas services for segmented scenarios, such as setting up 401k, Reimbursement System, VPN system, commuting expenses, dentistry, documents, meetings, and other scenarios. Each segmented scenario is an independent Saas product, which is why it was mentioned at the beginning that 23 Saas services were experienced. Behind these products, some are giants worth billions or even billions of dollars, and some are small and beautiful teams of 10 people who have worked for many years. Just as various parts can be assembled into a product in China’s strong manufacturing ecosystem, various Saas can also be found in the rich software ecosystem abroad to quickly build a company. When I personally experienced such a comprehensive, interdependent, and profitable software ecosystem, it really surprised and puzzled me. How did this ecosystem iterate and evolve into what it is today, and how will it evolve tomorrow? At least with my current knowledge reserve, I cannot combine macro and micro to answer this question.

But I can try to answer briefly. Let’s try to look at the problem from the perspectives of each character, which is very interesting.

Customers

When the cost of subscription is significantly lower than the R & D cost, customers will definitely be willing to pay. Taking Okta as an example, when customers need to establish a team of 15 people (with an average annual salary of $300K) for 10,000 employees to develop and maintain similar products, and the products are not necessarily better than Okta, then investing 4.50 million US dollars in R & D every year is obviously not as good as spending 3 million US dollars directly to purchase services.

SaaS Corporation

If the pain point scenario is found and customers are willing to pay, then the company’s revenue is good. Therefore, Saas company continues to dig deep into vertical competitiveness. One is to spend a high price to find top talents in segmented fields, and the other is to spend money to acquire competitive startups. After obtaining stronger technology and market competitiveness, Saas company gains more customers, revenue continues to grow, and enters a positive cycle.

Venture capital firms

VC see the revenue growth and profit margins of Saas companies and believe that the ROI of the Saas industry is high, with stronger willingness for initial and secondary investment.

Potential entrepreneurs

After discovering the pain points of the scenario, if the comprehensive judgment is that there are good financing expectations, revenue expectations, listing expectations, and acquisition expectations for startups, then potential entrepreneurs will believe that the entrepreneurial risks are shared across multiple dimensions, and have greater motivation to do startups.

The above can explain why it enters a virtuous cycle, but it does not explain why a diverse ecosystem is formed by the close combination of various vertical products. I think it is also because of the cost. For Saas companies to maintain the same leading competitiveness in multiple fields, it requires several times the additional financial resources, and most customers will only choose the top 1 product in a specific field. Instead of achieving 60 points in multiple directions, it is better to achieve 150 points in one direction, so as to eat more market share. After Saas companies lead in the vertical field, they open up this ability through API, allowing other companies to easily enjoy competitiveness by paying, creating better product stickiness. Once a company’s position in the entire ecosystem is stable, its market position is stable. As long as the market size is growing, the company will continue to grow. “Learning ecology”, “Integrating ecology”, “Optimizing ecology”, “Defining ecology” , “Harvesting ecology” Perhaps this is the growth process of many Saas companies in the US!

Okta#

Workday#

Glean#

Freshservice#

Other#

Summary#

Customers#

SaaS Corporation#

Venture capital firms#

Potential entrepreneurs#

References#