This article is a continuation of a series in which I attempt to apply a quantitative valuation model to high-growth technology companies. The first article* was about Slack (WORK) and contains an overview of my way of thinking quantitatively about companies that are growing quickly but are unprofitable today. As a brief refresher, my model relies on five key estimates: Total Addressable Market (TAM), the share of the total market that the company can reasonably hope to capture, the company’s net margin once it has reached scale, the PE multiple the market will be willing to assign to the company once it has reached its peak market share, and the annual rate at which the company will compound revenue growth until it reaches peak market share. The goal is to estimate what the company’s market cap will be by the time it stops growing and how long it will take the company to reach that point. The final result will be an estimated annual rate of return an investor can expect to receive if they invest in the company. In this article, I will be applying the model to MongoDB (MDB).
A Quick Overview of MongoDB
MDB is a market leader in noSQL database software and DaaS (database as a service) solutions. Historically, most enterprise databases have been relational databases, which require users to create structures and models for their data (think tables with rows and columns). This comes with some big advantages when you want to store, analyze, and update data that is related to other data in your database (hence the term “relational” database). A relational database allows one to use a standardized “Structured Query Language” (SQL) to read and write data. A relational database can be tuned to quickly retrieve related data from storage.
Relational databases are less ideal when working with unstructured data. A classic example of unstructured data is a social media post. Facebook and Twitter need to save all the posts their users make, but the content of a feed update or Tweet doesn’t fit nicely into a table structure. The whole message could be saved to a single column but, for reasons that are somewhat technical and nuanced, a relational database has difficulties retrieving and parsing that data in a timely fashion; it isn’t want relational databases are optimized for. Enter noSQL (“not only Structured Query Language”) databases, which are designed and optimized to store data more like documents without rigid structural requirements. Less like tables, noSQL storage more resembles an endless array of filing cabinets that can be broken down into folders and individual documents. With a noSQL solution, each of my social media posts can be stored in its entirety as individual documents and these documents can be searched, analyzed, and retrieved by the database quickly and efficiently.
MongoDB came to market with one of the first major enterprise-scale noSQL database solutions and is the market leader in this area. Customers can purchase software licenses and run the databases on their own hardware, or they can sign up for one of MDB’s DaaS solutions where MDB allows customers to create noSQL databases hosted on cloud solutions like Amazon’s AWS and Microsoft’s Azure. The DaaS model means users can create high-performing noSQL databases without owning hardware in a data center or hiring database administrators to oversee that hardware.
I go into this much depth because it is an important part of understanding who MDB competes against. I hope I have made it clear that a noSQL database is not categorically better or more efficient than a relational database; the correct solution depends on the type of data that needs to be stored. The demand for noSQL storage is growing rapidly (just think about how many Tweets are being created each second), but noSQL solutions are unlikely to strictly replace relational databases because relational databases are a better solution for certain projects. Oracle (ORCL) and Microsoft (MSFT) are the largest players in the enterprise relational database space, along with a host of open-source relational offerings (like PostgreSQL and MySQL). I would say relational databases companies are indirect competitors to MDB, rather than direct competitors. Direct competitors are other noSQL solution providers, which include Amazon’s DynamoDB, Couchbase, Apache’s Cassandra, and open source solution Redis.
Looking at MDB’s financials for fiscal year 2020, the company brought in $421 million in revenue and earned gross margins of 70%, compared to $261 million in revenue at 72% gross margins in FY 2019. MDB has not generated net income or operating cash flow as a public company, posting a $175 million net loss and $30 million operating cash burn in 2020. MDB has about $50 million in net cash, with nearly $1 billion in cash and marketable securities against $950 million in debt.
Some Quick Notes on Total Addressable Market
Estimating MDB’s total addressable market requires its own set of assumptions that I want to discuss briefly. Per MDB’s initial S-1 filing, the company cites IDC’s estimate that the world “structured data management software” market will be about $61 billion in 2020. IDC projects that this market will grow about 8% annually. As discussed above, I don’t think that entire market is truly accessible to MDB, given the differences between noSQL and relational databases. That being said, in my valuation scenarios below you will see that I use the $61 billion amount as my TAM value and assign slightly higher terminal PE values than usual to account for the relatively high industry growth rate. I admit this is a bit inconsistent, but I attempt to make up for it by adjusting market share estimates and assuming that the missing 8% a year of TAM growth accounts for the relational portion of the market that is inaccessible to MDB. As I show in my post-valuation section, the TAM estimates end up being less important to the MDB valuations than they were to my WORK valuations, so I am comfortable proceeding with the $61 billion number as a baseline estimate.
I begin with an optimistic scenario. In this situation, MDB ends up becoming the Oracle of noSQL solutions, with a similar market position and financials. This means MDB controls about 50% of the $61 billion total addressable market, resulting in annual revenue of about $30 billion. In this scenario MDB has an Oracle-level net margin of 25%, is assigned a slightly more generous PE of 25, and I credit the company with being able to grow revenue at an average rate of 40% per year. Putting these estimates together results in the following outcome:
|Current Market Cap (M)||$10,600.00||Year 1||$588.00|
|Current Revenue (M)||$420.00||Year 2||$823.20|
|TAM (M)||$61,000.00||Year 3||$1,152.48|
|Market Share||0.5||Year 4||$1,613.47|
|Terminal Revenue (M)||$30,500.00||Year 5||$2,258.86|
|Terminal Net Margin||0.25||Year 6||$3,162.41|
|Terminal PE||25||Year 7||$4,427.37|
|CAGR revenue||1.4||Year 8||$6,198.31|
|Final Market Cap (M)||$190,625.00||Year 9||$8,677.64|
|Years to Reach Terminal Revenue||13||Year 10||$12,148.70|
|Implied Annual Return (%)||25%||Year 11||$17,008.17|
(Source: Author Spreadsheet)
For a conservative scenario, I assume that MDB only captures 33% of the total market, their terminal net margin is only 20%, they are assigned a more modest PE ratio of 20, and most importantly their average revenue growth rate drops to 20%. In this scenario, the numbers look like this:
|Current Market Cap (M)||$10,600.00||Year 1||$504.00|
|Current Revenue (M)||$420.00||Year 2||$604.80|
|TAM (M)||$61,000.00||Year 3||$725.76|
|Market Share||0.33||Year 4||$870.91|
|Terminal Revenue (M)||$20,130.00||Year 5||$1,045.09|
|Terminal Net Margin||0.2||Year 6||$1,254.11|
|Terminal PE||20||Year 7||$1,504.94|
|CAGR revenue||1.2||Year 8||$1,805.92|
|Final Market Cap (M)||$80,520.00||Year 9||$2,167.11|
|Years to Reach Terminal Revenue||21||Year 10||$2,600.53|
|Implied Annual Return (%)||10%||Year 11||$3,120.64|
(Source: Author Spreadsheet)
Finally, my moderate scenario is similar to the optimistic scenario but with slightly reduced market share, terminal PE, and revenue growth estimates:
|Current Market Cap (M)||$10,600.00||Year 1||$567.00|
|Current Revenue (M)||$420.00||Year 2||$765.45|
|TAM (M)||$61,000.00||Year 3||$1,033.36|
|Market Share||0.4||Year 4||$1,395.03|
|Terminal Revenue (M)||$24,400.00||Year 5||$1,883.29|
|Terminal Net Margin||0.25||Year 6||$2,542.45|
|Terminal PE||20||Year 7||$3,432.30|
|CAGR revenue||1.35||Year 8||$4,633.61|
|Final Market Cap (M)||$122,000.00||Year 9||$6,255.37|
|Years to Reach Terminal Revenue||14||Year 10||$8,444.75|
|Implied Annual Return (%)||19%||Year 11||$11,400.42|
(Source: Author Spreadsheet)
MongoDB’s Primary Driver is Revenue Growth Rate
Using the ranges of estimates above to create valuation scenarios, it becomes clear that most impactful estimate is revenue growth rate. This surprised me because when I did my WORK valuations, total addressable market and market share were the dominant estimates. The difference is that MDB’s current revenue is proportionally much lower compared to its addressable market size. Even moderate assumptions about market share give MDB room to reach $25 billion to $30 billion in revenue, against FY 2020 revenue of only $422 million.
On the positive side, this means that MDB has a very long runway of growth ahead if things go well. On the negative side, MDB needs to be able keep up high levels of growth for well over a decade to return significant value to shareholders. This introduces additional risk to an MDB investment, as the odds of additional market entrants increases each year, as does the risk that existing competitors will take market share via new or improved products.
If one were to assume a much larger addressable market and/or higher market share, it matters less than it seems because if that market share target is paired with a growth rate below, say, 25%, MDB won’t be able to reach the revenue projections for 20+ years, adding significant risk to the investment thesis.
My model doesn’t have a clear mechanism for discounting the risk of future growth being disrupted; I will work on a better way of quantitatively representing that risk in future analysis.
MDB offers a valuable and differentiated product and has a long runway for growth. While the moderate and optimistic valuations look appealing on an absolute basis, in both scenarios MDB needs to maintain a high growth rate for over a decade. Given the number of existing competitors and the potential for more to enter the market over such a long time frame, I would approach an MDB investment with caution. I would particularly advise investors to focus on revenue growth rates in the future, as this metric will have the most impact on long-term valuation models. If MDB keeps up the growth, I expect the company’s shares to compound handsomely over time; if the company misses growth targets over a whole year or longer, I expect a sharp downward re-rating of the company’s shares.
*I have made my WORK article an “Author’s Pick”; it can be viewed in its entirety for free if accessed directly from my SA author profile page.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: This article should not be taken as financial advice, it is only an expression of my own opinions as an individual investor