On January 14, MongoDB (MDB) presented at the Needham Growth Conference. Michael Gordon, the CFO and COO, represented MongoDB. I always enjoy listening to Michael’s perspective because of his dual role – he can speak to both the financial position, as well as the business operations and long term strategy of the company. He has a strong grasp of the database landscape and how MongoDB offerings are positioned within it.

Listening to the presentation, I had a couple of take-aways. As readers know, I am long-term bullish on MongoDB. My optimism is driven by a few factors – 1) the overall size/growth of the data storage market, 2) the current “database refresh” occurring as a result of new app development and re-architecture of legacy systems, and 3) MongoDB’s popularity with developers and leading position in coding bootcamp curriculums.

The following is a bullet point list of my notes and reactions, as I listened to Michael respond to questions from the Needham analyst.

  • The database market is the largest in all of software – $64B spent in 2019, estimated to grow (per IDC) to $97B in 2023. This represents $33B in growth over the next four years. This growth is due to the fact that companies are increasingly trying to drive competitive advantage through custom software development in order to deliver unique customer experiences. At the center of this trend are new software applications, which include a database. The database market will support a handful of providers. There won’t be one dominant player, like there has been historically.
  • MongoDB leverages two dimensions for growth in enterprises – new app development and migration of legacy workloads.
  • Extrapolating projected $33B market growth over the next four years on an annualized basis, we get about $8B of new spend a year. Most of this is associated with new applications or workloads being brought online. MongoDB is well positioned to win a meaningful share of this business.
  • The $64B of existing spend isn’t being put out for RFP every year – that opportunity is driven by the legacy application lifecycle. Michael estimated for easy math that an existing application would be considered for a refresh/rebuild every 10 years. This would mean that about $6.4B of spend a year would have the opportunity to reconsider the application architecture, including the database choice. I agree that re-architecting does represent an opportunity to switch database technologies. The migration to micro-services, for example, where some application workloads and data structures favor a document database solution, will divvy this refresh spend among multiple providers.
  • Michael said that MongoDB is winning an increasing share of these database refresh cycles each year.
  • I think the three points above around growth in spend are significant. MongoDB is projected to finish out the 2019 calendar year with a little over $400M in revenue. Assuming they maintain existing engagements with their net ARR expansion of over 120% annually, then they need only a small percentage of the $8B + $6.4B = $14.4B of potential new spend to realize meaningful revenue growth.
  • MongoDB has incredible developer mindshare. Michael referenced the Stack Overflow survey showing that MongoDB is the most “wanted” database to work with by developers. If developers prefer to work with MongoDB, this would support the ability to win some share of new app development spend. As a CTO, I can attest to the fact that developer preferences do influence technology decisions. While I would require vetting of technologies and a limited set of choices, in today’s competitive environment for developer talent, it is critical for retention to bias software stack choices towards the preferences of the development team.
  • For enterprise workloads, Michael stated that MongoDB addressed the last feature gap blocking the movement of legacy relational workloads to MongoDB by addressing multi-document ACID transaction support in 2018. This has allowed enterprise customers with large workloads to feel comfortable on MongoDB. Michael referenced an investment bank that moved a trading application and Fortune 100 companies moving order capture applications to MongoDB.
  • MongoDB download count exceeds the number of developers on the planet – 80M times. Developers are driving the choice of MongoDB at enterprises.
  • One challenge for MongoDB is to make sure that decision makers are familiar with the current state of MongoDB and its latest features/capabilities. The MongoDB solution has improved substantially over the past 5 years. From personal experience, I can relate to this. For a large dating application that my team built in 2012, we tried to build a new service using MongoDB as the data store. Unfortunately, the MongoDB couldn’t handle the high volume of data writes that our app generated. MongoDB has since addressed this. However, MongoDB does need to improve their perception among hard-core software engineering leaders that the solution will scale and meet their needs, versus being a “toy” database for unskilled developers.
  • Michael explained that database market penetration is different from enterprise software. Large companies will have many individual applications – maybe hundreds or thousands. Each of these can have a different database. This is unlike ERP, CRM or HR systems, in which there is usually a single large software solution. The implication is that MongoDB has a better chance of getting a foothold in a large enterprise and then expanding, versus the overhead of a full enterprise software sale. This is an important nuance relative to new account penetration. I agree that this makes sense.
  • Partners provide an extension of sales reach and awareness of upcoming customer relational DB migrations. MongoDB works with Infosys, Cap Gemini, Accenture and others. Also, they are partnered with Cloud providers (as well as compete). Cloud partners are selling Atlas and their salespeople have similar incentives.
  • Michael emphasized that the multi-cloud capability of Atlas is important. Customers at large enterprises are exceptionally afraid of vendor lock-in. He said they worry about not being able to move from one cloud vendor to another, if needed. They don’t want a proprietary database solution provided by Amazon or Microsoft that can’t be migrated to another provider, or on-premise, if needed. It is relatively easy to move application code, but not the database system. This makes sense to me – not as strong an argument relative to cloud providers packaging open source MySQL or PostgreSQL, but certainly for their more proprietary NoSQL offerings.
  • The Alibaba relationship formalized in Oct 2019 provides access to the China market. None of the big 3 cloud providers have a presence in China, yet it is a huge market. Alibaba is now licensing a “legal” version of MongoDB. They were previously offering an unauthorized version. Alibaba is effectively adding cost to their MongoDB service offering by adding in the new licensing cost to make it legal. This is counter intuitive, unless Alibaba really values the opportunity, which Michael thinks is the case. MongoDB is very popular with developers in China – the largest number of downloads are occurring there. It is a multi-year deal, with annual minimum commitments. License fee increases as volume increases, so there is upside for both parties. The long term opportunity is large.
  • On questions around the licensing model, Michael clarified that MongoDB the company built MongoDB the software. They own all IP associated with it and the software development wasn’t crowd-sourced, like other open source projects. MongoDB created an open source version (freemium offering) to encourage developer usage/experimentation. This arrangement allows MongoDB the company to own the product roadmap. For investors, this is a key point, as MongoDB sometimes gets lumped in with other companies that support open source projects. Those companies can only offer services around the open source, like training or implementation support, versus actual licensing. MongoDB can generate revenue from licensing, because it is their software. The same rationale applies to blocking out the cloud providers, who can really only run the developer version of MongoDB at this point.
  • mLab acquisition closed on Dec 1, 2018. Q4 FY 2020 (calendar Nov 2019 – Jan 2020) is the last quarter of impact on year/year performance from the acquisition. This is important to note for investors, because mLab is actually having a negative impact on revenue and gross margins in the short term. This is due to the migration of mLab customers onto Atlas. It partially explains the low year/year growth projected in revenue for Q4. There has been expected attrition of mLab’s customer base. Also, MongoDB will give mLab customers lower pricing on Atlas, if they had it on mLab.
  • Finally, Michael commented on the recent capital raise – provides access to $1B. Allowed MongoDB to repurchase the prior convertible at more favorable rates. Will continue to pursue M&A opportunistically.

Overall, I left the call with more confidence around the long term opportunity for MongoDB. I think the comments around market growth and new spend on database solutions will favor MongoDB well. Even small slices of that new spend annually would represent meaningful revenue growth from the current $400M annual run rate. MongoDB needs to continue to build out their product offering and sales motion. Competitive offerings, primarily from cloud providers, will continue to improve. However, I think that developer preferences for MongoDB and large enterprise fear of cloud provider DB solution lock-in will result in continued wins for MDB.