MongoDB vs Oracle - Thoughts.

MongoDB management has repeatedly claimed their cost of license and maintenance is a fraction of the cost of Oracle.
How do they compare ?

  1. Licensing costs : From a quick google search I gathered the following:

a) Oracle: The license costs are: Enterprise Edition – $47,500 per unit (sockets * cores per socket * core factor) Standard Edition – $17,500 per unit (sockets) Standard Edition One – $5,800 per unit (sockets)

In this example we are going to set up a 2 node RAC cluster with DataGuard to a standby 2 node RAC cluster. We also want AWR/ASH for diagnostics. Each server (4 total) will have 2 Intel Xeon X7560 processors (8 core).

Our costs for Enterprise Edition would be (in list prices):

- 4 servers * 2 sockets per server * 8 cores per socket * 0.5 core factor = 32 units
- 32 Units Enterprise Edition ($47,500ea) = $1,520,000
- 32 Units RAC ($23,000ea) = $736,000
- 32 Units Diagnostics Pack ($5,000ea) = $160,000

Our grand total for this setup in EE would be $2,416,000. Hopefully you have a good discount!


b) MongoDB: MongoDB Enterprise comes in two tiers, with Core costing $6,500 per server per year and Advanced priced at $10,000 per server per year. Additional licenses are required for servers containing more than 512GB of RAM.


So MongoDB indeed is a fraction of a cost of Oracle. I think this is very important when evaluating Mongo as an investment For two reasons.

One Oracle has a classic Innovators dilemma here. Even if it puts in efforts to come up with NoSQL alternative (I have no doubt it can in some time if it wants) would it be willing to forego revenue willingly to compete with Mongo ? (Note: -30% of Mongo’s growth is from migration of relational to non relational databases).

Secondly, database is a very sticky business because of its mission critical nature and huge switching costs. Incremental improvements will not be enough for companies to migrate their workloads. The fact that Mongo licesnses are significantly cheaper than oracle, hopefully will create incentives to overcome this inertia.

  1. Maintenance costs : The whole point of No SQL database is dynamic schemas which gives flexibility and programmer productivity. It is much more closer to the datastructures of typical programming language, hence reduces the cost in maintaining it.

A snippet from earnings report transcript:

Before I review some of the operational highlights of our third quarter, I want to take a few minutes to reiterate the investment thesis we presented during the IPO road-show since this is our first earnings call as a public company. MongoDB was founded by developers who were tired of working on constraints of relational database. They started with a clean sheet of paper and built a database they always wanted, a solution by developers for developers to allow them to more easily and efficiently build mission-critical applications.

As software becomes more strategic, organizations are choosing technologies that drive developer productivity. In fact, the increasing importance of developers was giving them disproportionate influence on how technology is evaluated, adopted and ultimately monetized. And because database is at the heart of every soft application, slicing of database has become a highly strategic decision. But most of the applications today are built on database technologies that were introduced over 40 years ago and a lot has change in that time.

Organizations are moving applications away from legacy computing platform deployed on-premise to modern as our architecture that can be deployed either on premise or in the cloud. These changes made to database market right for disruption just as we’ve seen at the application layer, the management layer and the infrastructure layer of the IP stack in recent years. That’s what we think with MongoDB and is only scratched surface of the opportunity in front of us. MongoDB was built to address the needs of modern applications and to maximize development productivity.

Our unique platform combines best of relational and non-relational databases. We preserve aspects of legacy databases that developers have come to value like sophisticated access to the data, guarantees for data integrity and comprehensive management functionality to operate monitor management database platform. We combine this with the flexibility, scalability and always on reliability of modern approaches.

The developments of platform that enable organizations to build and deploy mission critical applications on-premise during the cloud for an incredibly broader array of use cases many times fastest than they could with the legacy products. This approach has completely changed the game for application developers and we believe this is reflected in our adoption and mindshare.

  1. Number of Customers: Oracle has 450,000 enterprise customers, MongoDB has around 4500 only so far ! 600 of them were added last quarter alone !! What this tells is acceptance of MongoDB in enterprise is only started to gain traction and the runway is really long. Database manage mission critical aspect of business. So it is natural any new database will face skepticism and resistance before gaining acceptance. They are held guilty till they prove themselves innocent. So though Mongo has 30 million free downloads, traction in enterprise has just really started.

Even if NoSQL does not gain traction over SQL, this will likely be a very good investment. And if it does gain traction, depending on the degree of its success it will be anywhere from a great to one of the greatest investments !!


One Oracle has a classic Innovators dilemma here. Even if it puts in efforts to come up with NoSQL alternative (I have no doubt it can in some time if it wants) would it be willing to forego revenue willingly to compete with Mongo ? (Note: -30% of Mongo’s growth is from migration of relational to non relational databases).


Thanks for that post. I also saw a few similar cost comparisons that favored Mongo and particularly on subsequent carrying costs which are much higher with ORCL.

I agree with your statement above and it is a point I made in the original thread from the transcribed investors conference post…at first blush, if one contemplates going up against ORCL, it might seem they would squash small competitors like a bug. But then one realizes that to compete, they would have to cannibalize their own business.

At that investors conference, the CFO stated they are NOT seeing “modern” new databases put out there by ORCL…probably for this exact reason.

It would seem to me that ORCL will likely/is likely trying to make their SQL databases better compete against NoSQL but even there for this behemoth of a company, when considering the TAM as we have stated with several different references, it is smaller than their main SQL database TAM and I doubt they are eager to make their product more price competitive.

Secondly, database is a very sticky business because of its mission critical nature and huge switching costs.

Agree and this is what I was referring to as a long sales cycle…if the legacy databases are working for you…there is no need to change and those potential customers are going to be VERY difficult nuts to crack for Mongo.

I referenced this Nov, 2017 article on the NPI that seems the most probable medium term outcome IMO:…

Stifel’s Reback, in his overview note, sees the two camps continue to co-exist:
From our discussions with both customers and vendors in the database market, we believe that while NoSQL databases will play an increasingly important role in enterprises’ future database needs given the accelerating velocity and volume of data types being generated by social, mobile, and IoT applications, relational databases will remain a crucial part of their overall IT stack and business strategy.

As a result, we expect enterprises to take a hybrid approach to their database deployments, and leverage a combination of relational and nonrelational databases in order to take advantage of each database’s respective strengths/weaknesses. We believe the opportunity presented by both of these segments of the database market is significant and continues to grow, with recent reports published by IDC estimating that relational and nonrelational database system spend will grow at a 5-year CAGR of 6.3% and 18.4%, respectively to $42.3B and $12.4B by 2021.

If such a “hybrid” model is correct (and the weight of evidence suggests it is), MongoDB’s design of using “familiar/similar features” to SQL would seem more attractive to a customer than multiple other niche NoSQL products all that require their own process/language/etc.

So if we put this TAM issue in context with MOngo’s present market cap of $1.9 Billion, we have seen a few estimates of TAM and growth rates:

  1. Relational and nonrelational database system spend will grow at a 5-year CAGR of 6.3% and 18.4%, respectively to $42.3B and $12.4B by 2021.

What would the market cap of Mongo be if it could acquire 20% of the nonrelational database market in 2021??? That would be a company generating $2.4 Billion in revenue with a present $1.9 Billion market cap??? That also would ignore any land and grab from the relational database that we know they are doing even now.

  1. Generic database TAM estimated at over $65 Billion annually with lower to higher hanging fruit being NoSQL ($4 billion), new SQL ($16 Billion) and legacy displacements ($45 Billion). The latter two categories are really mostly accessible after release 4.0 this summer.

That multidocument engineering feat took 3 years…incredible accomplishment but it also points out how hard it is to compete against relational that has been around for decades. The above numbers differ significantly from the referenced article but still reveal what I gather is a huge TAM in the latter two categories…rapid relational (data from mobile, sensors, etc.) and NoSQL…in total around $20 Billion (essentially acquiescing that the legacy databases would never switch over which isn’t true but we will just make that assumption for the moment since those are very long sales cycle projects). Again same assumptions as above with market share grab by Mongo…what would its market cap be if it could grab 20%???

IMO, this is why the original thread specifically stated that this stock could be greeted with decimation or multiple magnitude orders of appreciation…at least we have ferreted out that the TAM is large enough for the latter.

IMO, there is a reasonable probability that MongoDB gets bought out before the latter…specifically to either ORCL or MSFT. I base this probability on the fact that the founders of MOngoDB already did that with their prior startup DoubleCLick so they have a propensity to sell out but also that the potential suitor would almost surely rather retain/expand its market share than see it erode from a modern startup…again just speculation but considering historical and probability considerations.

BTW, I never invest on the basis of hoping for a buy-out, so to me, the above is not a reason to own the stock.

Unfortunately, the threads got distracted before we could try to bring more clarity to a crucial topic for Mongo - the actual business model itself. While I am satisfied that the market exists for orders of magnitude stock appreciation, does Mongo have a business model to exploit that market??

These three topics need to be explored further IMO:

  1. Fremium model - it is my understanding that while they open source the freemium product, they are in complete control of all licensing and can move features over the wall in either direction to paid vs freemium. I assume this freemium model of which there have been at least 30 MILLION downloads, is about building user familiarity and buy-in with a large mass of potential future paying customers. Judging from how many paying customers they have, how do they determine the effectiveness of this campaign?

  2. Atlas - this is a lower end Saas type model that seems to attract lower end customers that pay monthly fees. It is rapidly growing revenue now 8% of their total revenue from 1% last year. What is the lock-in for these customers and how will the three major cloud providers embrace or compete against them if there is less lock-in??

  3. Premium model - will the version 4.0 due out this summer really grab and expand previous relational business. Since this is not due out till summer…we will not know this answer until well in to 2019 IMO. We know already 30% of their revenue is from previous legacy databases but they are very vague on how that looks quarter to quarter.

Are these three business elements a succinct business model that promotes their growth into the TAM described above?? That may be an important discussion and one that Tinker helped identify with another favorite around here in SHOP…not all merchants are alike…and so with Mongo…not all of the 3 customer categories above are alike.

As an aside, I hope it is obvious that we have taken a very responsible and investigative approach to evaluating MongoDb…the good, the bad and the ugly. On the NPI with my many acquaintances in Tinker, Denny, AJ, Mauser, Ant, Darth, MC, and others, we ask and expect that anyone shoot holes in an investment thesis…what I call crowdsourcing investments.

I merely hope that we see posters produce links, research, corroboration of their thought angles just as TechnologyGrowth, DreamerDad, RhinCT have done and are much appreciated.

As GHonkaba has pointed out, stock lockup expiration occurs 4/17/18, earnings are next week (only their second ever as a public company), they have only $140 million revenue run rate…it doesn’t get any more risky in an investment at this moment in time.

I regret that the Mongo may have taken up so much bandwidth on this board and we can move the discussion back over to the NPI if Saul and others would prefer.