My Take on MDB

Investment Analysis Clubs Saul’s Investing Discussions
1

Sorry - I’ve been too busy even to lurk (and I’m happy with my current portfolio), but I did want to drop by and present my thoughts on the MDB concerning the competition from AWS DocumentDB.

So let me start with my basic thesis. MongoDb has won the NoSql wars, and it is only going to grow for the next 20 years. Really that is how long I intend to own it without selling. My reasoning is like this: Databases are a winner takes most market. Right now most enterprise data is stored on mainframes (Cobol,DB/2) and on db servers (MySql, Oracle, DB/2, Postgres). New datastores will be starting up on NoSql, mostly on the cloud, and there will be a steady movement of some old data stores to new environments.

Because databases are a winner take most market, MongoDb is going to benefit from most of that growth, which is why you want to own it for 20 years. They are very unlikely to be upset by any new technology - they will have the money and the market penetration to see any dangerous technology well before it takes their market share, and either build it or buy it to keep ahead of the others.

Now this doesn’t say that there won’t be competition. This is a huge market, and everyone would like a piece - but they are going to struggle for crumbs while Mongo surfs on the wave.

I expect every single major web player to try to take a piece of the pie. And every time a large company puts out a new offering, MongoDb’s stock price will swoon. But the stock will swoon with people taking profits, not because the MongoDb business model is actually affected.

Amazon’s offering is one example - this is a big company that knows how to win on the web. But Amazon’s offering is Mongo 3.6 - anything higher is covered by the new SSPL license. So how is Amazon going to keep up, when Mongo is already on 4.1? Amazon has had its own NoSql databases, with some limited success for a decade. But the industry is not interested in a wholesale move to an Amazon database - MongoDb is the clear winner, and Amazon DocumentDB 2.0 is not going to interest people who want MongoDb 4.2

I expect we’ll see lots of swoons over the next 20 years - buying opportunities. Just wait until Oracle buys a big NoSql database and reports that a very big client has decided to use Oracle NoSql (not mentioning that they will be moving from the expensive Oracle Sql db to an almost free version of the NoSql db). Mongo could fall 30% in a day. But that is just profit taking and emotion.

Now, I’m not saying that there isn’t risk. Mongo will need to stay on top of the technology, and on top of the price competition. But these company announcements aren’t going to affect the 20 year uptrend. The only risk that will make me rethink my ownership in Mongo, is if on an earning report their growth slows unexpectedly, at a time when other technology spending doesn’t slow. Except for earnings reports, I’m not even interested in Mongo news,

Just my view.

129 Likes
2

The question for me is, are there better places to put my 14% position.

3

The question for me is, are there better places to put my 14% position.

I would respond, are you looking for a trade or an investment? There may be better short term trades, after the almost triple from the IPO, but I don’t believe there is any better company to own for the long term.

12 Likes
4

I am also optimistic about this area for the future.

But how do you assess Lyft’s ‘dissatisfaction’? what was it for them to switch?

tj

5

Lyft has been using DynamoDB for their main data store since 2014. Any usage of mongodb was limited (trial, prototype etc).

7 Likes
6

Doesn’t read like that in mLab’s case study.

https://mlab.com/case-studies/lyft/

What’s your source?

7

Aws’s Case study where they profile lyft, where the cto of lyft said they began using DynamoDB in 2014 and been 100% aws all along.

2 Likes
8

Lyft wants to go public soon. Perhaps Lyft believes it is better PR to say it is using Amazon’s AWS versus an open-source MongoDB database.

The main concern I see is this is now twice in recent months we are hearing about subpar service provided by MDB. We have to keep an ear out for this.

That said, MDB is only 5.5B, who know, Oracle can buy them with cash to get into the NoSQL space.

Still holding a 3% position.

9

you might note that the mLab Lyft case study mostly discusses pre-Lyft. Further, the case study is from mLab, a company with less than $20 million in revenues.

But any excuse to start a panic. None of the panics have actually impact Mongo for more than a day or two, if that.

You know what we should do, just to make sure we are always in panic mode, is that every time a company loses a competitive, is replaced (even if just partially or for some insignificant business), or there is customer churn, announce it!

Take TTD. 95% customer retention. 5% customer non-retention. “The Trade Desk loses customer to Amazon DSP; they lose another customer to Amazon DSP due to dreadful customer service and inept product,” from discussion with insider at TTD. Oh no, The Trade Desk is going down the tubes!

How about “Palo Alto Networks loses bid to Checkpoint software with large international bank that previously had great relationship with Palo Alto.” Palo Alto’s last earnings report, just yesterday, it must be a fraud!

Or we could just say “business who runs mission critical applications on Oracle, and who is sick of Oracle, becomes a Mongo customer.” Oh no Oracle falling apart!

There is a post on the Mongo premium board actually claiming Mongo only has 1% insider ownership. Then when corrected he comes back the link correcting him was from 2018 and not now. Thus there was a historical and massive insider trading from Mongo where all insiders dumped their ~65% of insiders shares!

Obsess over something that has some meaning vs. tossing out noise that has no relevance.

Tinker

21 Likes
10

The “dissatisfaction” may have been nothing more than needing to have a separate skill set for a minority functionality.

11

how do you assess Lyft’s ‘dissatisfaction’? what was it for them to switch?

I don’t know this particular client or decision. The way that I look at it is like this:

Every single company in the entire world, that is online, needs one or more databases. In addition, many of them have existing databases on old technology that they are considering porting to something that works better for their needs.

MongoDb not only has mind share among developers, it also has mind share among decision makers. They also have a tremendous technical lead for generic NoSql database technology, due to their WiredTiger engine and their 4.0 capabilities including ACID transactions and Stitch Serverless API.

The database market is risk averse - databases are things that are under the covers that make your system work. This is not like an application or front end technology - your website isn’t going to look spiffier because of the database you use. Because of this, once the industry settles on the best generic database, this database wins most of the bake-offs.

Industry has settled on MongoDb as the best generic NoSql database, for a variety of reasons. So in most cases, where NoSql is an option, MongoDb will win. This will happen in, say, 2/3 of the cases. In the other 1/3 of the cases, one of the 50 other NoSql competitors will win. 1/3 isn’t a trivial number, it’s just divided up among a lot of competitors.

Competitors win for a variety of reasons. Sometimes the need isn’t the typical generic database need - you may need something special. Other times the salesman is really good, or is ready to sell at a loss, or the decision makers have already decided on a competitor.

This makes great news for the competitor and gets trumpeted. But unless we are talking about actually changing the leadership, one offs here and there are part of the landscape. It’s not real news - there are thousands companies that are bigger clients than Lyft for MDB.

Real news is if there is a sustained change in the leadership of db technology. Because of the nature of this industry, it is almost impossible for a competitor to take over the leadership with even significant incremental changes in technology - they are too easy to copy, and the leader continues to have all the other advantages. Only a radical revolution in db technology can change the leadership, which is occurring now as MongoDb takes over from Oracle. It’ll probably be 20 years for the current NoSql technology to be digested and a new revolution to occur.

So my short answer is that the Lyft story is not really relevant. Everyone in db already knows what Mongo 3.6 is, and how Mongo 4.0 differs from it. Very few companies are likely to put mission critical data stores on DocumentDb, as it won’t stay up to date with the Mongo releases.

DocumentDB will have its own forward path, and will get some customers, but it can’t threaten Mongo’s leadership.

So the best idea is to calmly let DocumentDB have it’s day in the news cycle. You would need 100 Lyft’s to have something to be concerned about.

If a real threat to Mongo’s leadership was to occur, I’ll probably hear about it six months before it hits the news cycles, and everyone will probably see it first in a slow down of Mongo growth on a quarterly report.

45 Likes
12

Thanks Steppenwulf and Tinker, I respect and appreciate both of your views on the MDB situation.

I’m not selling any, may actually add for the long haul!

2 Likes
13

SteppenWulf,
I mostly agree with you. But MongoDB will never “take over” Oracle IMO. There are too many use cases well suited to relational. None of them are going away anytime soon, if ever. NoSQL is an up and coming technology that better fits use case that employ various “document” formats, the stuff called BLOBs (binary large objects) in the SQL data world. The best relational can do with this stuff is to just cram it into a column for binary content (or in some cases, just carry the address of the BLOB in a column slot).

From what I read, Lyft had issues with performance. It was not clear if the issue was with vendor performance or DBMS performance. There’s a big difference. But, to be honest, I have some trouble understanding this complaint from either perspective. If it’s vendor performance, that belies almost all other reports regarding MDB. Customers seem to be largely happy with the vendor support. It also assumes they will receive more and better attention from Amazon, I think that’s a stretch. If it’s DBMS performance, I’m struggling to understand how locking themselves into an earlier version of the product ameliorates this problem, either now or long term.

Could it be that there’s something else going on with this switch? I don’t know, but it wouldn’t surprise me if there’s another game in play here.

2 Likes
14

Industry has settled on MongoDb as the best generic NoSql database, for a variety of reasons. So in most cases, where NoSql is an option, MongoDb will win. This will happen in, say, 2/3 of the cases. In the other 1/3 of the cases, one of the 50 other NoSql competitors will win. 1/3 isn’t a trivial number, it’s just divided up among a lot of competitors.

What can we say about the NoSql TAM, if this is already currently the case, and MDB’s annual revenues are under $300m/yr right now?

I can’t quickly find online any confirmed NoSQL market numbers, but see this link:
https://www.theregister.co.uk/2018/05/31/rise_of_the_open_so…
https://www.computerworlduk.com/galleries/data/best-nosql-da…

In a few places I seen numbers of approx $4b for NoSql in 2020 type of thing.
Your math tells me that Mongo would be about $2.6b in revenues…so I guess I am trying to understand the disconnect here?

If the answer is that the $4b isn’t just “generic” NoSql but also includes AWS or Oracle NoSql solutions and you are excluding those?

Just trying to make the math work to predict a path of MDB revenue growth next 3-5 years.

Dreamer

2 Likes
15

Dreamer, I am not sure about TAM, but when I read online, it says 80% of all business data is unstructured, which is also possibly an over exaggeration. So the pie in the sky theory is that MDB could actually some day be bigger than Oracle.

16

The only risk that will make me rethink my ownership in Mongo, is if on an earning report their growth slows unexpectedly, at a time when other technology spending doesn’t slow. Except for earnings reports, I’m not even interested in Mongo news

I’ve resisted commenting too much on MDB, because I’ve gotten to know a lot of sales reps and executives currently working there, and I don’t want to publicly comment about them in an unethical way. I’ve been in an enterprise software sales role for almost 20 years in Europe and have seen a lot of people come and go.

However, I think it’s okay to publicly say that these people are among the best sales people I’ve worked with over the years, and they are up to the many challenges they’ll face over the years to come. This was the primary reason I finally invested in MDB about a year ago.

I won’t post anything else about these people. But if this situation changes and I see a significant portion of them head for the exits without getting proper backfill, I’ll strongly consider exiting my position. If I do, I’ll post more.

That said, I expect the earnings reports to be pretty good for the foreseeable future.

DJ

50 Likes
17

Steppenwolf, Tinker and DJ -

Your writing is abundantly clear, logical and reality based. Thank you for your posts today; very helpful points of reflection and consistent with my long term approach to investing.

MongoDB remains a moderate sized position in my portfolio and I am still very excited about their market traction and potential to be among my very best long-term investments. It by definition is one of my high conviction holdings given I have less than 15 positions.

Fool on.

–Rockleppard

1 Like
18

“Could it be that there’s something else going on with this switch? I don’t know, but it wouldn’t surprise me if there’s another game in play here”

Maybe Amazon is making a move into the space?

19

Given that Lyft was a user of mLab and mLab was $20M total revenue, Lyft could not have been a major Mongo customer.

3 Likes
20

I just listened to an event that MDB spoke at. The question came up about Lyft. MDB basically said that they were an mLab customer, their spend actually was an outlier for mLab but that it would not mean anything to MDB as far as revenue. MDB said that most customers for mLab spent about $5,000/yr. MDB didn’t learn anything new from the news yesterday. Lastly, they said that Lyft was using an old legacy version of MDB and they were not on Atlas. I tried to take notes quickly so some things might not be exactly correct. A recording should be available soon.

I say keep your MDB stock! I added yesterday.

https://investors.mongodb.com/events-and-presentations/event…

26 Likes