MDB vs ESTC

Me: Maybe a better challenge for Saul: Why do you feel differently about ESTC than MDB?
https://discussion.fool.com/i-get-sauls-argument-and-i-get-dream…

Saul, I really am curious about this. Numbers-wise, both look very similar.

                 MDB      ESTC
Mkt Cap          7.2b     5.5b
PS ratio          21       18
P/GP              31       26
GM (last Q)       70%      71%  
OpEx/GP          154%     167%

What is it about the products or the competition or the market for each that makes you a fan of one of these (MDB) and not so much the other (ESTC)?

I think I’ve heard some say that MDB is the leader in its space and ESTC is in too many spaces and is not the leader. I think ESTC is the leader in search, though, and I don’t understand why we would ding them for having some optionality.

Just interested in your thinking.

Bear

24 Likes

What is it about the products or the competition or the market for each that makes you a fan of one of these (MDB) and not so much the other (ESTC)?

I think I’ve heard some say that MDB is the leader in its space and ESTC is in too many spaces and is not the leader. I think ESTC is the leader in search, though, and I don’t understand why we would ding them for having some optionality.

I think you basically have answered your own question. I don’t get much out of pronouncing same or different, yes or no, based on some numbers. Yes, there are a bunch of non-SQL databases, but MDB has an extremely commanding position. Yes, ESTC might be a leader in search, but search doesn’t feel to me like a product on a par with MongoDB. It is more an ingredient of a product. While that might provide optionality, it also creates an environment in which someone else can provide a product using ESTC search engine and ESTC not get much out of it. And, if someone else came out with a better search tool, those companies could probably switch to using it without a huge amount of rework since most of their product is not search. Whereas, switching to another database from MongoDB is a major rework.

8 Likes

Why do you feel differently about ESTC than MDB?

Hi Bear, Well as you know I don’t really feel great about either one. I’ve sold out of each of them several times, and got back into Mongo, but I don’t like the open source model at all. It sounds so noble but it seems to make it much harder to make money. With Mongo, they have Atlas at least which is not open source, but is all theirs, and all subscription, and while still smallish, is growing as a part of their system, and is now up to 37% of revenue. (I do know that Elastic also has parts which are paid, and I am constantly tempted by smart people saying nice things about them).

          MDB      ESTC
Mkt Cap          7.2b     5.5b
PS ratio          21       18
P/GP              31       26
GM (last Q)       70%      71%  
OpEx/GP          154%     167% 

None of those are the numbers I look at except Gross Margin. I look at revenue growth rate, movement to break even, cash flow, etc and stuff like that.

Gross Margin, as you say, is about equal.

Elastic, as I pointed out is losing more each quarter than the quarter before. But when I look at it, I see that Mongo is doing the same. Their last three quarters LOSS were roughly 10.5% of revenue, 13.5%, and 15% of revenue. Not a happy sequence either, but at least that 15% is improved from 30% a year ago.

As I also pointed out, Elastic’s revenue growth rate is falling each quarter, but I find that Mongo’s isn’t any great shakes either 71%, 78% and 67% in the three quarters since they shifted to ASC 606.

Elastic has considerable negative free cash flow, but Mongo isn’t any better. Their operating cash flow was minus $13 million this quarter and free cash flow was minus $14 million.

Well doing this calculation has succeeded in making me wonder why I have a 10% position in Mongo at all. I guess I’m banking a lot on Atlas but maybe it should be a bit smaller position. Thanks, Bear, for making me look at it.:grinning: Perhaps it would be better to put some of the cash in Alteryx, Okta, Datadog, Coupa, or Zoom. (Not racing back into a larger position in Zscaler yet).

Best,

Saul

33 Likes

Thanks Tamhas, I didn’t see your post before I posted, but I should reference everyone reading my last post to yours just before it. You hit the nail on the head.

Yes, there are a bunch of non-SQL databases, but MDB has an extremely commanding position. Yes, ESTC might be a leader in search, but search doesn’t feel to me like a product on a par with MongoDB. It is more an ingredient of a product. While that might provide optionality, it also creates an environment in which someone else can provide a product using ESTC search engine and ESTC not get much out of it. And, if someone else came out with a better search tool, those companies could probably switch to using it without a huge amount of rework since most of their product is not search. Whereas, switching to another database from MongoDB is a major rework.

Saul

7 Likes

Well doing this calculation has succeeded in making me wonder why I have a 10% position in Mongo at all. I guess I’m banking a lot on Atlas but maybe it should be a bit smaller position. Thanks, Bear, for making me look at it.:grinning: Perhaps it would be better to put some of the cash in Alteryx, Okta, Datadog, Coupa, or Zoom. (Not racing back into a larger position in Zscaler yet).

What about Crowdstrike? Their move up today, especially given the reasons people have offered for it (https://seekingalpha.com/news/3518839-crowdstrike-plus-6-per…), makes me think their whole huge drop is completely based on FUD. Do you feel differently? I recall you added quite a bit in October, but maybe you’ve changed your mind?

My guess is you’re changing with winds of the market. Jettisoning the inefficient and embracing the more profitable. I’m guessing your sentiment, like the market’s (at present), probably grows as the percentages descend on the following chart, which shows last quarter’s OpEx divided by Gross Profit.


ESTC          167%
CRWD          166%
TWLO          160%
MDB           154%
**ZS            112%**
DDOG          106%
ZM             98%
AYX            87%
TTD            82%

Unless Zscaler implodes when they report on Dec 3 due to a huge revenue slow-down (45% growth or below, in my opinion, would be pretty crushing), I could see them bouncing back in a big way, since their profitability/efficiency looks better than a lot of ours, and since that seems to be key at present. So I guess that could go either way, like you say.

AYX continues to look just so incredible. I really don’t see why the market doesn’t love it as much as we do. I think it’s 25 - 50% undervalued at this point.

Bear

19 Likes

Bear, quick question, why do you use that calculation vs something like Operating Margin?

Yes, there are a bunch of non-SQL databases, but MDB has an extremely commanding position. Yes, ESTC might be a leader in search, but search doesn’t feel to me like a product on a par with MongoDB.

You are correct that search can be thought mostly of as a “service”. But to say that it is “not a product” in comparison to Mongo is probably misleading. Mongo can be thought mostly of also as a “service”. It serves to store information and allow search and retrieval of this information. You build a product on top of Mongo - you do not sell Mongo as a product yourself (not typically anyway), just like you would not sell an Oracle database yourself - you use it to store and retrieve your stuff and build a product on top of it. Actually Mongo is a lot closer to ES than you think. ES is just a lot more flexible with what you can do with it, meaning, you can use ES to do “storage/retrieval in unconventional ways” that not even Mongo can touch.

It is more an ingredient of a product. While that might provide optionality, it also creates an environment in which someone else can provide a product using ESTC search engine and ESTC not get much out of it.

Same can be said of Mongo. They have generally different use cases with some degree of overlap. What I mean is if search is my main concern, then ES (or any search technology) is at the top of my priority. If unstructured storage with database-like capabilities is my main concern, then Mongo (or any other document DB technology) is at the top of my priority. Keep in mind this is not either or, you can use ES to do some Mongo use cases, or vice versa. It’s just that ES is strong in some areas and Mongo in others. I can self-host Mongo and build a product and not pay for it, the same way as I can do to ES. Or I can pay for Atlas and build a product on top of it, the same way as I can pay for Elastic Cloud.

And, if someone else came out with a better search tool, those companies could probably switch to using it without a huge amount of rework since most of their product is not search. Whereas, switching to another database from MongoDB is a major rework.

Although I do not dispute your assertion, it is again a bit misleading and is likely not based on experience (no offense). First off, if you know anything about enterprise search, there have been a lot of commercial products that dominated this space for a long time. At the same time open source search have also evolved for a long time (specifically Lucene and anything based on Lucene such as SOLR, and of course the newcomer ES). It just so happens that the (search) space is at a point in time now where the open-source solutions have reached mainstream that they are now dominating the space (yes even in the presence of existing proprietary search solutions from the biggest cloud vendors). What I am saying is sure it is possible for a better search technology to come along but it is unlikely (at least for the forseeable future). The reason for this is a company like ES is on top of the industry and they are improving their capabilities at breakneck speed since they started the project about a decade ago, and look at where they are now (I can say the same for Mongo BTW - they have matured a lot since they started). Elastic provides open-source products that literally powers some of the world’s most widely used applications and use-cases and not to mention is used to solve some of the most interesting problems today. The amount of effort to switch from ES to a different search solution is likely no easier (or harder for that matter) as switching from Mongo to another document database. It is not something that can be generalized and you’d probably be surprised at how “easy” (relatively) it is switch things around these days given the more open nature of the APIs (REST and HTTP).

Anyway, I am typing this not to be a pro/con going one way or the other in terms of investment. I am just adding some clarification to some of the misconceptions and nuances surrounding these technologies.

31 Likes

The difference is two fold: (1) Mongo’s license no longer allows a cloud titan to deploy and commercially sell a version past 3.6 (which is ancient in feature time for Mongo) - whereas, although it appears AWS, for example, has not done a real good job of it, Elastic remains deployable through its most recent version by say an AWS; (2) Mongo commercially dominates its space with no real alternative - Elastic, you can go with Datadog (built on Elastic - but for free), Splunk, and all the other alternatives, of which are many, and not miss a beat.

We went through this exercise last year or so when AWS and Azure were promoting their versions of Mongo and the change in open source licensing issue. We looked at, if Mongo did not exist what substitute exists?

The answer: NONE. If Mongo never existed, there is no commercial product out there today that can do materially what Mongo does. No real substitute. Postgres can do some use cases. Couchbase some use cases. Cassandra some use cases. No one who is capable of being the modern general purpose database in the world.

Elastic: the underlying code that makes Elastic unique is out there for free and commercialized everywhere BY OTHERS who pay nothing to Elastic; e.g. Datadog. If Elastic was simply an open source society and not commercially running Elastic, so that the underlying search engine technology and related features that are its core was maintained, there is a multitude of commercial companies that you can use instead of Elastic for everyone of Elastic’s use cases. EVERYONE - and to do so in totality (unlike just a subset of use cases here or there, but big compromise here or there if you don’t use Mongo).

Then comes Datadog…who puts both Mongo and Elastic to shame given its viral adoption, relatively minor sales and marketing costs, and self-service ability requiring little in terms of professional integrators to get you up and running. That is the iPhone of Elastic products (iPhone came along and made complex applications and features so intuitive that no user manuals required - even for the stupid - to become an iPhone expert user).

As for the valuation issues…come over to my site and have a say! I use to say that the best of the best are almost always able to print cash even earlier in their life cycle (Arista, Nvidia, Microsoft, Cisco, Juniper, Network Appliance, EMC…all were cash printing machines early in their existence as examples) then came the SaaS and we tossed that notion out the window. Instead all we wanted was land grab while increasing financial leverage…

Well, that is Datadog, that is Alteryx, that is Zscaler, that is Zoom and so on. Is that Elastic? Is that Mongo? {well, I think it is Mongo but for further discussion}.

Nutanix and Pure are so instructive. Growing like a weeds (or was with Nutanix, and still is with Pure) and yet look how tiny the multiple can grow to. That will be Elastic if it cannot prove financial leverage in its business model (and it very well may as there is a chance the growth rate for their SaaS offering is going to accelerate…but that is another complex discussion).

But back to the point, Mongo is singular. There is no present substitute if you want what NOSQL can do. With Elastic there are many alternatives, and many of them are built (and more valuable in the hands of someone else like Datadog) on the core tech that is Elastic. Elastic, the commercial entity, is not singular per se. The commercial Elastic is a choice amongst multiple valid choices to accomplish what Elastic accomplishes.

If that makes sense.

Tinker

Tinker

57 Likes

Elastic has considerable negative free cash flow,

Saul, I am not sure I would term it that way (considerable). I see only about -$3.3M for the quarter ended July 31st from page 10 of the most recent 10Q. With a current cash balance over $300M, 1% burn in a quarter is not necessarily “considerable” if the company is still in a hypergrowth phase.
10Q link: https://ir.elastic.co/Cache/399539602.pdf

This was a big improvement from the quarter ended April 30th, which had a FCF burn of -$20.9M.

Here are the press release links for the 2 most recent reports for Elastic, for quick reference for anyone. They list the FCF number on the next to last bullet under the highlights section near the top of each.
Quarter ending 4/30/2019 - https://ir.elastic.co/file/Index?KeyFile=398213094
Quarter ending 7/31/2019 - https://ir.elastic.co/file/Index?KeyFile=399380814

FCF is not, of course, the be all and end all, but it is a very important metric (as is Enterprise Value, of course).
https://discussion.fool.com/evfcf-ratioa-superior-valuation-metr…

volfan84
long ESTC
long EV/FCF considerations in looking at valuations as superior to merely P/E (or P/S)

1 Like

Understand that I said what I said from the point of view of someone who spent a large portion of their business life as an ISV for a company that supplied a database and 4GL in which I supplied an ERP application built on that foundation. When I sold my application, I sold the DB with it and, in many cases, the customer was only dimly aware of the source of the database, something I am sure is even more true with the larger ISVs for the same source who had more established brands and thus did not sell the virtues of the DB technology as strongly as I did.

I have no personal experience with either of these companies, so I can’t speak authoritatively about the vendor lock-in which either creates, but I would be very surprised if the lock-in with MongoDB was less than substantial while the lock-in with ESTC notably less, which was my point.

1 Like

I am not sure I would term it that way (considerable). I see only about -$3.3M for the quarter ended July 31st from page 10 of the most recent 10Q. With a current cash balance over $300M, 1% burn in a quarter is not necessarily “considerable” if the company is still in a hypergrowth phase.

volfan,

You only looked at one quarter and compared it to the comparable quarter. CFFO and FCF can be lumpy and can vary wildly from quarter to quarter. I like to look at TTM numbers for these metrics to smooth out once a year costs from things such as annual user conferences. Below are these stats for ESTC for as far back as I could find:


CFFO	FCF	TTM FCF
  4.1	  6.3	
 -1.8	 -2.8	
-23.9	-25.0	
  5.1	  4.8	-16.7
 -0.6	 -1.4	-24.4
 -8.7	 -9.9	-31.5
-19.8	-20.9	-27.4
 -1.7	 -3.3	-35.5

TTM burn is more than 10% of revenue and more than 10% of the cash balance. And these numbers don’t count share based compensation. More importantly, the burn is not shrinking (in fact they are trending up). By comparison, CRM and TEAM were FCF-positive when they were much smaller than ESTC.

Chris

13 Likes

AYX continues to look just so incredible. I really don’t see why the market doesn’t love it as much as we do. I think it’s 25 - 50% undervalued at this point.

Well Bear, I certainly hope so, as it’s my largest position by far. It’s at 21%, while Datadog and Okta are clustered at about 16%. I think the analysts got all tied up worrying about billing numbers instead of revenue numbers.
Best,
Saul

15 Likes

CFFO FCF TTM FCF
4.1 6.3
-1.8 -2.8
-23.9 -25.0
5.1 4.8 -16.7
-0.6 -1.4 -24.4
-8.7 -9.9 -31.5
-19.8 -20.9 -27.4
-1.7 -3.3 -35.5

Chris, thanks for that extra look-back. That does show a different picture than simply the latest quarter, and there is clearly a bit of a cyclical pattern there with the big cash burn quarter every year. They do need to start making more progress on improving that soon, as FCF does already take out the non-cash expense of stock-based compensation.

-volfan84
still long ESTC

1 Like

Hi Bear, Well as you know I don’t really feel great about either one. I’ve sold out of each of them several times, and got back into Mongo, but I don’t like the open source model at all. It sounds so noble but it seems to make it much harder to make money. With Mongo, they have Atlas at least which is not open source, but is all theirs, and all subscription, and while still smallish, is growing as a part of their system, and is now up to 37% of revenue. (I do know that Elastic also has parts which are paid, and I am constantly tempted by smart people saying nice things about them).

At first I stayed away from Mongo because it was open source and depended on Freemium income. That model had flopped with the most popular open source SQL database, MySQL. But Mongo changed their license and it is no longer truly open source. The open source community was (is) very unhappy with Mongo’s v 4 licence. But the real kicker is Atlas that is fast becoming their main source of revenue and the reason to invest in MDB.

Open source is a wonderful idea but a lousy investment.

Denny Schlesinger

BTW, now that Oracle owns MySQL my hosting service switched to MariaDB, a MySQL fork. I’m still using MySQL on my Mac

More than you ever wanted to know about MySQL/MariaDB

https://kinsta.com/blog/mariadb-vs-mysql/

2 Likes

(1) Mongo’s license no longer allows a cloud titan to deploy and commercially sell a version past 3.6 (which is ancient in feature time for Mongo) - whereas, although it appears AWS, for example, has not done a real good job of it, Elastic remains deployable through its most recent version by say an AWS;

Tinker, I don’t think this is correct. Elastic has a free but proprietary version called Basic which a cloud vendor or anyone else can deploy and sell. Elastic continues to create more and more differentiation between its OS and Basic version - think of the OS ver as similar to Mongo 3.6 and the Basic ver similar to Mongo’s newer versions.

https://www.elastic.co/subscriptions

2 Likes

I don’t think I’m wrong. Because AWS has been offering Elastic, Elastic moved their former pay feature, security, to core free stuff. Security is the #1 basic must have pay for feature. By now offering security for free it removes a reason to go w AWS and they are suing the #1 security alternative. Thus competitive pressures forced them to do this.

Any company can use all the latest and up to date features and offer for resale for free if they want. This can no longer be done w Mongo.

This said, yes. It is not easy to offer the premium features. AWS offers many, but apparently poorly. Meanwhile Datadog, a direct competitor, does it extremely well, perhaps even better, while paying Elastic nothing.

Elastic creates the core Datadog product for Datadog, who then creates all their own proprietary features and makes more money off core open source Elastic than Elastic.

In comparison AWS has anointed Mongo as the NOSQL winner.

So there is a clear difference. Elastic has their own strategies and lets see how they do. But in each market they face strong competition, including some running their own software for free!

Tinker

12 Likes

Not to belabor the issue. You maybe mixing up the OS and Basic version of Elastic. The basic version is free but proprietary to Elastic. AWS cannot sell it. The basic version has several features that OS version does not. That is the reason why AWS created Open Distro in the hope that the OS version of Elastic can achieve parity and of course they can sell it:)

I think Elastic has developed many new products in the last year. I agree with Darth that there should be no sales elongation issues. The bigger question though is whether customers will use all their new capabilities. If they do then Elastic sales will grow.

4 Likes

The answer: NONE. If Mongo never existed, there is no commercial product out there today that can do materially what Mongo does. No real substitute. Postgres can do some use cases. Couchbase some use cases. Cassandra some use cases. No one who is capable of being the modern general purpose database in the world.

I have an alternate view. :slight_smile:

As a side, for Cassandra check out Datastax Enterprise.

In a polyglot world one uses the best database for the job, which is generally a specialist database - think graph databases, NoSQL, NewSQL, traditional RDBMS (like Oracle, Postgres etc.). Mongo trying to be all things to all men (transactions spanning the cluster from version 4.x anyone? Urgh) means crowbarring in the feature on top of an architecture that wasn’t built for it. This results in a suboptimal implementation.

But then…

…how good does the implementation have to be to be good enough?

I think that that is where Mongo my have it right.

Good enough for most companies’ use cases.

Then there is support.

I believe they have an excellent 24/7/365 support model which is essential for critical applications Mongo is deployed as part of and for their larger clients that are willing to pay for this.

Disclaimer: Long MDB (although it may not seem like it hehe )

1 Like

I see that Saul posted his reasoning, turns out he’s not all that positive about either product.

I have a different take. Mongo is the premier No-SQL general purpose DBMS. Elastic is not. Large companies (like the one I worked at) tend to standardize on a given product (especially middleware like a DBMS) rather than going through a new evaluation for each application. The reasons for this is that it minimizes the internal support requirements and provides a lot of flexibility in terms of allocating DBA’s who can move from project to project without having to learn the ins and outs of a different DBMS every time they support a new project.

Combine this with the dog’s breakfast of products that Elastic has developed fairly recently. Each of these products must compete in a field already bristling with competitive products and the Elastic offering with respect to the security products (and probably their other offerings as well) are not the run-away the leader in the field, they are all also ran products. Agreed that at first blush Elastic’s end point software offers a sort of revolutionary pricing model, but when you dig under the covers what at first appears to be a blanket license irrespective of how many endpoints are engaged, you still end up paying for each implementation.

The way I see it, Elastic has to pour money into the R&D needed to grow each product and retain a competitive posture without corresponding revenue. While not a dilution of revenue, it is a dilution of resources.

It is primarily for these reasons that I recently sold all my Elastic positions and used the funds to buy Roku. I won’t go into any details or my reasoning for buying Roku, several folks that follow this board have already made the case for a Roku investment more eloquently than I could.

8 Likes

Brittle,

Agree with what you are saying.

Combine this with the dog’s breakfast of products that Elastic has developed fairly recently. Each of these products must compete in a field already bristling with competitive products and the Elastic offering with respect to the security products

Elastic has already existed in all of these markets for some time with an expansive list of users leveraging their platform to perform these use cases. They have an inherent presence. Many with open source version connected with a variety of other open source tools to perform APM, SIEM, or other observability use cases. Because the logging platform is so good at it. Some may self manage or host and some may use free or paid licenses. Those free ones are also customers that Elastic is trying to win over to their paid versions with their ever increasing collection of out of the box solutions from the source. And also for customers that may use paid licensed for one use case to simply expand into other use cases, expanding their existing deployments. A lot of opportunity and a lot of challenge.

Both Elastic and DataDog are investing heavily in R&D to expand use of the product. Non-GaaP

ESTC:

R&D expense in Q1 was $29.4 million, up 76% year over year, representing 33% of total revenue.

DDOG:

Research and development expense increased by $14.5 million, or 103%, for the three months ended September 30, 2019. 30% of revenues

5 Likes