MDB Revenue Breakdown

I realize I am late to the MDB earnings party. I read the posts immediately discussing the earnings report and like most I was very much satisfied. But all the IPO talk caused MDB and Atlas to keep bouncing back into my head, as Atlas almost seems like an IPO in and of itself. So I decided to revisit MDB and parse out the specific growth pattern for Atlas by breaking the last 6 earning reports apart. I used 605 numbers because it was easier to get them going back 6 quarters and there seemed to be only a minor impact switching between 605/606 on Atlas revenue. I only went back 6 quarters because Atlas starts to get very small any further back. Breaking down MDB revenue looks like this:


                Q1     Q2     Q3     Q4
Revenue
FY18                          41.5   45
FY19            48.2   57.5   65     83.1
Atlas
FY18                          3.3    5
FY19            6.7    10.4   14.3   28.3
Non-Atlas Sub
FY18                          38.2   40                 
FY19            37.9   42.5   50.7   49.9
Services
FY18                          3.6    3.2
FY19            3.7    4.6    4.9    4.9

The biggest thing that jumps out to me is that Atlas has finally moved beyond “growing from a tiny base” insignificance and has become the primary means of revenue growth. Non-Atlas Sub growth was 25% yoy and essentially flat-lined between Q3 and Q4. Atlas grew by 98% Q3 to Q4 and 467% yoy. More significantly, Atlas is finally big enough to have an outsize impact on overall revenue numbers. Breaking apart the revenue numbers clearly shows that Atlas was the driving force behind revenue growth accelerating from 56.6% yoy in Q3 to 84.6% in Q4 as Non-Atlas growth actually decelerated. Now that Atlas has established a larger revenue share, the impact of Atlas growth rate on overall revenue growth rate is only going to increase.

Taking Saul’s advice of ignoring MDB’s earning forecasts I simply ran the numbers for FY20 assuming they maintain the same growth rates Q to Q to see what would happen. For simplicity, I also assumed Non-Atlas Sub and Services revenue remain flat. Using those assumptions, projected revenue looks like this:


                Q1     Q2     Q3     Q4
Revenue
FY18                          41.5   45
FY19            48.2   57.5   65     83.1
FY20            90.4   111.3  133.4  212.6
Atlas
FY18                          3.3    5
FY19            6.7    10.4   14.3   28.3
FY20            37.9   58.8   80.9   160.1
Non-Atlas Sub
FY18                          38.2   40                 
FY19            37.9   42.5   50.7   49.9
FY20            47.5   47.5   47.5   47.5
Services
FY18                          3.6    3.2
FY19            3.7    4.6    4.9    4.9
FY20            5      5      5      5

That produces a FY20 annual revenue of 547.7m and a further accelerating growth rate of 115% yoy. Is that achievable or absurd? I think it is more likely than the current laughable company/analyst guidance of 370m. Even if Atlas growth slows dramatically (let’s say the sequential Q to Q growth cuts in half), revenue will clear 410m with an Atlas growth rate of 167% yoy. So bracketing a dramatic slowdown (on the low end) to maintain current growth rate (on the high end) would be between 410-547m. If TAM is truly close to 80b then there is plenty of room to run here. Others would know better than I do. The nearly 100% Q3 to Q4 growth tells me that Atlas is not approaching a plateau and maintaining growth rate is certainly in the realm of the possible.

Anyways, enough for the late night thought experiment. I will sleep soundly long MDB.

Joe

49 Likes

This Q Atlas growth has been helped by the MLAB acquisition. So, better to model the non MLAB growth and then add back the MLAB acquisition effect.

Those are pretty amazing figures Joe, and food for thought. Thanks for posting them. Even allowing for part of the big jump in Atlas coming from the acquisition it’s still impressive. You bracketed between $410 million and $548 million. Even taking just 29% of the difference, up to $450, more than $100 million below your top estimate (the top estimate does seem unrealistic), but even taking the $450 estimate would give revenue up 77% (using ASC 605). That would be a significant further acceleration (which would be due to Atlas being a larger part of total revenue, of course).
Saul

14 Likes

Atlas revenue was around $5M. Doesn’t really change the story too much.

I find your numbers a bit too optimistic but Atlas “IS” the investment thesis. I bought MDB once I saw the Atlas results.

Denny Schlesinger

1 Like

From Q1’19-Q3’19, Atlas increased 3.7mm & 3.9mm quarterly. If we start with the 28mm Q4 and tack on 4mm quarterly for FY’20 and keep the rest of your assumptions, we basically end up right around managements guidance of 370mm. This seems to be the ‘conservative’ estimate, but I think that’s much more realistic than the high end scenario put forth. We should all be so lucky that that high scenario comes to be, but it seems a bit too rosy to me.

Marc
Long MDB

5 Likes

Do we know what the products that make-up the Non-Atlas Sub revenue are ---- realizing that this is NOT part of the MDB growth story?

If you back out the $5M from Q4 from MLAB you get only $113.6M ((23.3/5*23.3)+5) for Q4 2020 in Atlas rev instead of the projected 160.1M. Also, Q1, Q2, Q3 Atlas rev. Will go down by the same proportion as he has used the Q4 19 Atlas growth rates for those as well. The net reduction for the 2020 will be quite substantial.

1 Like

Do we know what the products that make-up the Non-Atlas Sub revenue are

Mostly licenses to use the non-free (premium) parts of the database.

Denny Schlesinger

Freemium is a combination of the words “free” and “premium” used to describe a business model that offers both free and premium services. The freemium business model works by offering simple and basic services for free for the user to try and more advanced or additional features at a premium.

https://www.investopedia.com/terms/f/freemium.asp

1 Like

Got it, thank you Denny. That helps.

All, I get that non-Atlas was essentially flat from Q3 to Q4, but do we really see those non-free/non-Atlas products truly just being flat? What about all those customers that choose to use Mongo in a hybrid non hosted platform? I imagine that those users continue to utilize more non-Atlas related services. Remember, even AWS is starting to build local clouds/hybrid clouds offerings. Seems like non-Atlas should be more than just flat ---- but perhaps not quite the 400%+ Atlas YoY growth rates.

Secondly, do we view all the Products on their Product page to be part of the Atlas revenue stream ---- or the non-Atlas revenue stream? Those include MongoDB Stitch, MongoDB Charts, Cloud Migration (sounds like a product, not just a service), MongoDB server, MongoDB mobile, MongoDB compass, Ops Manager, MongoDB Charts, MongoDB Connector for BI, MongoDB Connector for Spark.

Seems like non-Atlas should be more than just flat ---- but perhaps not quite the 400%+ Atlas YoY growth rates.

All I can say is that without Atlas I would not touch MDB. Great database but Freemium is not a good business model for investors. There are better places for my money.

Denny Schlesinger

1 Like

Hi jsammur,

Thanks for the post and interesting food for thought.

However, please double check the math on your “Non-Atlas Sub” number as in some quarters you are simply calculating Total Revenue LESS Atlas but in other quarters you are calculating Total Revenue LESS Atlas LESS Services. If this is truly a subscription number less Atlas it needs to consistently match the second calculation. See below example of the variance.

Q3’18:
$38.2M Non-Atlas Sub
$3.3M Atlas
$41.5M Total Rev
*Since A+B=C, the $3.6M Services must be included in your Non-Atlas Sub in order to tie

Q4’19:
$49.9M Non-Atlas Sub
$28.3M Atlas
$4.9M Services
$83.1M Total Rev
*In this quarter the Services must NOT be included in your Non-Atlas Sub in order to tie

This causes a rather meaningful impact in the subsequent analysis. In particular, Non-Atlas Sub was not flat Q/Q in Q4’19 but was up ~$4M Q/Q and +37% Y/Y which has been consistent with the last several quarters of Non-Atlas growth. I don’t doubt that Non-Atlas revenue will slow over time as Atlas scales, but given the TAM I believe Non-Atlas still needs to grow at +25-30% Y/Y for some time.

Regards,

Erik

2 Likes

Also to echo the sentiment of the replies, I don’t believe Atlas revenue will be even half the Q4 FY20 amount you have extrapolated but I still see MDB as a great investment despite this. The actual historical growth numbers are awesome, for sure. Q4’19 was a blowout, no doubt. But when beginning from a small base you should much more heavily use the $'s to extrapolate growth rather than %'s.

470% Y/Y growth is eye catching and wonderful, but it is off a $5M base and includes ~$6M of mLab acquisition benefit with no prior year compare. So, without mLab they sold ~$17M more Q4 this year than last year. That is great! I own a lot of MDB, I am not sneezing at it. But to think next Q4 they will sell $130M+ more Y/Y is crazy imo.

This may be rather contrarian and scare some on the board (please run your numbers!)… but I hesitate to think Atlas will grow at even much more than 100%+ Y/Y in Q4’20. If they do exceed it will not be by much. They have to grow ~$8M Q/Q for the next four quarters to grow 100%+ Y/Y in Q4’20. To date the most they have grown Q/Q was last quarter and it was +$7M Q/Q w/o mLab. Management has indicated mLab will be flat to down as they transition over the pricing to standard Atlas pricing so it will not help Y/Y.

The good news is that with the legacy business growing 30%+ Y/Y, MDB overall can still grow ~50% Y/Y in Q4’20 with the Atlas scenario I outlined above.

5 Likes

All,

Thank you for the feedback and discussion on MDB revenue. It certainly is an optimistic forecast. Anytime you are projecting ridiculous growth numbers like 467% yoy it has to be categorized as “optimistic”. It was mainly a “what if” thought experiment for Atlas to see what maintaining current growth rates would look like.

I would like to particularly thank Texmex and Erik (FlyFisher22). My previous post used the straight Atlas numbers from the ERs without stripping out mLab. Removing mLab absolutely makes sense if you are trying to parse out the “organic” Atlas growth rate (which I was, so thanks Texmex!). There were also a few errors in my numbers, as I did not subtract out the Services revenue properly in some quarters (thanks Erik!). For full disclosure, I fixed the numbers and they look like this (I also stripped out mLab):


                Q1     Q2     Q3     Q4
Revenue
FY18                          41.5   45
FY19            48.2   57.5   65     83.1
Atlas
FY18                          3.3    5
FY19            6.7    10.4   14.3   23.3 (mLab removed)
Non-Atlas Sub
FY18                          34.6   36.9 (fixed errors in Q3 and Q4)                 
FY19            37.9   42.5   45.8   49.9 (fixed error in Q3)
mLab
FY18
FY19                                 5
Services
FY18                          3.6    3.2
FY19            3.7    4.6    4.9    4.9

This changes some things from my previous assessment of the revenue breakdown. As Erik pointed out, you can see that Non-Atlas Sub revenue does not flatten but maintains a ~35% yoy ramp. Stripping out mLab lowers the Atlas growth rate from 467% yoy to 367% yoy. Atlas is still becoming the driving force behind revenue growth rate but not to the same dramatic extent as in my first post. Although seeing the non-Atlas growth maintain 35% can only be a good thing.

So what happens in FY20 if Atlas maintains 367% yoy, non-Atlas maintains 35% yoy, mLab disappears and Services remain flat?


                Q1     Q2     Q3     Q4
Revenue
FY18                          41.5   45
FY19            48.2   57.5   65     83.1
FY20            87.4   110.8  133.4  181
Atlas
FY18                          3.3    5
FY19            6.7    10.4   14.3   23.3 (mLab removed)
FY20            31.2   48.4   66.6   108.5
Non-Atlas Sub
FY18                          34.6   36.9                 
FY19            37.9   42.5   45.8   49.9 
FY20            51.2   57.4   61.8   67.5
mLab
FY18
FY19                                 5
Services
FY18                          3.6    3.2
FY19            3.7    4.6    4.9    4.9
FY20            5      5      5      5

Using the lower Atlas growth rate and 35% non-Atlas growth still lands you at 512m for FY20 and an acceleration to 101% yoy revenue growth. Of course this assumes Atlas maintains exponential growth. If you assume linear growth for Atlas as some have mentioned, it looks more like this (I assumed linear 7.1 Atlas growth for Q1-Q3 and a “blowout” Q4 of 12m growth):


                Q1     Q2     Q3     Q4
Revenue
FY18                          41.5   45
FY19            48.2   57.5   65     83.1
FY20            86.6   99.9   111.4  129.1
Atlas
FY18                          3.3    5
FY19            6.7    10.4   14.3   23.3 (mLab removed)
FY20            30.4   37.5   44.6   56.6
Non-Atlas Sub
FY18                          34.6   36.9                 
FY19            37.9   42.5   45.8   49.9 
FY20            51.2   57.4   61.8   67.5
mLab
FY18
FY19                                 5
Services
FY18                          3.6    3.2
FY19            3.7    4.6    4.9    4.9
FY20            5      5      5      5

This linear growth rate lands you at 427m for FY20 and a 68% yoy growth rate. Nothing to sneeze at but a bit of a deceleration from the current growth rate. This also assumes a massive deceleration of Atlas revenue from 367% yoy down to 100% yoy.

As several have pointed out, it may be too early to call Atlas growth exponential or linear. Both numbers look good and will substantially beat current guidance, but the linear number will start to die off pretty quickly in FY21.

The biggest lesson I learned from this thought experiment is this: If you have a thesis, expose it to the sharp minds of this board. They will find the weak points, lop off the errors, and hone it into something better. Still long MDB and interested to see what Q1 FY20 will bring. Thanks again!

Joe

42 Likes