Splunk

Investment Analysis Clubs Saul’s Investing Discussions
1

Apparently during analyst day (1/12), Splunk managment announced slightly lower guidance for next year than analysts were expecting. I’m sure they’re going to beat their guidance, but you know what happened next. The announcement drove the share price back down to the low 50’s…so I’ve been adding to SPLK. My position has gone from something I just dipped back into in December (just over 3% at year end) to now a top 5 position (surpassing HUBS, which I trimmed a little when I realized I didn’t understand the business very well).

Because this is now a top 5 position, I’m committed to understanding Splunk better. So I went to check out this infamous Analyst Day presentation. Well, the thing is 4 hours long. However, check this out: The 188 slides are linked to the video in a unique way. When you skip to the next slide, the video follows suit. I thought this was so cool I just had to share.

Click “Listen to Webcast.” http://investors.splunk.com/events.cfm

The financial portion of the presentation is about 3 hours in. If analysts didn’t like the trends for Splunk, then I just don’t think they were listening.

  • Splunk’s TAM is 55B.

  • Over the next three years, they expect to go from 12,700 customers to 20,000 customers.

  • Billings this year will be about 1.1B. 3 years from now it will be more than double: 2.3 billion.

  • Perpetual 25% / Subscription 75% by 3 years from now, and about a third of the subscription will be cloud. It takes a while to figure out what’s going on. Bert’s article helped, but I had trouble even following it. Basically they used to sell a license as a one-time revenue event. They had maintenance each year after that, but it was only a small amount compared to the upfront cost. They still offer software on this “perpetual” basis, but now they have subscription products too, which are just the opposite. The upfront costs are minimal, but the customer pays for the subscription annually. This is of course recurring revenue. Now I’m not sure why 2/3 of the subscription revenue will be non-cloud, or what that even means, but…baby steps. (Seriously, if anybody knows, please share.)

  • Three years from now operating margin will be 12-14% (it’s 6% now). After that, they expect cloud to hit an inflection point, where margins will really start to increase in a “non-linear” way.

So far from being concerned about growth because of analyst day guidance, I’m more excited than ever about Splunk.

Bear

9 Likes
2

Now I’m not sure why 2/3 of the subscription revenue will be non-cloud, or what that even means, but…baby steps.

Do not know much about Splunk. But can answer this question from work experience. Software licenses can be perpetual - one time upfront cost or subscription based. Usually it is an annual subscription. Software does not have to be offered as a service in the cloud. My old company heavily relied only on annual subscriptions and discouraged perpetual licenses by pricing them ridiculously high. This way there was predictability to revenue flows. It was not cloud based software, but a complex desktop based application. Hope this helps.

3

OK, I’m going to look at this from a skeptic’s POV:

  1. TAM (Total Addressable Market): I saw no indication of how they came up with the $55B number. For all I know they’re including any company with two or more computers.

  2. Customer growth: The slide “Customer Growth” (backed with more data on the previous slide) doesn’t show an acceleration in new customers. They’ve been flat for a few years at about 500 new customers/quarter, and before that flat about about 350 new customers/quarter.

  3. New Customer Growth Outlook: This is a slide later on. It shows them going from their current “Linear Growth” to “Accelerating Growth” and from “8 Quarter Stair Step” to “Faster Step Up” and from “~2K/year” to “~3K/year”

Now the kicker. The next slide “Customer Acquisition Drives Growth” shows them going from 12,700 customers today to 20,000 by 2020. At 500 customers/quarter, they’ll reach that in 3.5 years, which IS 2020/ So, there’s no accelerating growth here.

They expect the ASP to increase from $55K today to $80 in 2020, but that seems primarily driven by what they call “Mega Orders” (which are over $1million). That means they’re depending on a small set of customers to make big purchases.

Yeah, they’re going from a subscription model on the customer’s premises to a cloud model. But, the “Continued Bookings/Top Line Growth” slide shows the rate of increase will be decreasing. If you don’t understand onprem versus cloud, I could do another post - it’s something you should really understand. But, know that with either model, Splunk is essentially charging by the amount of data being processed - it’s not just buying some software and then you as a customer put it on as many machines as you want, each bit of growth costs you, so you’re actually incentivised to not use it more.

Splunk actually sums it up on the last slide “Path to $2B”, which shows 3 graphs:

  1. Customer growth staying at current rate, not increasing
  2. A more than doubling of “Mega Orders”
  3. ASP going from $50K to $80K. (ASP = Average Selling Price)

But, #3 includes #2. In other words, the ASP is increasing because they have more “Mega Orders”. So #2 only shows the risk they’re taking by betting on 160 new really large orders the next 3 years. To be fair, they also claim they will have increasing margins, but that presumes they continue to have pricing power.

Now, this isn’t a bad thing, but is it enough growth from a company this risky? We’re talking a company that hasn’t turned a profit, that is moving to new costs by providing a cloud solution (meaning they are absorbing the DevOps costs), and which is seeing companies like Microsoft (Azure), IBM, Oracle, and even Amazon compete. Being able to compete will mean more R&D, which means lower margins, not higher.

I’m not excited. What am I missing?

19 Likes
4

If you don’t understand onprem versus cloud, I could do another post - it’s something you should really understand.

I don’t want to take away from the Splunk discussion.
I, for one, would be interested in what further you have to say on the subject.
I wouldn’t think I’m alone either.

Hope to hear more from you.

Thanks,
A.J.

1 Like
5

Smorg,

Based on a couple of your questions, I’m guessing you just scanned the slides and didn’t actually listen to the presentation? I’ll give quick answers here, but I’d strongly recommend listening.

First big note is that the quarter ending January 2017 is fiscal Q417. Kind of weird, but really important. So the end of fiscal 2020 is only 3 years away. Makes a difference.

Anyway:

  1. I believe they said the TAM came from a data analysis…maybe external? I’m sure Investor Relations could point you to it if you’re interested.

  2. So?

  3. They won’t be at 3000 each of the next few years. They plan to get to 3000 in fiscal 2020. Remember, that’s only 3 years away, so 500 per quarter would only get them 6,000 in those three years and they’d be at 18,700 instead of 20,000.

But, #3 includes #2. In other words, the ASP is increasing because they have more “Mega Orders”. So #2 only shows the risk they’re taking by betting on 160 new really large orders the next 3 years.

Again, in the presentation they demonstrated that this is actually the case – as customers are with them for 3, 5, and more years, the customers spend more with them. Each year they have more and more customers who have been with them longer, so there’s no reason to think large orders won’t ramp along with the customers they are adding and retaining.

But, the “Continued Bookings/Top Line Growth” slide shows the rate of increase will be decreasing.

You’re referring to the infamous revenue growth for next year (fiscal 2018) which is guiding for “only” 27%. I mentioned that. I honestly think they’re just being really conservative and that they’ll beat it handily, but even if not, the reason I’m excited is that it sounds like they’re building a great network of long-term customers which are increasing their spend and finding new uses for Splunk all the time.

Bear

2 Likes
6

I believe they said the TAM came from a data analysis…maybe external?

Yeah, like I said.

…they’d be at 18,700 instead of 20,000.

So they go from 500/quarter to 600/quarter. For a company without profits yet, that’s a yawn, no?

…there’s no reason to think large orders won’t ramp along with the customers they are adding and retaining

OK, but nothing shows big growth, and nothing shows profitability on the near term horizon. They have costs switching from the onPrem to Hosted model. Seems like even they admit it’s at least 3 years before they hit that “inflection point” and a lot can happen in 3 years, like competition.

I honestly think they’re just being really conservative and that they’ll beat it handily, but even if not, the reason I’m excited is that it sounds like they’re building a great network of long-term customers which are increasing their spend and finding new uses for Splunk all the time.

Splunk’s a cool tool for some things - we played around with it at my previous company, but frankly the costs were simply far too high for us as a small company. As a customer, you end up pre-filtering your data since Splunk charges by how much data you put through it. And once you go through the trouble of pre-filtering you end just continuing along and doing most analysis outside of Splunk. If you’ve got some weird issue, you’ll fire it up, but for normal day to day analysis, Splunk is just too expensive.

Splunk is great at data mining, looking for trends or systemic problems that you haven’t seen before. Essentially, you slice and dice the data in different ways, looking graphically for anomalies or outliers. It probably does even more than that, but that was its main use for us.

BTW, a similar product at a different scale, Platfora, was recently acquired by Workday. And there’s also Tableau in a similar space.

4 Likes
7

Splunk is great at data mining, looking for trends or systemic problems that you haven’t seen before. Essentially, you slice and dice the data in different ways, looking graphically for anomalies or outliers. It probably does even more than that, but that was its main use for us.

Check this out – esp the chart under “the opportunity.” This addresses the TAM question too. One thing that interests me at present is the Security use case.

http://seekingalpha.com/article/4037882-splunk-path-2-billio…

The valuation bit might also interest you – don’t forget they are cash flow positive.

I really am not the least bit surprised you aren’t excited. I agree with you that the numbers don’t jump off the page. If you watch the video, you might feel differently.

Bear

1 Like
8

The valuation bit might also interest you – don’t forget they are cash flow positive.

That is true but 285 million of that is stock based compensation. If you take that away you are approximately 213 million in the red. Splunk is so dilutive I will be surprised if anyone makes any money on it except for the employees.

Andy

5 Likes