hortonworks deep dive

PaulWBryant posted that hortonworks might be a good investment right now. http://discussion.fool.com/hortonworks-opportunity-33020461.aspx… . I must confess that HDP is a stock I briefly owned but never got around to researching it so I figured I should sell it. Bear (paulwbryant) sparked my interest so I figured I’d take a look.

So I actually had a bit of a hard time figuring out what hortonworks does. I get tech, I have been peripherally involved in “tech” since I was in highschool. Funny story. I typed in, “what does horton…” into google. And google autocompleted “what does hortonworks do”. So apparently I’m not the only one. Basically hortonworks is a major contributor to Hadoop which is opensource. Similar to how redhat supported and serviced a linux distro (fedora was a spinoff if i remember correctly) . Hadoop is an open source software program (with major support by HDP at this point) that allows a large groups of computers to work on huge datasets relatively easily. There are a bunch of surrounding programs that do things like machine learning, data transfer, nonrelational databases, etc etc, many that are listed as hadoop competitors but HDP supports and services integrating them.

HDP ipo was at the end of 2014, with a marketcap of around 1 billion dollars and a Price to sales ratio around 15. They now have a market cap of 1.6 billion with a p/s of 5. HDP did a secondary share offering in jan 2016 because operating losses in 2015 and 2016 got to the point where they needed money. At the same time they purchased what would become Apache NiFi (from a NSA tech transfer program). Investors in the IPO got kind of screwed with this stock offering. Looks like they diluted shareholders at least 10%(but probably a fair bit more) and they were only able to complete the share offering at 9.5 dollars or less than 40% of the IPO price of 24 dollars. Mr. Market was very unhappy with HDP. Now their stock price is up from 7.5 to 20.5 in a little over a year with their p/s increasing from 2.5 to 5. Seems like Mr. Market is less unhappy with HDP but still very wary. So now, with all that out of the way is HDP a good investment now?

MarketCap 1.57 billion

Dollar Revenue Expansion Rate (something to watch)

2015 159%
2016 131
2017 122

Revenue growth% % subscription
2015 121 64%
2016 183 49% 69%
2017 261 43% 76%

Customers (increasing nicely)
2015 800
2016 1000
2017 1300

Operating cash flow (is free cash flow not including capital expenditures)
2015 -99.3
2016 -82.4
2017 -29.8 million
Q4 2017 was +6.4 million
They are guiding to be operating cash flow positive for 2018

Cost of revenue
2015 45%
2016 39
2017 31

CAC (including stock options, I know this isn’t ideal )Sales and Marketing costs/new customers. A good portion of this is stock grants to their sales team.

2016 181million/200 = 905k per customer
2017 200million/300 new customers = 666k per customer

Gross margin is increasing
2016 68%
2017 74%

I’m going to do some guesswork here but if we figure HDPs churn is very very low since it should be for most enterprise business. Using a gross margin of 74%, average account value of 204k (from their call) we get a lifetime value of a customer of around 7 million. We should apply some sort of discount and some sort of time limit on that. Since this is a guess lets say that number could be anywhere from 2 million to 5 million. A good LTV/CAC for a SaaS business is 3:1. 2/0.66 is 3. These numbers can drastically change depending on churn rate guesses, how long you keep the customer etc. I wouldn’t put too much stock in them. You can see if they spend 600k acquiring a customer and they are growing really quickly then they are going to continue to show negative cash flows for a while since they don’t recognize all the revenue for a customer up front. They get that money over years.

So what do i think? I think that HDPs numbers show that things are improving. They are starting to reach a size where they are self funding and probably won’t need to raise additional capital unless they decide to acquire something else. The person at the helm during the bad years is gone and things seem to have materially improved since then. HDP doesn’t sell software and i think there is some uncertainty out there if they have a good business model since everyone has access to the same software. In fact many business roll their own hadoop and support it themselves. I’d guess that is why both cloudera and HDP carry relatively low p/s ratios, 5 and 6 respectively. One additional point. I think HDP went public too early. They should have waited until they weren’t so cash flow negative. I wonder if their initial management didn’t have experience running what is basically a SaaS company Anywho I digress. People will have to decide that HDP’s business model is good before they catch up to other cloud (SaaS) companies P/S ratio. That will have to come with continued operational performance.


P.s if anyone catches math mistakes or thought mistakes please let me know. I’m doing my best but have much to learn.


Mr. Market was very unhappy with HDP. Now their stock price is up from 7.5 to 20.5 in a little over a year with their p/s increasing from 2.5 to 5. Seems like Mr. Market is less unhappy with HDP but still very wary.

This is why I sold out. I haven’t looked at HDP in a while. I don’t trust that their won’t be more dilution and excessive SBC. All the growth in the world can be counteracted by excessive dilution.

I prefer TLND as my big data bet. Not much dilution with great growth.



Great deep dive, Ethan. Just in case anybody missed it in your numbers, that cost of revenue decrease is tremendous. Here’s the effect on GAAP gross margin the last 8 quarters:

Dec15: 58.05%
Mar16: 60.44%
Jun16: 58.61%
Sep16: 58.39%
Dec16: 65.40%
Mar17: 68.10%
Jun17: 66.90%
Sep17: 69.07%
Dec17: 71.43%

Now that’s improvement! There’s an obvious reason for this too – their high-margin subscription revenue is growing SUPER FAST, and the boring low-margin professional services revenue is not. (Btw, I’m guessing your 74% gross margin number was non-GAAP? Still, they are marching toward that even on a GAAP basis!)

Also the deferred revenue was up from 185M (Dec16) to 275M (Dec17) – so it increased even faster than recognized revenue! And that’s just a large number.

I think I agree with you, they IPO’d too early – they were just losing SOOOO much money. It appeared (I think Saul even said this) that they would never be profitable. But lo and behold – they were cash flow positive last quarter. They’re just improving so fast all around – especially in trimming the losses. This is why I’m not so worried about future dilution. It will happen, to some extent – SBC is high. But the stock is still only at a market cap of just over 1.4 billion…whereas the company is becoming more and more valuable every quarter.



Here is the latest report from Bert that I could find. July 17


In summary, the investment is questionable due to management change up, slowing rev growth.

Final comments,

“It is hard to give advice in an information vacuum-I will not be taking action in our portfolio regarding HDP without more data as to what is wrong and the likely duration for a cure. Overall, given that the shares are up more than 55% so far this year, even based on the after-hours pullback, I am not inclined to hang around this name absent some far more reassuring commentary than is currently available.”