PSTG thoughts

There has been some really great discussion about PSTG over the past few days. Great information and analysis. Thanks to everyone who contributed to providing information and analysis.

At first I had serious reservations about investing in PSTG. The main reason was commoditization leading to ASP declines and margin erosion. While the revenue growth rates over the past few years look really outstanding, how could a company continue grow profits with declining price points. If you look at other companies that sell hardware you can that companies that continue to invest into better next generation products can maintain margins. NVDA is a great example. The cost per computing operation or per speed is rapidly declining. However, NVDA sells a new, better version every year enabling the company to maintain gross margins.

What will happen with storage. Seems that storage pricing will drop 50% over the next 3 years. Then over the following 3 years prices will drop by 50% again leading to a 75% drop over the next 6 years. That amounts to a decline of annual price erosion of slightly higher than 20% per year. I’m assuming that these figures represent pricing on a price per TB of storage.

So the key 2 questions are…

  1. How will PSTG’s cost decline? If the cost also declines at the same rate then margins can be maintained.

  2. How fast is the demand for storage growing? If companies won’t need more storage then PSTG’s revenue growth will be determined only by market share gains. If the market is growing by only 20% then PSTG’s growth rate will be about zero unless there are market share gains. If they market growth rate (we’re talking growth in TBs (terabytes) not dollars spent) is 40% per year then PSTG’s revenue growth rate contributed by increase percentage in TBs growth would be 20% (40%-20%).

So what are the answers to #1 and #2? Can PSTG maintain margins by COGS declines that need to match pricing declines? Will the market grow fast enough to make up for price drops. The answers to these questions will provide the answer whether PSTG can be a good investment or a not so good investment.



In my personal experience (which means little!), density rate of growth appeared to be following Moore’s Law - doubling every 2 years or so. That’s just from observing the consumer market where today I can get about double the size in storage for the same price 2 years ago.

I also reckon there has to be a butt load of sunk storage in the world. Data stored somewhere that’s probably never going to be accessed ever again. But given the rate of growth, this sunk data diminishes into insignificance as the years go on.

So here’s an actual study from 1986 - 2007.…

The study found that data storage grew 23% annually between 1986 and 2007.

General computing capacity from 1986 through 2007, grew at an annual rate of 58%, the report said.

So if that same rate continues going forward, it will at least match the decline in cost of storage.
That’s really an incredible growth rate in 31 years! But I guess that kind of answers question 2? PSTG will have to grow by taking market share, and that will happen whilst the switch to SSD occurs, which looks inevitable barring a new technology disrupting the transition. However, I noticed a source on the other thread indicating the market is growing at 40% annually.

Imagine you’re in charge of all this storage. And the amount you are responsible for is growing exponentially every year. Every year, you can expect a % of your drives to fail. You need to have inbuilt redundancies so customers don’t lose their data. The bigger you get, the more often you’ll have to replace your drives. Therefore costs will increase. But if storage density is also increasing every year, and thus the price per TB is decreasing, you’d then hope the costs will be keeping in line with the subscriptions and growth.

Companies store their data with PSTG not just because of the costs (well okay, everything is about costs as the price of developing and maintaining the expertise in-house would most likely be more expensive). They are paying for their data to be protected from disasters. They want/need the redundancies. I don’t see COGS of storage affecting their margins (I think the pricing will match it). It’s competition that will/could affect margins.


Interesting observation about stale data.

Very early in my career, before digital documents had overtaken the world, I read a study by a company that made microforms equipment. Microforms was (maybe still is) the collective term for microfiche, microfilm and similar. Their study revealed that once a closed document (“closed” means no longer subject to revision) was placed in a file cabinet it had a less than 3% chance of ever being removed except to physically move it to another location.

This resulted in massive amounts of wasted space and all the attendant costs of HVAC, underutilization, insurance, etc. The answer was micro-encode all those documents and store the microform in a much smaller space. There was not even a hint that possibly a better strategy might have been to purge all the unnecessary documents.

Nothing has changed, really. Where I worked near the end of my career I was the champion of building or buying a managed digital archive and establishing a disposition process for the stored documents. I had all kinds of arguments about how the added cost of the archive and associated management would be outweighed by the benefits, many of them legal. I even had corporate lawyers who would agree with me in writing. But, when it came to finding budget, no one organization would foot the bill. Not even legal. Engineering was looking at the possibility of funding an archive, but they had unrealistic expectations of what was achievable. There was nothing on the market and never would be that could fulfill their requirements. In a nutshell, they wanted to be able to pull documents from the archive years hence and be able to readily revise it with whatever the current technology was irrespective of the state of the technology at the time the document was archived. We couldn’t even migrate a digital design from one version of our design software to the next. The only way to keep those designs “active” was to retain with few exceptions virtually all the machines, os, dbms and application with which they were originally created.

Even when I left the company about 30% of the boxes of paper documents stored at Stone Mountain were orphaned. Nobody had any idea who was responsible for making a disposition decision, the default decision was to continue to pay for the storage - or pay to digitize it and pay less for storage (digitization is the new microform). The junk was never purged - at least not while I worked there.

Data is incredibly tenacious. Email has the half-life of granite. If you think deleting an email somehow erases every trace of it, you haven’t any idea about what companies do to make sure it can’t be disposed of by accident which pretty much guarantees that it can’t be disposed of on purpose as well (notwithstanding HRC’s missing emails).

Storage requirements will continue to grow for the foreseeable future, not just due to retention of stale data, but because there’s just a lot more of it to store and the growth is exponential.


“Stone Mountain”

Iron Mountain? That said, I agree with everything you wrote about data archiving.

Most likely “Iron Mountain.” When I closed down my business in California I had the legal obligation of keeping my accounting records for ten years. To play it safe I deposited them at Iron Mountain and paid ten year’s storage fees with instructions to dispose of them if they didn’t hear from me.

Denny Schlesinger…

Seagate - which is a HDD cigar butt play on spinning disk storage released some really strong numbers. If the market for storage is so strong that the legacy obsolete players still do well agains the flash headwinds, it only adds to my confidence in Pure’s market opportunity.



Yeah - that should have been Iron Mountain