Q. What do you call a sales-growing letdown? A. Pure Storage All-flash shipper Pure Storage delivered higher-than-expected revenue for its first fiscal 2017 quarter, but also had the first quarter-on-quarter revenue decline in its history since the IPO. Analysts thought it could have done better. Is this seasonality emerging for the first time or was there some other reason its turbo- …
Is Pure depending entirely on VMware's horrible software define storage offerings to boost their value or at least sustain them?
Hyper-V 2016 is entirely hyper converged and insanely fast. OpenStack has a dozen different high performance hyper-converged solutions.
What's their value point? Is their product intended for companies who would rather spend a million dollars more instead of two million less?
The anal-ysts are not trouncing Pure because of some lack of future vision, or that Hyper-V hyperconvergance is going to wipe away an enormous addressable market in primary block level storage. They are in a lather because Pure did not beat their own projections by as much as they did last quarter. Pure still beat projections, but the anal-ysts can find the turd in the punchbowl every time.
If Pure can continue to beat projections and gets bold enough to also raise guidance as some of their competitors have, they will do ok long term. However, woe be unto their shareholders if Pure ever misses their projections. One bad quarter can wipe out a lot of their valuation - a bit silly if you can take a longer view in that the company fundamentals don't change with any single quarter, good, bad or just marginal as this one was for them.
The take away for Pure leadership should be - when it comes to your earnings projections, do not over-promise and under-deliver. That only works for your marketing department.
Not sure if you're responding to my earlier question or not. I was honestly wondering what the benefit of Pure's systems are.
It's been about 5 years since any data center I've worked with would benefit from increasing storage performance in such a centralized fashion at insane costs per byte and more insane costs per IOP. This whole market simply died at almost the precise moment that Cisco bought Invicta.
Hyper-converged isn't new, I've been deploying it for years... though back handed since VMware isn't very good and we had to wait for vendors to stop locking into VMware. It's been years since I've built a rack that would operated suitably on such incredibly slow systems like Pure... I mean, it's fast for one or two servers, but for a rack filled with 88 core/6TB servers, it doesn't even have the bandwidth to serve one server let alone 12 in a rack. To run proper servers, hyper-converged is the only possible way as it requires local dedup and compression to overcome the bandwidth limitations of Ethernet and Infiniband.
So... who would have the money to buy Pure and simply waste it by buying Pure as opposed to building their data center properly so they can get more performance with less hardware and cost?
Besides stiff competition in AFA's, the growing server-SAN market will keep any of the smaller external array vendors from seeing long-term success. Get your money from PSTG if you can now, as having no long-term plan is not a good way to make profit.That goes for Nimble, Tegile, Tintri, and NTAP too. If you don't have a server-SAN package to offer the masses you are, IMHO, probably looking to acquire before you become irrelevant.
Looking at their costs, Pure has taken on a couple of high cost issues in the last year or two.
1. Going public - auditing isn't cheap... I'd expect as they grow this would start to consume $10s of millions per quarter. EMC on the other hand with the Dell merger is going to be saving on the order of $100s of million per quarter on doing away with auditing for being publicly traded.
2. R&D - The //m series and Flashblades are fully custom in house engineered products. Engineers, software developers, and scientists aren't cheap ...
3. Manufacturing - With only part of the above arrays being commodity components, the rest have to be manufactured by Pure... which again isn't cheap.
4. Sales - As much as sales people make, they also spend a TON. Starting small they spent an incredible amount of money on travel and expenses on their sales / SE teams moving them around their rather large GEOs. As they've grown they are now able to keep account reps and SEs in a more localized GEO, but depending on the performance of that GEO they may be making or losing money.
5. Marketing and swag - I have more pure swag around my cube than any other vendor I deal with. They give it away free, but it all costs money to make.
From a business standpoint, they need to spin off their losses into a different company or a non-profit subsidiary if possible. Wall street isn't interested in products, they just care about numbers. Being publicly traded requires financial engineering to make the books look as good as possible by any legal means necessary.
[caveat: I work for a company that positions and sells storage from multiple vendors]
Pure has an attractive sales model:
1. all flash
2. dedupe
3. buy it and we'll upgrade for free every 3 years.
4. executive engagement sale (the old EMC model)
Obviously, nothing is for free - you pay for it one way or another .
However, Pure is buying business, either by competing on price (which is hard to do against the big boys, who have deeper pockets) or greasing the channel - fat reseller margins and bonus spiffs for uncovering opps.
Thus, the more they sell, the more money they lose. This cannot continue forever and Pure - (and Nimble too) will either be bought, raise prices, or go out of business. There's no option D.
Worse, their time to market has evaporated: it's not a bad product - it works and does what they say it does, but EMC (with XtremIO/Unity), Dell (Compellent all flash), HP (3PAR) have all caught up and in some respects exceeded the functionality of Pure. In Dell's case - the performance they can deliver is at a significantly lower price and I know who I'm betting to be around 3 years from now. Nimble is delivering fair performance at lower price point than Pure too, for those that think they need all flash, but in reality don't.
"Worse, their time to market has evaporated: it's not a bad product - it works and does what they say it does, but EMC (with XtremIO/Unity)"
Unity doesn't have data reduction and won't until a promised bode update "later this year". Current pricing is not remotely competitive with Pure on Unity and XtremIO achieves real world data reduction numbers that aren't low enough to touch Pure for cost per GB, which makes sense given XtremIO isn't technically designed to compete in the mid-range AFA space. Sure XtremIO is faster at scale but the lets be realistic on how many enterprise shops actually care about the difference. Pure is selling deep into VNX/Unity territory today and EMC is giving them at least another 6-8 months before Unity 'might' be cost competitive.
I recently pressed EMC on price and suggested they should be selling me an all flash Unity config for the same price per GB as a Pure array without Unity data reduction if they want me to consider buying more gear from them. They sidestepped the question because they just can't do it and maintain a margin.
The other problem I see for EMC (and benefit to Pure) is the Dell merger. The shakeup in staff and possibly shuffling of sales teams between customers will create openings for Pure to sales teams to exploit. The uncertainty around the merger in general will also hand Pure (and competitors) customers who feel neglected or worried that Dell might decide to walk away from product direction (IE: Unity data reduction perhaps in favor of another array developed inhouse post merger? Who knows?).
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
