What’s In a Number?

When is the last time that anyone relevant paid attention – let alone based a buying decision – on ridiculous market share claims? IBM says it stole 1.2% of HP's storage business in Bolivia. Emulex/Q-Logic has 52.4% of the hotly contested Fibre Channel HBA in a server that sits next to a poster of Christie Brinkley in her day (she still looks great, and she's 87) vs. (Insert vendor here: Emulex/Q-Logic). Cisco has garnered 56% of the director market on Tuesdays following a snow storm – Brocade is vehemently denying this, and counters with the fact that Cisco is counting food service Ketchup/Catsup bottles twice. I made up all those statistics, but don't be surprised if you see them in a press release soon.

Who cares about all this? What possible value is it to anyone with a real job to listen to this stuff? It seems to me that the only folks who care are vendor employees who for some reason or another, continue to be measured in some way based on numbers derived from processes that are so fundamentally flawed that the only metric that matters is "do I get my bonus?". A peer review is one thing, but this has gone too far to be useful.

The only way any of this would be meaningful is if it were accurate and true – and as far as I can see, most of it is neither. Vendors don't report any numbers anymore, and those who count things use "models" – i.e. wild ass guesses based on mood or sponsorship. No one sits and counts physical shipments. The only valid way to do any of this is if you had a statistically significant sample of buyers and users that you monitored over time. Until then, it's just sort of silly.

Vendors use the numbers in three ways: 1. if any number makes them look good, they announce it to the world as proof positive that they are the king of whatever it is. They don't care if they get credited with being responsible for someone winning a medal at a beer festival or for discovering a new planet, nor do they care if the source is really a three legged dog in Mumbai with a parlor trick or two. If it's positive, it's true. 2. They have a Bolivian country manager that somehow gets paid if their number is positive – regardless of whether the overall company results just tanked worse than Waterworld, or 3. They use the inaccuracies and flawed methodologies to completely pooh-pooh the results as absurdly one sided and entirely useless. Those are the ones who didn't fare as well in the hyper-scientific analysis.

The fact is anyone can come up with any number anyone else wants. That's why it's stupid. No IT guy worth a dime will actually make a buying decision because "according to ABC, EMC has 98.5% of the worldwide NAS market" or "Cisco now has 87.4% of the world under its switching control". Any investor or hedge fund manager who puts any weight into these numbers will eventually be fired, and forced to live the remainder of their lives on a sweet Caribbean island counting their money. Tracking Microsoft's revenue by product in conjunction with total server shipments might yield some interesting results. Listening to numbers showing how Star Office is killing the Seattle folks won't.

Now, if the results were all true and accurate – and not wordsmithed to an outrageous extreme, they might be useful. I still don't think anyone would make a real decision to part with their money because of it – but it would at least help folks decide who gets a chance to play. I haven't met every single large shop in the world personally, but I refuse to believe that a Brocade/McData shop is going to dump what they have for Cisco just because Cisco is gaining share, or they are going to toss out Netapp because someone told them that EMC owns the entire landscape of Bakery supply NAS devices. Folks will apply some basic common sense most of the time. For example, there are only 2 real combatants in the Cisco/Brocade (McData) fibre channel switch/director war in the fortune 1000. Is it really reasonable to think that anyone would eliminate the only real competition because of some goofy report? Of course not. They would look at both, as the numbers clearly do signify that both are worth looking at. They would weigh each proposal on its own merits – and if the decision was made to really consider removing the incumbent – they would then go through the pain vs. pleasure analysis. At the end of the day, the buyer makes a choice – but I will lose my faith in humanity if that choice was made solely based on some report that showed that Vendor A was more Dolphin Friendly than Vendor B. Of the 467 things that should be considered on a major IT purchase, market share data should come in at about 468.

QLogic and Emulex enjoy a relative duopoly. Cisco has made huge advances into the core of the largest data centers with their director products – but Brocade still owns the overwhelming percentage of switch ports. People are buying more Brocade/McData than they did last year, the year before that, and so on. The McData acquisition questions have been answered – the sky is not falling, and conversely, looks pretty bright. Wisely, Cisco tried to capitalize on the fear that any major merger drags, and who can blame them? Regardless of the fuzzy numbers, the fact is that Cisco is also selling more than they did last year, and so on. EMC knows that in a pure NAS play they don't sell as much as Network Appliance, but why would they correct the perception? EMC sells more everything storage related than anyone else, not to mention a few fairly successful non-storage things – so who cares? IBM, that's who. Real people? They don't care.

It would be helpful if someone started counted things that mattered – like what works and what doesn't, and how painful it was to Yankee fans to see that the Red Sox are now America's team – according to statistics published by USA Today, which unlike all those shoddy research outfits, clearly is an unbiased, above reproach example of how things should be done…..

Comments

[Open Systems Guy]

Tech managers can have a herd mentality- I have seen shops choose one vendor over another precisely because that was the way the market was trending, and if others are doing it, there must be a reason for it.

Numbers shouldn't matter more than product, but sometimes they do.

[Joerg Hallbauer]

Another reason that vendors focus on these numbers is that if they are gaining share on their competition they want you to believe that it's proof that their product is better. Of course, this is patently ridiculous, but it happens all the time.

The only thing I use these numbers for is to see who are major players are in a market, and who are the little guys, and then include all the major players in my decision matrix.

[TM]

You're absolutely right here, Steve, but let's not forget you are a seasoned industry analyst and a technology purist; not so easily influenced by these Jedi mind games. But lots of people do indeed put stock in them - and as long as that's the case, vendors are going to make a big deal out of them. And as long as *any* vendors are doing so, the rest will try to keep up with the Joneses.

I think these single-digit share tug-of-wars between the 800 lb. gorillas offer little value and are rather boring. I think it’s interesting, though, to see a new guy on the block begin to take share – because that is, in my opinion, meaningful (and very much needed) validation that there is a real story here and that perhaps the game has changed.

Totally agree that the publishers of these market share reports should report the relevant facts rather than creating 300 different categories in order to ensure that every vendor has at least something to be happy about. It’s fairly comical: the day after a major report hits the street, every major storage vendor issues a release about its successes:

“HP Leads Storage Market” (units shipped)
“EMC Leads Storage Market” (capacity shipped)

“NetApp Leads NAS Market” (NAS, not including CAS)
“EMC Leads NAS Market” (NAS + CAS)

I guess Benjamin Disraeli had it right when he wrote, “There are three kinds of lies: lies, damned lies, and statistics.”!

-----Good point on the little guys, but even so, I think it's more impactful when a little guy talks about where they win and not rely on the numbers. As soon as NewCo takes any meaningful share, the big guys get to change the category. IDC counts EMC's Centerra in thier NAS numbers, which is completely wrong in my opinion, but what are you gonna do? -- Steve