I have covered the “why?” angle enough I think, and I still can’t come to grips with any scenario by any party that makes obvious sense to me.  Thus, it probably is more relevant to think about the “what now” scenarios than the “why the hell…? ” questions.

If EMC wins the battle and buys Data Domain, here are positives:

1. They keep it out of the hands of arch-enemy NetApp (right or wrong is not the point).
2. They become the 800lb gorilla in whatever the dedupe backup target market is, and what it will be.  They will get first and last look, and that is good for them as they have a sales force that has historically proven successful when they can control the sales process.  As long as people are interested in buying dedupe (again, right or wrong is not the point), it will be impossible to bypass EMC in the process.
3. EMC will be able to market themselves as the king of all things dedupe, and as such will be able to position themselves as the “solution” provider regardless of the implementation and application.  Want it at the source?  Avamar.  Want it as a target?  Data Domain.  Want it from someone else?  Quantum, Falconstor, and lord knows how many others.  You want dedupe?  We have dedupe, baby.
4. EMC doesn’t do much in the pure lower midmarket outside of Dell, historically, and Data Domain boxes, for all their yapping to the contrary, are sold into environments where it could make for a compelling “surround” strategy.  EMC sits in the core with big Symms, NetApp is all over the place with their systems, and EMC could come in with a surrounding DD product set aimed at first stealing % of spend from NetApp, and second, trying to kill them by making a primary play at the NetApp market.
5. I can’t think of another overt positive other than they get to brag that they beat NetApp in the deal – unless the deal becomes an albatross, which is entirely possible (if not probable).

The Negatives for EMC winning;

1. Huge amount of dough to spend and very hard to show a positive return on the investment in any reasonable time frame in my estimation.  They do have the cash, but it’s still a huge bite.
2. Which one do I sell?  It’s already an issue, and this will compound it.  There will be much cloudiness before all see the light.  Unless you are a distributor, absolute focus is required for victory, and how that can happen here is not clear at all.
3. The confusion and chaos will make it relatively easy for others to FUD the giant, which is a lovely turn of events.  “Weren’t they just in here telling you to buy Avamar, or was it the Quantum box, or the FALC system?  Now they want you to buy this one.  I see.  Well, they sure are committed….”
4. This may get them downmarket, but it sure ain’t going to help them seal up the play at their core market: the high end.  The real competitive threat seems to be if/when IBM, HP, or HDS gets into the game with a product geared right at the Symm market as a means of disrupting the flow inside these critical accounts.  In this game, it seems the value/risk/reward is NOT at selling a million small boxes, it seems to me to be able to offer/protect the much fewer but massively critically profitable larger account deals.  This is what Data Domain and NetApp both covet – but this move doesn’t help EMC here.  What happens when old friend Dave Donatelli, who already has a Sepaton relationship, decides to really push them right at the heart of EMC?  EMC has no answer.  (Ironically, when I suggested to Dave several years ago that a play in this space at the high end would be pure accretive profits and an excellent defensive strategy to boot, he saw it differently.  Now, at HP, one would imagine he may reconsider that position – only from a server perspective of course.).  The high end is the battle ground for this space in the value acceleration game – if HDS or someone else gets into those big Symm accounts with a big iron solution to this very real problem, it will be a huge opportunity cost and open up the possibilities of legitimate competition in the installed base.  This is a giant risk as I see it, and one that can’t be addressed with any of EMC’s current lineup, including Data Domain.
5. EMC still sells core storage products based on CAPACITY – and this whole play is contra-capacity.  The reason DD had a business at all began with the simple fact that the PROBLEM being solved (i.e., one that someone will actually pay money for) was one of physics – there simply were no more minutes in the 24 hour day and as soon as it took 24 hours + to complete a backup, I have a problem – and it is time.  DD’s dedupe made the ECONOMICS of disk capacity work to enable a target that solved the real problem (time) without bankrupting the system.  Guess what?  Those economics are no longer the same.  Disk capacity is effectively free – thus, there is no real value in dedupe if the only problem you are attempting to solve is  TIME that you are given by backing up to disk.  Capacity is no longer a truly valid metric, because I can buy 100 TB of disk for $10,000 today – so why would I pay $80,000 (or more) for 100 TB of “virtual” capacity?  The buzz will leave this market soon – because if it’s purely economic, the math doesn’t make sense anymore.  Capacity is a stupid metric on either side of the equation at this point.  The “problem” that was originally solved no longer requires dedupe to solve it – thus there had best be another problem that needs to be solved.

So, that means the only reason to buy DD or any other dedupe device is either that having duplicate data (which I am not advocating is good) causes OTHER problems beyond the cost/capacity argument.  Too many boxes to manage, too many people to deal with too many boxes, etc. are completely legitimate problems.  Having said that, can I not solve those issues by A: not backing up duplicate data (process changes meet common sense) and B: throwing cheap disk at the problem?  Perhaps the real problem is that you simply can’t back up up enough fast enough even with disk – which is true in very large accounts.  That is a totally legitimate problem – but not one solved by Data Domain (at least not yet).  If the answer is that I need a ton of smaller Data Domain boxes, well, have I really solved any problem at all?  If it cost me 5X the money versus being dumb and cheap, but only gets me 2x the benefit, did I buy myself anything?  I’ll just buy cheapo raid arrays for 1/10th the cost of current Data Domain systems.  Worse yet, I’ll go back to tape.  Remembering what problem you are trying solve is often the first thing you overlook when things get excitable.

EMC sells most effectively to very large customers, but this move would not help them there.  Instead, it would spread the market to smaller customers which is against the core competency of the company.  This split personality syndrome could cause more problems internally than it solves.  It is NOT an easy fix, no matter what.

Positives for NetApp if they win:

1. While many have commented that they think EMC has more synergy with DD than NetApp, I disagree.  If they do get it, all numbers aside, they have the opportunity to shore up the midmarket storage space for primary and backup targeted devices.  They will inevitably bridge both worlds and, all said and done, will have the opportunity to own the landscape – for as long as there is value to be derived from that landscape.
2. NetApp proved that once the box is in, they can make lots of money selling add-on software function/features to that box – probably better than anyone in history.  They have maintained 63% gross margins on a product that at the conceptual level, is 20 years old.  Brilliant.  If they can do the same for the DD world, they will effectively eliminate others who want to grab the same footprint, and at least have the opportunity to repeat their previous success.  The key factor is total domination of the footprint.  They would need thousands upon thousands of new customers for the DD product set, and they need them really, really fast.
3. They have a chance, if they go full bore right away, to keep EMC and others off balance – and to dominate this landscape.
4. It will definitely (right or wrong) put NetApp on the top of the watch list, if for nothing other than beating EMC in this deal.  Of course, that will shine the spotlight on every move they make unlike any they have ever experienced before.  This is a BIG deal for NetApp.

Negatives for NetApp;

1. Everything I (and others) have already stated.
2. See 1.  This is super high-risk.  The deal went from $750M cash – roughly 20% and stock to what will end up being at least $2B all cash.  Will or can NetApp really afford to lose 50%+ of its operating cash?  If EMC is playing them, this is pure brilliance, as it will effectively mean NetApp won’t be able to do any other deals of substance, and thus will enable EMC to focus on a very specific target with a very specific set of moves.

Other Considerations:

1. EMC has opened up the door to let NetApp off the hook, which personally, I think is the best possible outcome for them.  I would rather keep EMC distracted with DD and use it against them. This is a free pass from a potential huge problem – there are other, potentially more significant possibilities that exist for them.  They can turn the tide on EMC and reverse the spotlight – and firmly put all the pressure of the move squarely in Hopkinton.  They can even say “we knew EMC would react like this, we are brilliant!”  This does NOT have to be about NetApp losing – quite the contrary.  Although I think reality is that ego will dominate the deal from this point forward, I think it’s a lot like two idiots fighting over a girl even after the girl left the dance.  They forgot what they were originally fighting about, and thus the potential benefits, but continue to beat the hell out of each other anyway.  Everyone loses.  Except the girl.
2. Is NetApp more or less attractive as an acquisition target with or without DD?  I suggest they are more attractive without DD and the expense.
3. HP is going to buy someone.  They will deal with the NAS space AND the dedupe space one way or another.  HP is so big that small deals don’t move the dial.  NetApp is big enough to move the dial, but not so big that it’s out of their buy range.  It would almost double HP’s storage business overnight.  In a consolidation era, size matters.
4. If EMC is playing NetApp to spend more money than the $1.5B, the move really is genius.  At $2B, it makes a huge dent in NetApp’s cash position and severely restricts it from making other moves.  It’s basically going to put them in an “all in” bet (for all you Texas Hold-em fans).
5. It’s interesting to see EMC reacting in this way.  I’ve never seen it before.  You would think the reason why they are doing it would be completely obvious to justify such a reaction, but I still don’t see it.  It can’t be defensive – can it?
6. How great is it right now to be Exagrid in the mid/lower end market, or Sepaton on the high end?  I don’t see how either does anything but get more valuable with all of this going on.  Exagrid is adding customers at roughly 25% of the rate of Data Domain, with 1/25th the $ and people. (I made that up, but it’s probably not that far off).  Sepaton doesn’t have the reach (without HP) but when they play at the high, high-end, they win.  I gotta believe HP wakes up and puts some wood behind that arrow, and it wouldn’t even surprise me that HP (but not Donatelli!!) goes after it directly and buys up Sepaton to hit EMC where it really hurts.  Actually, I like this idea, as it would also put direct value add on top of the HDS play that HDS would have no answer for, allowing HP to garner more leverage from the relationship.  NetApp, or EMC could probably snap up Exagrid for a few gazillion bucks less and be in just as good a position to drive value as they would with Data Domain, truth be told.  Once they focus on the real problem, the means of solving it are almost secondary.  The battle will be won by having a bunch of sales people violently waving their swords in the same direction – not because one sword is better than another. If someone takes Sepaton (or if IBM really shows up with Diligent) to the high-end peak of power, there is NO WAY that EMC can ignore it for long, as they will have their very way of life attacked.  I don’t think EMC can ever let that happen.
   6A.  Maybe EMC takes out DD AND Sepaton, and, why not, Exagrid too.  No way it fails the DOJ monopoly test with all the other alternatives, but in reality, it gives them the entire market landscape.  Still too much money, but that’s an interesting scenario.
7. I applaud Chuck Hollis for publishing all the comments to his rationalization of the deal, especially since they are all negative.  Having said that, where are the true believers in the deal for either EMC or NetApp?  I see lots of reasons why not, but very few reasons why.  I’m a dope, but can everyone be blind?  Sometimes the war becomes more important in the moment than the reason we fought to begin with.  My guess is ego will propel the deal from this point far more than logic or reason, which is unfortunate.
8. No matter who loses the fight – Data Domain wins.  So does dedupe and we’ll all be buying de-duped ice cream and pajamas just like we did with Sponge Bob.  The noise will compel the masses to start deduping everything, which is just fine for an industry desperately looking for any reason to be interesting.
9. Either of these guys would find buying Isilon for $150-200M more in line with what they really need, and either ignore or compete in the dedupe space as they currently do.  I see much more long term upside in having scale-out architectures that work universally regardless of the “application” of the target.

So, my takeaway after all that is that Dan W. should not react emotionally, and let EMC take it, because I don’t think either of them can make it a winner because the girl they are fighting over not only left the dance already, she was ugly.  No offense Frank, you just became prom queen, but for all the wrong reasons!

• Share