Marketing Memos

I had the good fortune to sit in on a presentation by MailChannels. Their evangelist and old time friend of Mine, David Greer, presented both the problem and the potential solution of email spam.

In David’s mind, spam has become warfare. As in warfare, tactics on both sides are ever changing. Today the spammers are winning:

* Estimated 100,000,000 zombie PCs in botnets ready to send spam

* 8,000,000 active zombies at any moment

* 87% of all email traffic is spam

For years I laughed at legislation such as the CAN-SPAM act, which is as toothless as a wino. There are too many spammers, not enough trained cops, and a labyrinth of international regulatory, legal, and jurisdictional obstacles to ever be effective.

For as many years I have said that the only way spam will be reduced is if we can lower the profitability of sending spam. Only when it becomes tough to make a buck sending spam will it start to abate.

Here is where MailChannels comes in. They are in the crowded server-side spam filtering space. They provide a mail transfer agent (MTA) proxy that doesn’t delete any spam at all. But what it does is slow-to-a-crawl any suspected spamming sender, causing the spammer to give up on sending their solicitous email.

Using a number of configurable triggers, their Traffic Control product will hold open a socket connection from a suspected spammer. Since spammers make money by sending millions and millions of emails, they cannot wait on sluggish receiving MTAs, and these spammers will drop connections to slowpoke recipients. Hence the spam does not get through, and the spammer loses money via failed sends. Over time the results of spamming campaigns is reduced, dropping sales of spamming services and margins.

Needless to say more than a few companies would have to implement Traffic Control. But if just a few of the top email hubs (AOL, Comcast, Road Runner, etc.) were to do so, we might well see a number of spammers leave the market and some sense of sanity return to our work lives. Since MailChannels tested and discovered that 78% of spammers drop their connections within 60 seconds, we can estimate how many will go out of business once Traffic Control gets traction.

Finally, through measuring the number and types of publicly accessible MTAs in operation over a period of time, Mail Channels discovered that there is a 20% churn in MTA technology. This is strong evidence that system and email administrators are struggling with the problem daily. Given that other reports claim that spam costs companies about $500/employee/year, it is little wonder — and even more pressure to finish off spammers as soon as possible.

Bookmark to:

Tell a software developer that the real money has always been in services, and you beg being punched in the nose.  Those software types are a violent bunch.

The truth is that services have always been the largest profit center in the software world.  We may faint when presented with a software licensing invoice, but happily scribble our signatures on the adjoining support services contract that will cost 15% of the license cost every year, knowing that in technical support hours we will never come close to regaining the cost.

Services — regardless of if it is simple technical support or an implementation master plan from IBM Global Services — are products with value for the customer and even more value for the vendor.  Service margins tend to be fat, and the selling of services requires almost no additional effort once the product is installed.  This is nothing new, and has driven the software business since ones and zeros were invented.  Heck, Computer Associates made a fine living from acquiring existing products and the associated installed base, and milking the support revenue alone.

I bring this up because we have witnessed some recent and otherwise seemingly bizarre examples of major vendors dancing the Service Shuffle.  First came Oracle proffering Red Hat support service as discount pricing.  Then HP offered support for Sun’s Solaris.  And everyone seems to be offering support for MySQL.

There are two market(ing) dynamics at play.  First is that people will pay good money … or, in a pinch, American currency … for the insurance policy that is technical support.  When you risk the very operation of your business on a piece of software that cannot instantly be replaced by a competing product, a 15% annual fee is cheap insurance.

The other dynamic is the desire to have "one throat to choke" when it comes to support.  This is why HP can charge for supporting Solaris running on HP computers, because HP customers would like to have just one vendor for the support of their preferred hardware (HP) and their required software (Solaris).  It is not a bad strategy either in as much as Sun makes nothing on Solaris, and having HP take hardware sales from Sun adds to Sun’s market miseries.

As for Open Source, service has always been the economic driver given that charging for software licenses ranges from illegal to non-competitive.

These market dynamics do not adequately explain Software as a Service (SaaS), which breaks traditional tech support models.  Customers tend to view it as an ongoing, volume-based support contract without any of that messy software licensing business.  Vendors also view it as such, and also providing the added service of hosting the software and removing that technical tedium from the customer.

The take away here (yes, you knew there would be one) is that services have always been, and will likely always be, part of your revenue model.  This should not be an after thought, but should be part of your go-to-market strategy.  Heaven help the start-up pitching to a venture capitalist who has not included services revenues as part of their business plan.

Bookmark to:

I do love playing technology markets.  At times picking future technology survivors and corpses requires less brainpower than ever seeped from Paris Hilton’s alleged mind.

But right now, I wouldn’t handicap the virtualization market on a plugged nickel bet.  In this week alone, there has been enough news and intrigue to rattle Robert Hanssen’s cage.

First came word that Intel, and Microsoft’s bride Novell, had concocted a way to virtualized Windows Server on top of Xen.  Between them they developed some drivers that bridged Xen’s modus operandi to present it as real, live hardware to Windows.  In previous announcements Microsoft made it painfully clear that they were positioning their Virtual Server as being reluctantly tolerate of a guest Linux OS, but not for Windows to be a guest under anything else.

Apparently Microsoft relented, for they and Novell announced that Windows would eventually become paravirtualized on top of Xen.  This requires modification of Windows itself, showing that Microsoft has bowed to certain market inevitabilities. I say this because Microsoft never cheapens their recurrent "we control everything" stance until the market slaps them up side the corporate head.

All this is happening as VirtualIron dumps some Open Source alternatives into the mix to keep up with Xen Source, and VMWare starts to exit the hypervisor market while seeking an IPO.

So who, if anyone, will win at this game?  I can’t say.  But some things are becoming obvious:

1) Virtualization — at least the hypervisor part — is becoming commoditized, and will likely have Xen as its core.  It is not unthinkable that due to  Novell connections that Microsoft may well sell a Xen-based virtualization package in the future … one tuned to support Windows,
naturally.

2) With commoditization, most virtualization vendors will exit the hypervisor market, switching their attention to virtual machine management suites.  Insiders tell me that VMWare is doing this as a strategic directions, and VirtualIron has done so already.

3) The real winners are not the virtualization companies — they will fight one another for differentiation, that will become less and less available over time.  Companies that offer other products that can be optimized for dynamic and virtualized environments will thrive.  These include DBMS, systems administration, network monitoring, and identity management.  These product groups have untapped opportunities with customers who have densely virtualized servers.

Strap in boys and girls.  This will be a wild ride.

Bookmark to:

I sit at an odd intersection of the Linux tsunami.  Not only was I deeply involved in making SuSE Linux a viable brand within enterprises before the Novell acquisition, I also work in Alameda, California (a.k.a. Silicon Island), home of Wind River.

Wind River was and is the acknowledged leader in embedded operating systems.  For years their proprietary kernel was to be found secretly snuggled within automobiles, routers, satellites, and some military projects in the wink-wink, nudge-nudge category.

When Linux started to chisel away at their lead, Wind River did what a nimble firm should do — they changed the rules and became the predominant force in embedded Linux as well as their own operating system, and started bridging their world class development tools for embedded Linux bigots. 

I tell these tales to prelude a point:  Linux is a virus, and is spreading.  We know it has been a hit on servers, and depending on what metric you use, it is either a huge threat, an indomitable force, or the world champ.  But because it runs everywhere, it is getting into everything. 
As it did with servers, the spread of the virus begins with isolated infections, then rockets into a pandemic once it reaches large populations.

Two recent Linux news items are worth noting.  First, the mobile phone industry is strongly embracing Linux.  When we see Motorola, NEC, Panasonic, Samsung Electronics, NTT DoCoMo and Vodafone band together to standardize Linux for handies, we know the Linux virus has reached a crossroads from where it can infect the population at large.  Since the cell phone business is highly competitive, with margins thinner than Kate Moss, such a cable is interesting — as is any time competitors band together.

Second, we see
Peugeot adopting Linux for upwards of 20,000 desktops.  Until recently, mainly government and Linux focused technology firms had mandated Linux for desk workers.  But now we see a major commitment by a corporation in traditional manufacturing taking the plunge.  Being hard nosed mechanics, they likely take such decisions with serious deliberation, especially since 20,000 desktops covers 22% of all their workers … most of which don’t sit at desks.

Some may write these news stories off as anecdotal.  But so did others with early reports of Linux being adopted for servers.  I recall the stories well, with reporters and pundits claiming that such Linux adopting firms were  “bleeding edge” or “taking calculated risks.”  Perhaps they were, but it paid off because the economics of Open made ROI workable.  So too will be Linux cell phones and desktops, for when the ROI is proven in one
instance, eager competitors will mimic the success.

Bookmark to: