Newsletter
Cloud Ain’t Easy, It’s Never Easy: The Future of Managed Services

© 2011 Service Leadership, Inc. All Rights Reserved

(Synopsis: The world of cloud computing is still full of unknowns, which translates into risk, and correspondingly, opportunity. What is known is that historically innovations in technology - many of which have been made in the name of simplicity and reduced cost - have been accompanied by increased support complexity, with the resulting need for IT services increasing. Evidence indicates the same holds true for cloud computing, which means we can describe a clear path for today's Managed Service Solution Providers to continue to grow and prosper.)

In the movie Cadillac Man, the main character (played by Robin Williams), says, for a bit of comic relief, "It ain't easy. It's never easy."

The current meme on cloud is: "It's easy."

Those of you who have been in the business for a few decades have heard "easy" before:

You get the idea. I personally lost track in the 2000's. I have not noticed that IT services revenue has gone down.

Who Wants "Easy"?

Strong constituencies want zero-support computing.

Customers want zero-support. Unless they have a spouse with a career in IT services/operations, no customer ever bought a new piece of software or hardware and said to themselves, "Oh boy, I get to spend more time and money with IT people."

Venture capitalists want zero-support, for two reasons. First, VCs know customers don't want to spend money on IT operations/support. So products which can be advertised as zero-support sell better.

Second, the majority of VCs do not like investing in people-based businesses. They fear business models where "the assets walk out the door at 5:00 pm." So, those seeking venture capital tend to promote their new IT offerings as "lights-out" operations. You can bet every would-be cloud vendor on a VC roadshow is trying mightily to give the money-men the impression that humans will need to be hired to build, yes, but not to sell, deliver and support their disruptive new cloud service.

In 2001, we successfully raised venture capital for a client to start a Managed Services business. As usual, we pitched about two dozen VCs around the U.S. The most common objection by far: "Your idea has a limited life span. Everybody knows five years from now IT won't need support." Have you noticed any technology-related drop in IT services demand since 2001? Me, neither.

IT Needs Support, No Matter Where It Is

The fact remains that information technology - being necessarily and insanely intricate not far under the surface - will forever remain complex to build, put into production, operate, maintain and upgrade. Wherever there is hardware and software, IT jobs are safe. The nature of the jobs may change - more on that in a moment - but the jobs will remain.

Another example of this lack of understanding of how IT operations work, is in how IT vendors look at different markets around the world. We often get the question from major IT vendors: "Will Managed Services be popular in (insert non-North America geography here)?" For the past five years, we've been giving Managed Services workshops to Solution Providers on every continent, in cultures and economies as diverse as India, Russia, Brazil and Germany. Solution Providers everywhere - whether they've ever heard of Managed Services or not - express the same problem: "How do I most effectively support IT and users with my costly resources? And make money?"

Vendors view localization in a clinical way: adapt to the local volts and hertz, translate the screens and manuals, and that's the end of the job. For some reason, they believe that what happens after that might somehow be different in that market from everywhere else.

IT operations come down to solid state physics and physical logistics: when the bit doesn't flip, there are only so many ways to fix the problem. The laws of physics and logistics are constant throughout the universe. Users are the same everywhere, too: when the user can't figure out how to get imperfect software (from their point of view) to do what they want, they ask for help. Imagine that.

Cloud Gazing

So let's look at cloud.

Everyone knows cloud has been around for a while: hosted Exchange and backup has been around for the better part of a decade. When we founded our own company eleven years ago - having done IT service for a long time - we decided we wanted no IT assets to operate and support. So we were early cloud adopters. What's been the experience?

While we decided on laptops and desktops with traditional Microsoft Office and Quickbooks Pro, everything else we got from the cloud:

First, the good news. As a startup, there's no question cloud gave us a multitude of capabilities on an easy, pay-as-you-go basis. Little infrastructure investment, and minimal infrastructure support costs from our chosen local Solution Provider. A relief, really, from the usual burden of IT ownership.

And, we've largely stuck to that model to the present time. I don't see us changing.

Low cost? No. Worry-free IT? No. Let's look at the reality.

Cloud Complexity

If you include the various ISPs we use, and the telcos and wireless companies associated with having employees in three states, we write checks to 26 different IT/telco providers each year; many of them we pay monthly. We try to watch them, but really:

Cloud means more transactions with more vendors. That's complexity, even before you get to the level of integrating applications or data across the cloud.

Cloud Cost

Cloud providers for the most part are not dumb. They know that their customers are transferring risk and cost to them, so they charge for it. Pay-as-you-go is the lowest leverage form of retail. A well-run, fully-scaled cloud offering earns about 70% gross margin. There is no question we are good customers for our cloud providers, from a gross margin point of view.

Could all of this be provided in-house for a lower cost? Probably not until we're bigger. But that doesn't diminish the size of the checks I write. And for the most part, it's linear: add an employee, add a unit of cost.

In addition, many end customers will find cloud applications to be more expensive than server-based alternatives. Why? Talk to application vendors (both ISV's and the major application houses): they're salivating over the move from the traditional "buy once, 18% annual maintenance and maybe I'll upgrade someday if you force me" to the cloud's "pay every month no matter what" model.

Cloud is not cheap.

Cloud Flexibility

Other than the joy of adding users easily, cloud applications are for the most part, not very flexible. The primary reason is common to all shared systems: the economy comes from sameness, not uniqueness. Everything about a cloud provider's operating economics drives them towards treating every customer the same. You get what you get. Will they eventually get better? Yes, but only slowly.

Lack of flexibility requires expertise to get around. Either the customer hires his own expertise, or he hires you. No cloud relief here; in fact, the cloud creates demand here.

Cloud Risk

This is the biggie. There are serious risks in going to cloud applications. They all center on "stickiness" to use the polite term.

Availability

The cloud fails, or the Internet fails. Then your business hurts.

Here's an example from our own recent experience. Subscribers to our Service Leadership Index® benchmark services know we use Intuit's cloud database offering, Quickbase, for much of our front-end data collection. Intuit is a publicly traded company with $3.5 billion in annual revenues, so a company that has the resources to deliver a solid platform.

We've used it for years, it works well enough, and until a fateful week late last year, availability was never in meaningful doubt.

Then, in the course of one week, during one of our customer data entry windows, Quickbase...disappeared. Gone from the Internet. Due to problems at Intuit's world-class billion-dollar datacenter, Quickbase disappeared for about four hours.

My staff was anxious. Calls and e-mails from confused subscribers started coming in: they'd set aside time to enter their data right now. Thankfully, Quickbase came back after four hours. We received a sincere and lengthy apology from Intuit's general manager of the Quickbase unit.

Two days later, still in our data entry window, Quickbase disappeared again. For twenty-five hours. Anxiety among my staff and my customers went up.

We received many earnest apologies from the Quickbase general manager during the 25 hours. When it was finally over, we received a four page email explaining the situation, why it occurred, what they were going to make sure it would never happen again, and compensating with some free service.

Now, there were several things wrong with this final "I personally apologize" e-mail:

Too little: a rough calculation shows the outage cost use around $7,200, if you assume an hourly labor cost for various staff whose productivity was impacted between 10% and 75%. So $2,600 was not only too little, and felt a bit gratuitous.We could not calculate the cost to our customers, but it was more than just patience.

Too much: Quickbase gave up one month's revenue across its entire customer base, let's call it 10% of their year's revenue ramp, for a day's interruption. This is a bad financial and legal precedent to set. Financially: If the next outage is two days, will they forgo two months of revenue? Legally: No doubt their EULA says they're responsible for no damages of any kind from an outage. I'm no attorney, but it seems their general manager would disagree.

The whole response came across as overwrought and naïve. The $2,600 did little to reassure me that cool heads were in control. Better to have said: "Sorry. You have my commitment it won't happen again. Please keep those checks coming in."

Consistency of Features, Capabilities and Interfaces

If I license an application and put it on my server, I decide when it gets upgraded. I can be an early or late adopter, as the situation requires. In the best cases, I can decide which features I want from that upgrade, and which I don't.

No doubt lagging upgrades bedevil many a software vendor's revenue budget. Plus, they have to support old versions, it probably seems forever.

In the cloud, these business problems are neatly resolved: the old code is retired immediately, everyone gets the new code, at the same time, included in their monthly fee. Won't they be happy!

Well, no.

Another first-hand example. We provide documentation for some of our online services, by linking to documents served securely by a four-year-old cloud provider called Scribd. It's a free service though we would be fine to pay. But the VC-approved cloud app playbook is, give it away and revenue will come. Judging by the plethora of documents posted by companies small and large, neither visibility nor traffic was a problem. Maybe the problem was revenue?

Over the course of a year or so, we integrated Scribd screens into our website and number of our cloud services. Until about eight days ago, everything worked great.

Then, unannounced, Scribd changed its technology and its business model. Why should they announce it to us mere users? After all, it's free? We should be thanking them!

Scribd changed from Adobe Flash to HTML 5. No doubt for good reasons. But half our integrations stopped working. No real reason HTML 5 should not work, but IT being touchy stuff, our customers started calling to report - what's the fashionable phrase right now? - their "user experience" wasn't the same. Those aren't the exact words they used: "I can't see the document," "the page is unreadable," words more like that.

Scribd changed from a no-revenue model to - you guessed it - an advertising-based model. Hey, it worked for Google. Suddenly, the nice, clean, business-like user interface our customers were seeing was chock-a-block with random ads. At the moment, the top banner ad shows three grisly mug shots along with an offer to run a background check on anyone I want, one side banner invites me to three free nights at a second rank Las Vegas "palazzo", and the other banner ad - refreshingly professional - invites me to try Cisco WebEx (full disclosure: of the three, thankfully Cisco is the only one who is our client). Eight days ago, when we discovered the change, the ads I recall were - I'm serious - for condoms and a political fund raiser. This is what our customers are seeing.

We called Scribd. Try getting to a live human in a modern, cloud-based business model. How we reached the very pleasant and helpful Director of Product is another story, but when we asked if we could pay for an ad-free version, his answer was, "We're thinking about that."

Fair enough, it's their company. But it's our - and our customers' - cloud experience.

Such abrupt feature changes are not limited to free cloud services. Cisco's WebEx, through whom we sell recorded and live webcasts, recently changed how it works with PayPal. At first it appeared the PayPal interface disappeared altogether, then a few days later it reappeared in a different and less functional form. While many of Quickbase's new features are welcome, virtually every time they have a major release, something breaks in our system.

Do these companies give us advanced warning? Sometimes yes, sometimes no. Do they allow us to test? No. Do they allow us to stay at N-1? No. Why not? It would hurt their business model.

Understandable, but these have material implications for our business. These are client-facing applications we're talking about. We have a development pipeline all planned out, our dev team hard at work. And we can't control when an upgrade will be applied, or even test our code in advance?

Another example is what we call "value deflation." For a few purposes, we've used SurveyMonkey for the last six or seven years. We've been at their highest fee level, to get the most features. A feature that has been in even the "free to try" level is, at the end of the survey, we can redirect the user to our own website (as opposed to the default of being redirected to SurveyMonkey's public site). It's a nice touch from a professional-look standpoint, and it probably drives a bit of traffic to our site.

The point isn't that these kinds of things won't eventually get resolved in the cloud world. The point is, a whole host (pun intended) of other things we haven't imagined will crop up. The nature of IT support changes, but the amount of it that's needed, just keeps growing. More software, more users, more devices, more connected: simple, it is not.

Customer Profile

The third risk you incur by going to the cloud, is that your profile as a customer, may over time diverge from the target customer profile of the cloud provider. Yes, this could happen with an on-premise solution - my local Solution provider might outgrow me or vice versa - but once I own it, I can run it for just about as long as I'd like.

In the example of Scribd, above, other business-oriented users likely will react to Scribd's new business model with the same unease we have: do we really want to send people to see our document on a website that advertises who-knows-what? Whether Scribd knows it or not, one suspects they have changed target customer profiles, and left me behind.

Another example. When we got our first database server in 2004, we put it into an outstanding colo facility in Dallas. They had just spent $20 million refurbishing a data center originally built for $60 million back in 1999. The original owners had gone broke when the Internet bubble burst, and the facility sat unused until new management was able to raise the venture capital to buy and restore it.

We said we needed just a few "U"s of space. Amazingly, they invited us down for a tour of the 200,000 square foot facility. More amazingly, it was the CEO who gave us the tour. Now I love data centers, and this one was big and beautiful, and I love CEOs, and this one was sharp, but how could our puny requirement merit such attention? To this day, we don't know, but we liked the $270 a month price.

Fast forward to six weeks ago. We're buying new servers, we need a few more "U"s. Simple, right?

We called for a price quote. Three times.

"Our" sales representative finally, grumpily, called back. New price: 20 times higher, for twice the space. By the way, even if we didn't want any more space, the price was going up to that amount. Take it or move out by the end of the month. (If we hadn't attracted their attention, we'd have been able to stay at what they obviously had concluded was a money-losing price, possibly for years.)

Skipping over the antics required to get a civilized explanation out of the colo company, the gist of course, was that their business model had changed. No longer were they accepting small customers like us. Clearly, "our" sales rep had a compensation plan which encouraged her to spend as little time as possible with small clients. Perhaps she had a good client in need of our space right away.

Who knows? The point is, their business model changed, regardless of our business needs. That's their prerogative. We eventually agreed on a rational migration plan out of their facility, but basically brother, when the cloud gets ready, you've got to move.

Application Stickiness

The last risk is the most consequential. When you use a cloud application, any modifications - whether you do them yourself or you pay the cloud provider to do them for you - stay with the cloud provider. Ask all those folks who have paid Salesforce.com for specific mods: how likely are they to leave? What's their switching cost? It's enormous.

That low cost of entry is the siren's song, beckoning you onto the rocks of lost application investment.

Some cloud platforms do allow on-premise alternatives: Microsoft's Office and Azure, for example. But the great majority of them do not. Indeed, the smart ones are trying to offer you ways to build your own code within their cloud, so you can't easily switch away. The siren's song...

A skeptic might portend that the sheer stickiness of cloud applications, in fact, makes it easier for cloud vendors to be abrupt, to be less considerate, to take customer satisfaction a little less seriously.

The End of "Easy"

It won't be long before you start seeing cloud vendors coming clean about the true complexity and risk of cloud. You can imagine their newly sincere admissions, coupled with announcements of new initiatives to enable the channel, as they differentiate by being the first to say, "Hey, we know it's not easy, and we're here to help."

Don't hold this against them. It is a truly magnificent feat to bring any new hardware or software to market, and all such products are so far advanced in ease of use and administration compared to even ten years ago, as to be magical to the primitive denizens of 2001. Vendors should rightly be proud of their advances and we should be happy for their progress.

What Future for Managed Services?

That said, rumors of Managed Services' death are greatly exaggerated. That's because cloud is less an IT support model than a new solution. Like any other solution, cloud is comprised of hardware and software (albeit at a different location). Like any other solution, cloud is an option available to be integrated by the Solution Provider to meet the customer's needs. Like any other solution, cloud needs support.

Doesn't the fact that cloud removes the hardware and software from the customer's premise, mean the support revenue will move with it?

Not really.

What does the small or medium-sized customer need an MSP for today? And in the cloud?

  1. Understanding the customer's business needs and budget. Needed in the cloud? Check!
  2. Picking from the myriad of available solutions. Needed in the cloud? Check!
  3. Arranging for their provisioning and integration. Needed in the cloud? Check!
  4. Migrating from the current solution to the new one. Needed in the cloud? Check!
  5. Monitoring systems for proactive support. Needed in the cloud? Check!
  6. Supporting end users. Needed in the cloud? Check!
  7. Supporting management's IT decisions and planning. Needed in the cloud? Check!
  8. Getting vendor support quickly and efficiently. Needed in the cloud? Check!
  9. Managing and reporting on vendor performance. Needed in the cloud? Check!
  10. Consolidating purchases from multiple vendors into one invoice.Needed in the cloud? Check!
  11. Migrating off of the current solution when that solution or its vendor, no longer fits the business need. Needed in the cloud? Check!

Does the cloud mean you'll have fewer server jocks on staff? Probably. But you've already either let go or retrained the large number of Exchange engineers you once had, as Exchange evaporated into the cloud.

You will have more people doing the classic Level 1 and Level 2 Managed Services tasks: supporting users, managing systems and vendors, helping management plan. Lest you think having "mere" Level 1 and Level 2 people is somehow counterproductive to profits, here's a dirty little secret from our benchmarking (not so secret if you subscribe to the Service Leadership Index®):

Remember, a multiple of wages of 2.0 is roughly profit breakeven; 2.5 or better is the Best-in-Class threshold. In short, low level engineers generally make you money, and high level engineers generally lose you money. You're used to having senior engineers because: a) they're usually more fun to hang around with, and b) in the pre-Managed Services business models, they helped you differentiate your company. As you know, in Managed Services, it's the efficiency and effectiveness of your operating processes which differentiates you, not your rocket scientists.

To use conflicting metaphors, in the cloud you'll have fewer rocket scientists.

But most of the things you've invested in, learnt and built in your Managed Services business, will be happily utilized in the cloud world. Your revenue and gross margin will be fine.

How?

Let's say the big hardware sales stop, and convert to monthly payments. That's a problem if you're product-centric, as about 24% of Solution Providers are. No question. But for the 76% of Solution Providers who are services-centric, does it matter if hardware sales stop? Let's see:

Now, will there be a dollar for dollar equivalency? Likely: historically clients spend about the same proportion of their budget on IT, regardless of what technology generation we are in. What they spend it on may change, but you're there to help them, right?

Will you need fewer server engineers? Yes. Will those cloud margins, which are essentially product resale, be subject to erosion? Yes.

But if you're doing your job right, you're more than offsetting any threats (revenue or margin erosion, lack of differentiation) by delivering the values listed in 1 through 11 above, as well or better than anyone else in your neighborhood, from your existing Managed Services operation. And by making sure you charge a win/win price for them.

Won't your customers just buy cloud directly from the vendor? Sure, a few will, but most won't. Why? For all the usual reasons: they don't have the time or the expertise, and they know they have less sway with the vendors than you do. If they want to take IT on by themselves, cloud or no cloud, let them. When they come to their senses, they'll be back.

Because IT, cloud or no cloud, ain't easy. IT is never easy.

And therein lies your value.