The recent price cutting from the big public cloud vendors (Google, Amazon, Microsoft) has re-energized the familiar debate about which is more cost-effective: using a public cloud service, or doing it yourself?
While any intelligent answer involves a healthy amount of “it depends”, I found myself thinking about another time and place — where we as an industry were debating the merits of outsourcing vs. keeping IT operations in-house.
Then as now, the primary argument for outsourcing was economic. A lot of people took the bait — but it didn’t always work out well.
Of course, things are certainly different in 2014 — but are they all that different?
I can’t precisely recall when it started in earnest — maybe ten years ago? — but all of the sudden the notion of outsourcing IT production to specialists became very popular.
As I was working for EMC, this was a matter of some concern at the time. Our primary customer engagement model was to work closely with enterprise IT groups to design and implement rather sophisticated storage and continuity solutions.
If it was all going to an outsourcer, we could sometimes end up being on the wrong foot. Naturally, I became curious, and made an extra effort to meet with customers who had either made the decision to outsource, or were seriously considering it.
Before long, a standard picture started to emerge from my conversations.
The primary motivation — of course — was financial.
The outsourcer would contract to deliver an inventory of specified IT services at a unit cost significantly below what was being done internally.
As a sweetener, the outsourcer would frequently offer to “buy” the IT assets already on the floor (who wouldn’t love to get a big check?) as well as “hire” the existing IT staff as part of the outsourcer’s operations.
Initially, I found this curious. Given that the outsourcer had contracted to "buy" the existing assets and "hire" the people, where was the sustainable cost advantage going to come from?
I quickly found out how this was done: by doing exactly what was contracted for, but doing it for less: fewer/cheaper assets, fewer/cheaper people.
Additionally, the outsourcer could potentially gain an economic advantage through pooling and sharing infrastructure. And — truth be told — outsourcers usually had a better grip on IT costs and processes than many of the companies they were selling to.
However -- as a for-profit business -- not all of these savings would be passed on to the customer.
Beneath the ostensible economic argument, there was sometimes a harsher truth: IT management had lost the confidence of the executive team. And by the time an outsourcing proposal was on the table, the die had been cast.
Hard Choices Abounded
From a business perspective, whether or not to outsource IT was a strategic decision made at the highest levels.
On one hand, a path of continuing to invest in internal capabilities -- with all that is implied.
On the other hand, a path where you would trust a business partner to do run IT on your behalf — but losing control in the process.
I clearly remember how my customer conversations would go, usually along these lines.
#1 — Is Being Good At IT Strategic To Your Business Model?
Every company uses IT to some degree, it’s how they use it that’s different. Some companies need to be good at IT to be ultimately successful. Others see it more as a cost that should be minimized.
You’d be surprised at how many people hadn’t thought about it in those strategic terms. For those that could conclude that being good at IT was more of a cost to be minimized vs. a capability to invest in, I reminded them that things could change quickly, and reversing their decision would not be easy.
#2 — Is The Company Prepared To Invest?
“Invest” means that you’re prepared to spend now in hopes of a payoff down the road. That means spending on good people, good technology, good process and good projects. Not every investment will pay off, but the portfolio should do well over time.
When it comes to a fast-moving game like enterprise IT, you’d better be prepared to invest — otherwise you’re better off simply handing it all over to someone who can run your mess for less. You’d be surprised at how many companies back then weren’t prepared to invest: either they saw more attractive investment opportunities, or they were constrained financially from doing so.
#3 — Can You Live With Loss Of Control?
For all of its warts, in-house IT has one redeeming characteristic: you’re in control of your own destiny. The dissatisfaction with outsourcing soon became commonplace, mostly around this theme.
For example, if the outsourcer wasn’t performing to agreement, your remediation options were somewhat limited. After all, they had all your applications, all your data, all your infrastructure and all your people.
Not exactly a strong negotiating position.
Or — more pragmatically — let’s say your requirements changed: new functionality required that had to work with the existing legacy. Many outsourced customers found that trying to modify what they had contracted for to be an entirely new level of IT hell: as the outsourcer had all the leverage in any negotiation.
Things change — that’s part of being in business. If your circumstances change, how quickly can your outsourced IT capability respond?
The Grim Reality
At the same time, I also became interested in how outsourcers did what they did.
To be accurate, I met a few very good outsourcers who arguably did IT production better than most IT shops I knew. Their technology was better, their processes were better — and they had the numbers to prove it. Good for them.
But I also saw a lot of grim reality — outsourcing shops where continual cost-cutting was the order of the day — not in a strategic sense, but in a let's-be-really-cheap kind of way.
Production gear that was well past end-of-life and no longer supported. Corners cut on availability and recoverability. Poorly documented processes coupled with apathetic and demoralized staff.
Basically, a disaster waiting to happen.
I realized that their customers probably didn’t have the knowledge — or perhaps the interest — to probe behind the curtains.
To be clear, today I meet plenty of customers who have formed deep and productive relationships with their outsourcing partners.
There are many examples where two have fused into one — and both parties have done well as a result. But — historically speaking — I would views those exceptions rather than the norm.
But for those who realized it wasn’t working out, their options were daunting. Find another outsourcer, and move everything to them?
Yet another level of IT hell, and no real guarantee that the experience would be any better as a result.
What I did notice a few years later was a small backlash, including more than a few organization who realized that the outsourcing model no longer fit their needs, and they had to bring most or all of IT production back in-house.
The challenge of creating an enterprise-class IT organization essentially from the ground up is no small challenge, but that’s where they were.
Parallels With Today’s Public Clouds?
To be completely fair, there are big differences between that historical outsourcing wave and today’s public cloud wave. But there are also similarities worth considering.
As before, the lead headline is “save money”. And as we’ve appeared to enter a period of price wars, there’s plenty of fuel for that debate.
But let’s not forget, enterprise IT suppliers (well, most of them) are also continually lowering their prices, especially when it comes to hardware infrastructure. They don’t issue a press release every time they do this, though — so perceptions here might be skewed.
I also see IT operations of decent scale delivering non-trivial unit cost savings simply through operational improvements: just-in-time provisioning, optimized service catalogs and transparent pricing.
I do think public clouds have an inherent cost advantage when it comes to transient or unpredictable workloads. However, when infrastructure requirements are more predictable, you’re back to the rent-vs-buy discussion. Parts is parts.
While it’s true that all of the public cloud providers seem to be in a race to deliver ever-more functionality, I’m skeptical.
If I were running a public cloud — and didn’t want my customers leaving me — I’d want to get them hooked on services that only my cloud offered and others didn’t — the ultimate in stickiness.
With one notable exception, there’s little motivation to make public cloud services, functionality and operational models compatible with either each other, or — ideally — their customers’ operational model. That would make it easy for customers to leave!
Paul Maritz once famously described public clouds as the Hotel California -- “you can check out, but you can never leave”. Actually, leaving your public cloud isn’t impossible — just like leaving your outsourcer — but it’s going to be a major challenge.
I do have to point out — the primary exception in my mind is VMware’ vCHS. The mission here is to deliver public cloud services that are functionally and operationally compatible with existing vSphere environments. Should vCHS ever branch out to a franchise model, I would presume that all vCHS-compatible instances would be highly compatible with each other.
The Road Ahead — And A Look Behind
Our industry’s experience with outsourcing taught us that the appearance of short-term payoffs can sometimes result in longer-term pain.
We can optimize for today’s circumstances, but what about tomorrow?
Choices may be hard, but there are historical lessons we can learn from — if we choose.
Because this isn’t the first time there’s been an option to have someone run your IT for less.
Like this post? Why not subscribe via email?