On Sunday I attended Brian Gammage’s session Workplace Transformation through EUC Transformation: How to Plan, How to Start. This was a 30-minute session.
Brian joined VMware in 2011 after being Gartner’s top analyst in the VDI space, so him moving to VMware was the first of a multi-year wave of famous desktop people joining VMware who “get” the desktop space. Now he leads a team that supports customers doing larger VMware EUC projects.
The session was not a technical session or about VMware products, rather, it was about how EUC and VDI projects start in customers, and how (from the business standpoint) you can make these types of projects successful.
Roughly speaking, EUC projects have four phases:
- Project identification, setting goals
- Project scoping, which is where you turn goals into plans
- Validation / testing / proving. You have a plan, now pick the products, suppliers, etc.
- Economics analysis. Does it all make sense? Does the business case jive with the investment?
Brian had a cool quote about the business case: “Everyone understands cost. Virtually no one understands value.” :) This makes me wonder whether we should all be talking about TVO instead of TCO? Actually, Templeton said this in the Synergy 2011 keynote.
Anyway, the challenges organizations face when starting EUC projects today are:
- So much choice.
- This whole mobile/cloud just makes everything tough. People expect everything to be really hight scale, right? Cloud means that everything is super cheap now, right?
- There’s a productivity / growth agenda. All the pressure is to increase productivity, we have to grow features and capabilities, etc.
- Most EUC projects are based around a complex set of assets we own. The “big change” of a device is when it breaks and it’s replaced. The “big change” of OS migrations is because our current OS expires, etc. This means that we have to make these big change decisions based on the timing of assets, not based on whenever we want.
The interesting thing here is that none of these challenges are about technology. The EUC products today (not just from VMware, but Citrix, etc.) give you a range of options that mean you can do pretty much anything you want.
So the challenge is not about a lack of options, rather, it’s living with the things that you’ve chosen to deploy and with the management model you’ve chosen.
Brian went on to say that, broadly-speaking, a desktop environment is going to be about 20% capex and 80% opex. So opex is the big part which means you get the biggest bang for your buck by focusing on opex. So how have we controlled desktop opex over the past 20 years or so? We try to standardize. Standardize the processes and standardize the tools. Standardize the desktop images. This means that, by definition, the “best practices” we’ve built up over the past 20 years are the very things we’re trying to break through today.
So today everything is fast and diverse, but we’re still trying to standardize everything to squeeze out costs and maximize value… what???
Oh, and by the way, in the middle of all this, we also have to do this today on all this stuff we don’t own. :)
Brian had another great point about user choice or BYO. See said it’s something that exists in every organization—it’s just that the line is different from company to company. Paraphrasing what he said, “In a commercial organization, the CEO always uses whatever he wants, and the janitor gets whatever we give him, and somewhere in between is a line." Every organization has a line. The only thing that changes is where that line is.
As for this ownership question, that also applies to the back-end. There’s this perception that the “as-a-service” options are always a better deal. But are they? Maybe yes. Maybe no. But how do you know? (Brian also said the biggest thing that affects whether you go with “aas” versus “own” is (1) what’s your company’s balance sheet strategy (trying to get things off the balance sheet? Ask your CFO if you don’t know), and (2) seasonality of workers.
Anyway, so there are three “drivers” of IT in general now which also apply to EUC projects.
- We want to withdraw from managing “things”. (Note he says “withdraw,” not “stop,” because it’s a process.) Reduce that overhead, focus on cost, not results.
- Downsize the unit of investment. Make it smaller which can give you more control (and easiest places and chunks to audit). So as it is now, you’re doing this “big steps” every 3,4,5 years, then in between those big steps you only make small adjustments. The problem is budgets go up, down, change, etc. So if you can take shorter steps, you can shift away from these platform-oriented huge steps and towards application-oriented continual improvement small steps.
- Make it all less visible for users. “Changing devices should be like changing shoes. Who cares?” The drive is to shift the focus to the results, not the means needed to get those results (which is how it is today).
When those are your drivers, you’re designing to change.
This is a big change since historically we designed things not to change. (Everyone is the same. If they diverge, re-image them so they’re now the same.)
The challenge to all this is the in today’s world, no one is starting from scratch. We all have baggage. So we have to transform from what we have to what we want.
If we could magically start from scratch, it would be easy. We’d build a strong single sign-on, we’d create a service catalog, and we’d have a great policy engine on the backend.
But how do we apply this on top of our existing baggage?
The idea is we just take all of our existing applications and make them good. :) haha. But many (most?) or our current apps are tied to a specific OS. So we have to figure out how to expose or “surface” those apps to other platforms and devices.
By the way, we talked about the 20/80 balance of desktop costs being capex/opex. But what does that look like for mobile? Probably more like 90/10, i.e. mobile is 90% capex and 10% opex. That means that while in the desktop world, only 20% of our spend results in change, in the mobile world, 90% of our spend results in change!!
Anyway, so how do you get started with this? How do you build a plan?
Next Brian explained how VMware handles this when they do customer engagements. He was clear to say this is not the only way or not even that this is helpful in every environment, but anyway, here’s what they do.
To build a plan for EUC transformation / VDI / desktop whatever, they figure out the following:
- Governance. Who makes the decisions? Does IT have ultimate central control? Do individual departments have the power to pick what apps they use? Can business units pick their own devices and platforms? Are you okay with how things work now, or do you want to shift governance?
- Audit. If you don’t need to know, don’t keep it.
- Elasticity. How much elasticity is there in your environment? Do your users change? What’s your turnover? Seasonality? etc.
Also, by the way, most customers need a lot more than “task workers versus knowledge workers” to segment their users. Brian said that most companies end up with 7 different types of users:
- Productivity task worker
- Communications task work
- Content/media worker
- Office-based information worker
- (I’m missing one.. I’ll try to find this and add it later)
Brian closed the session with this summary: “We build plans because things change, not because they stay the same."
- content/media worker