With Windows 7 breaking all box office records on Amazon and the 64-bit version selling out in hours in Japan, corporations across the globe are gearing up for their imminent Windows 7 migration. Windows 7 is the destination of choice for many ambitious IT managers. Many even want to visit exotic virtual LANs. But, be warned. As with many mass migrations, a lot of people are going to get hurt, some fatally.
But it doesn’t have to be like that. Indeed, it shouldn’t because Windows 7 can not only make your company degrees more efficient, but it can lay the foundations for future generations that will be even stronger, using 64 bit computing and virtualisation. Get it wrong and you will waste millions of man hours, not to mention pounds. It’s all about the preparation, but how can you prepare when you don’t know what to expect? Here’s some pointers.
There are many classic traps that Windows 7 migrators sleep walk into. We asked seasoned Migration expert Frank Foxall, CEO of Camwood, to identify the most common mistakes that could bring your enterprise to a halt, to make your support desk light up like a Christmas tree and bring your name to the attention of the board of directors. Foxall says the following 11 are the mistakes you should avoid.

One: Treating applications as an after thought
Information technology is about end users, writes Foxall, and the software tools they use to work more efficiently. So we should never forget that IT is all about applications. But curiously, in many migrations, applications are the last thing anyone thinks about. Which is why one bank found itself on the verge of moving to a 64-bit App-V platform on which seventy per cent of its business critical systems would fail. Luckily, they discovered us in time and we saved them from an awful fate.
People forget that a new operating system affects every part of your IT operation: every piece of hardware, node on the network, server farm or data centre. So when enterprises install the new operating system, upgrade the hardware and address infrastructure issues first. Then they think about applications. But decisions about apps should determine decisions about infrastructure – not the other way around.
Besides, their application estate will be entirely different from the one involved in the last big migration. It will have grown far bigger and more complex.
So your application strategy is your user strategy.  Everything starts with your users and what they need. You build IT around users. No-one should build an IT system and expect the users to adapt to it.
Without a clear apps strategy, your infrastructure decisions will be wrong.  But people do it. That’s why you see people today building virtualised data centres for applications  that won’t run in it.
Two: Missed opportunities through going in with your eyes closed
Many IT projects are dogged by lack of foresight. For example, Windows 7 may be the immediate priority but in the longer term everyone’s thinking about virtualisation and 64-bit computing, for which a move to Windows 7 would be the perfect opportunity to lay the foundation. If you were alert to the possibilities, and brave enough to implement them, you could reach this computing nirvana sooner and at less cost.
So a serious Windows 7 migration process should at least analyse 64-bit and virtualisation strategies and options.

Three:  Bringing needless baggage – have a clear out first
Every application that’s ported onto the new system costs around £3,000-£3,500 just for packaging and migration. So it makes sense to dump the useless ones. You’d be amazed how many big companies waste millions migrating applications that they don’t even use.
In a 3,000-app migration one company included 250 apps it didn’t need, costing it an extra £750 to £875k! That’s before you factor in the cost of licensing, support and management – all wasted.
Every IT department knows it’s a waste, but few cull rigorously. At Camwood, our scientific rationalisation process includes discovery, categorisation, context and a usage survey that reveals how much each application is actually used.
Is your app’s journey really necessary? Can it even work with the new target platform? Will those 16-bit apps work on your 64-bit environment? If not, it could be time for an app retirement.

Four: Lack of compatibility testing
A few migrations ago, you might have got away with a ‘fix on fail’ program. Today retrospective compatibility fixing will not work. Tragically, it’s still popular. And when an enterprise with over 300 applications needs serious, automated application compatibility testing, they will often make do with low-end tools on the market. But sophisticated tools, that use heuristics to identify the components in each application, are needed.
This can give an instant report on which apps are ready for Windows 7 and its virtualisation platform.

Five: Too little, too Late
We’ve seen hundreds of major migrations take place and in many cases they failed because there were not ‘front-loaded projects’. They try to match a new environment without knowing too much about themselves.
The discovery phase is where you collect everything you know about your application estate and bring it together into one place. We envisage Total Application Awareness, which inform you on six levels: Users & their usage patterns, applications, business processes, change processes, target platforms and the relationships between apps and databases.
Know these and you’ll migrate faster and avoid all those outsourcing and packaging costs. And disrupting the end users.

Six: Overlooking your web apps
You wouldn’t think it possible, but people still under estimate the importance of the web. There are hundreds of web apps and website services that people depend on now to do their jobs.  But these are sadly overlooked in Windows 7 migrations, where everyone will be moved Internet Explorer 8 (then 9). But many web apps won’t be able to go with them. So web applications need testing just as rigourously as the packaged software. Otherwise you could end up like one manufacturer, which inadvertently locked out all its suppliers, at a critical time.

Seven: Migrating by hand. Manual labour is not enough
A major migration in action is like a complex machine with lots of moving parts and process dependencies. Manual migration is never going to be good enough for such a complex operation.
Never run a major migration without a collaboration and workflow platform – ideally one designed specifically for migration, packaging and testing. It’s best to keep a single data depository, used by every step in the migration chain, which gives a single source of truth about your application estate and the progress that each app is making through the migration.

Eight: Understand what you’re getting into
Before you go anywhere you need to know about your new home. But it’s not easy to understand a platform inside-out if you’ve never used it.
So what do you do? Devote your time to research and working with Windows 7? Or find a company that’s crawled all over Windows 7 already and knows the implications for applications.  At Camwood, we were closely involved with the Windows 7 launch and a key partner in Microsoft’s Technology Adopter Program. To prove it, here’s some hazards to look out for when migrating to Windows 7:
Windows Help (.hlp) files are longer no supported in Windows 7.
If you deploy Windows 7 as 64Bit, you’ll need alternative serving strategies for any 16Bit applications.
Applications designed for an older OS, will find the runtime files they take for granted are missing on Windows 7.

Nine: Asking outsourcers to do what they’re not built to do
It’s great that packaging and sequencing apps can be done by outsourcers.  But they can’t do up-front discovery, planning and validation work. They’re specialist tasks. But they’re essential for a successful migration.
In other words: outsourcers are like computers. They’re only as good as the instructions you give them. If you throw a bunch of apps over the wall without detailed guidance, you’ll get back a significant portion of app packages that fall down on the new platform.

Ten:  Throwing bodies at the problem
The point about IT is that it’s not supposed to be labour intensive. But time and again a big company migrations hits trouble in the applications phase and panic sets in. The knee jerk reaction is to throw bodies at the conflict caused by compatibility, testing and packaging issues. Unfortunately, people are expensive. More to the point, people were never the problem, information was the problem. We have run huge migrations with few in-house staff.
 A pound spent in up-front discovery and analysis saves 8-15 pounds in packaging and testing.

Eleven: Failing to use migration by products for the business
Like any journey, your Windows 7 migration program will teach you some lessons. In fact, it will unveil some information gems. Intimate knowledge of your entire app estate, for example. Your newly created workflow process for fresh applications will be invaluable too. As will your new reporting environment. One company used these best-practice processes to cut application acquisition from 105 days to 17 days by use of our process benchmarking exercise.
It’s a disaster when companies ditch these best practise efficiencies once the migration is complete.
So there you have it, the promised LAN holds many treasures. But make sure you don’t get lost on the way.