Whether it’s an off-the-shelf desktop application or a full blown enterprise solution, all software comes with hardware recommendations for optimum performance. Everybody knows that. So when I say “don’t forget the hardware”, I’m not for a minute suggesting you have taken leave of your senses and are planning to go live with a 200+ user PLM system, running on the old server buried under the half eaten cheese sandwich and last month’s Stuff magazine in the back of the IT cupboard. Of course everyone (well almost) makes sure the recommended hardware is in place for go live.
This is about when not to forget the hardware. In short that sandwich hosting server at the back of the IT cupboard should not feature at any point during your implementation. All too often hardware installation features towards the end of the project plan; at best it is in place for the pilot or final testing. Process workshops, demos and most other user-involved implementation activity, even user training, is regularly conducted on existing hardware used as test servers, usually considerably below-par of the recommendations.
So why is this a problem? Well as the old saying goes, if I had a pound for every time someone asked me “but will it be quicker in real life?”, I’d be quite rich. One of the fastest, most reliable ways to turn people off a new system, particularly those from the design community, is to plonk them in front of a screen with a never-ending egg timer. Reassurances and platitudes of “oh but it won’t be this slow on the real system” are met with suspicious stares; you might as well be telling them their laptop will be spitting gold nuggets from it’s USB drive after go live. All they are thinking is this looks like a mighty fine pain in the arse and it’s going to take me much longer to do my job if I have to use this.
User acceptance is critical to a successful implementation and once those seeds of doubt are planted it’s pretty darn difficult to weed them out. Show them something to give Speedy Gonzales a run for his money from the outset and you have far more chance of keeping them favourably onboard with the project. The required hardware has to be installed at some point, so the benefits of shifting this to the start of the project plan and using the recommended specs for the testing and development servers outweigh the negatives. If you know you will have invest in the hardware, get sign off and source it from the outset of the project so the installation is not delayed.
Now it’s not just an important factor for customers implementing PLM, or any other enterprise solution for that matter, but also for those selling them. Back in my presales days laptops with high specs required you to take an amiable donkey with you to the customer to demo. Nowadays machines with the specs of beasts are extremely portable so there is no excuse not to be fully charged and racing to go when demoing solutions to potential clients. If that dastardly egg-timer (or any other time measuring icon or blank page) appears on your screen, know with every second your content fails to appear your audience looses just a little bit more of the will to live.