These 15 reasons that Cloud fail represent the most frequent reasons that are documented by organisations. There are many more no doubt, however because the area of Cloud is somewhat new, particularly in New Zealand, the list of public failures surrounding the implementation of Cloud is low. That and people don’t want to be embarrassing themselves of course.
It is important to recognise what fail means as well. We are talking about everything from a catastrophic implosion of service through a service that fails to deliver what was promised, is more expensive than originally stated, is inflexible, or just generally annoys the hell out of your business.
The failure reasons are then broadly categorised as:
- Vendor is not really selling Cloud service or is immature
- SLA’s not Cloud ready from vendor
- ICT organisation not ready
- Going Public if you are big
- Going private if you are small
- Not engaging with the business users
- Forgetting security and data sovereignty
- Poor Project Management and Sponsor Engagement
- Not understanding requirements
- Early adoption
- Doing it to purely save money
- Insufficient Testing
- Insufficient Skill Set
Vendor is not really selling Cloud services or is immature
There is a lot of “Cloud washed” services being sold in the market at the moment, particularly in smaller countries (hello New Zealand) with a lesser mature market than the larger countries. Cloud washed services are regular services that are rebadged and sold as true Cloud services, when they are anything but.
Cloud services are on-demand, have broad network access, resource pooling, are rapidly elastic, and have measured service. They are not facilities management, outsourcing, or capacity on demand services that typically lock you in to long contracts with very little flexibility in how you receive the service.
SLA’s not Cloud ready from vendor
As well as the usual SLA’s that you will need to negotiate make sure that the vendor a) offers and b) can deliver on some additional ones that are either more critical with, or peculiar too, Cloud services.
Disaster recovery is particularly important. For example, if the vendor is hosting fifty companies on their Cloud how do you know where you are in the recovery pecking order? How do you know their DR actually works? When did they last test it?
In addition pay careful attention to other SLA’s such as data location and exceptions based on location, trade rules, and other legal aspects.
ICT organisation not ready
The move from traditional in-house services, or outsourced services, to Cloud services demands a different approach to management of ICT services. If you don’t get your ICT organisation structured and operating correctly then there will be process breaks and potentially missing functions that will cause the service to be degraded. An end-to-end support model will highlight the areas of risk.
Going public if you are big
If you are a very large organisation or company then running your services as private Cloud as opposed to public Cloud is likely less risky, easier to manage, easier to change, and more agile will still retaining the same economy of scale cost savings that public Cloud brings.
Going private if you are small
Conversely, if you are smaller organisation or company then investing in a private Cloud may be problematic whereas taking advantage of a public Cloud service will likely make more sense.
Not engaging with the business users
A recent study found that 64% of users within organisations already are using Cloud services of some description, with email and DropBox at the very top of the list. In addition, the rise of people bringing their own devices (BYOD) is increasing.
If you are talking to your business about what Cloud services you are going to stand up for their consumption you run the very real risk of simply being ignored and the service not being used at all.
Start your strategy for Cloud now. It’s here for the medium to long-term and you will be adopting some Cloud services at some point in the future. Waiting may force you into a reactive adoption.
Forgetting security and data sovereignty
These two characteristics can kill any project faster than a bolt of lightning. If you don’t understand what your policy is up front on these two issues then you will be in a world of pain further down the track.
Just because you’ve bought a shiny new Cloud email service doesn’t mean that it will work with your existing applications and ICT services. Make sure you get the architects, or seek advice, early to ensure that what you propose to deploy will integrate with your existing ICT services.
Poor Project management and Sponsor Engagement
Cloud is a complex project. Failure to get the project and programme management right will cause a major problem. In addition, without a strong, enthusiastic sponsor who has the authority to clear road blocks you will run into problems.
Not understanding requirements
Just like any project you need to define requirements up front. The chances of you finding a one size fits all Cloud service for your organisation are pretty much zero. Make sure you have clear requirements so you can at least negotiate with your business if you need to change them.
Cloud is still relatively new with analysts picking 2013 as the year that people just start to stick their toes in the water to actually see what it is like with the majority of the rest of us starting to plan for it. That, coupled with the absolute explosion of “Cloud services” onto the market means that early adoption is risky and must be carefully managed.
Doing it purely to save money
Always a bad idea. There are plenty of benefits to Cloud, of which one may be reducing some costs. Make sure that all benefits are balanced.
Pilot, pilot, pilot, test, test, test, and then test some more. The number of projects that cut down testing is astounding. The first place to look in a failed Cloud service (or any ICT service for that matter) is testing.
Insufficient Skill Set
Cloud is new. Don’t expect your existing staff to understand it without significant upskilling. Pick people who you trust and have a demonstrated background in Cloud service management. Sadly, there are few, however the number will grow.