Factors to determine which platform of ERP deployment is suitable for your business

Cloud is the current buzz word in business software. It is a new delivery model that brings both benefits and limitations. Beyond the hype, this blog is a basic guide to define the most suitable delivery model for your ERP system.

There is a dominance of cloud and hosted delivery models over on premise with 65% of new implementations using cloud or hybrid models. But this one-size fits all approach don’t work and value the opportunity to listen deeply to customer needs and revealing the best solution, and for some clients, on premise deployment can't be replaced.

Here are five points to consider when determining the ERP platform that best suits your business. 

1. Multi tenancy and version control

Pure cloud solutions run in multi-tenancy infrastructures with shared resources delivered over the internet. The benefit for the vendor is having all customers on the same version of the software.

From a customer’s perspective, the benefits are instant delivery, abstraction of the software platform and in most cases, a monthly subscription avoiding upfront capital expenditure. It also simplifies the upgrade cycle which is done automatically. Multi-tenant software is adequate for simple applications that do not require customizations, therefore suitable for small businesses. Modern applications mitigate this point by using configuration over customization.

For medium and large organizations, the inability to control the version lifecycle brings business disruption that makes multi-tenancy inadequate. The best option for companies with sophisticated business processes, custom development and multiple integrations with other systems is to run their ERP in a private cloud environment.

2. Locations with poor internet service

Networks are improving all the time, but based on the business location, a pure cloud solution may not be the best fit.  For some of these customers, there is limited, intermittent or no internet connectivity at all, pointing to the need for an on premise solution. The cost of a private connection would be excessive and hard to justify. In this case, an analysis should be done to assess the performance and reliability of local networks. Further, a cost versus benefit analysis will determine the best deployment, either cloud, hybrid or on premise.

 3. Data privacy and security

For some companies, cloud is simply not an option. There are companies that require complete control of their systems and data due to the high security nature of their operations.

Working with an ERP vendor that provides a choice of deployment models is preferable. Hybrid and private clouds may be an option to benefit from the cloud delivery model while keeping critical data secure.

4. Service level agreements (SLA)

Service level agreements need to be reviewed in detail to ensure adequate service and availability. When using a cloud solution or storing data and applications in data centers, quality of service varies greatly. A level 1 data center provides a 99.671% uptime while a level 4 data center has a 99.995% uptime.

In plain English, a level 1 data center may have up to 28.8 hours downtime a year while a tier 4 data center only has up to 26.3 minutes. There are other criteria that vary greatly like fault tolerance, data redundancy and power outage protection. The estimated cost of an outage and loss of data need to be compared with the cost of the service provided by the data center or cloud provider. For some companies, to have their system down for half a day is simply unacceptable.

5. Data ownership

When using cloud services, data retention and ownership need to be defined in the contractual agreements. An option to access data in a usable environment after the service is terminated is essential for all business, both from a legal and operational perspective.

Some vendors provide a local, limited copy of the software to be used on premise or within a hosted environment. It is therefore essential to check the fine print and ensure your data’s protected.