What are the main requirements and design principles for an omnichannel architecture?

From EcommerceWiki
Omnichannel ArchitectureOmnichannel Architecture AdvancedWhat are the main requirements and design principles for an omnichannel architecture

What are the main requirements and design principles for an omnichannel architecture?


For designing, setting up and managing an omnichannel architecture, the following design principles are to be taken into account:

  • Construct the architecture out of multiple single domains with autonomous functions ('modularity'). This also counts for multiple domains that are supported by the same software solution. This enlarges the flexibility and manageability. Furthermore, for some functions it is important they can operate autonomously when (temporarily) communication with other involved systems is down.


  • Consider the channels as 'thin' front-end-applications when possible, and do not try to add an abundance of complex logic. This keeps them flexible for changing content and interaction. Keeping the front-end thin usually has a positive effect on the system performance.


  • Try to put complex business-transcending topics in a flexible 'in-between layer' as much as possible. For business logic it is often the case that it should be adapted to changing market circumstances and learning effects.


  • Comply to industry standards in the following areas:
    • Techniques and protocols (REST, SOAP, Web services, http(s), (s)FTP);
    • Message structures (market standards, technique standards such as standard Idocs, Bapi's);
    • (Product)codes and characteristics (GS-1 standards).
By complying to these standards, flexibility and integration in cooperation can be achieved.


  • Use the 'single source'-principal for (master)data. When replicating data is necessary, work with a clear master/slave construction.


  • Fathom the data model of used software solutions and bring the model in line with as many applications as possible.


  • Secure data(base) entry for cloud solutions and for Software-as-a-Service(SaaS).
Be aware that for this there is an important difference between being able to use data, and actually owning the data. This could possibly be necessary for combining this data with data of other systems.



Up - Next

Page contributors

Host
TCBYC.png
Curators
Axel Groothuis
Partner
Magnus
Robin Gabriner
Consultant
Magnus
Experts

Share this paragraph

Comments