Taming the Enterprise part 1 discussed the problem of siloed systems and decentralized information in organizations. As a continuation, this post lays out a pragmatic approach to phases for moving to a centralized and integrated system.

Phased Approach

As enterprise systems mature, data can get fragmented across different sources i.e. SORs, spreadsheets, etc. With each SOR using their own data storage, data entities, relationships, etc., getting a complete picture becomes very challenging, if not impossible. As mentioned in the previous section, introducing an API gateway to build integrations between existing systems – that will not directly integrate otherwise – can help with getting the data from multiple sources, but the transition can be painful for the users.

Change is hard.

Large ships turn slow. The same is true for large organizations. Throughout the organization, plenty of new systems and workflows introduced by business units are met with poor adoption. Often these new systems cause users frustration to increase. As a result, the users abandon them. Any major impact to business workflows demands research, stakeholder buy-ins, value-added scrutiny, and rational emotions to be fully adopted. Therefore, it is beneficial to take a phased approach.

In the initial phases, wean the users off of the shadow systems. Move them to use a centralized solution that bridges any functionality gap. During the transition phases, allow users continued direct access to the SORs to ensure uninterrupted business flow.

Followed by integration with SORs in a gradual manner until the user is fully interacting with the engagement layer only. Users drive the majority of the design and requirements for the engagement layer since they will be using the applications and fill any gap to avoid workarounds. Once the application development cycles mature, the users are comfortable accessing the system only in the engagement layer.

After that, further improvements and enhancements can be applied to these layers.

Single Source of Truth

For organizations with a mature enterprise system and advanced understanding of their data, a holistic data modeling approach such as a Single Source of Truth (SSOT) system can be implemented to improve data quality and accuracy.

An SSOT implementation requires custom data modeling and data relations. Begin by consolidating data models and attributes from all the existing SORs in use into a single data model. Define custom data models to match their business needs. As a result, any solution – internal or external – integrating with the SSOT must be compatible with the APIs and data models. Use data mapping to map SOR entities to custom data models. Finally, position SSOT between the APIs and the data layer. It acts as the data source for the entire enterprise. All data flows into, and from, the SSOT. Most importantly, sync the data – either periodically or by action triggers – between the master database (that holds all the custom data models) and other data sources such as ERP, CRM, etc. to ensure accuracy. SSOT will in covered in detail in another post.

Challenges

As an organization matures its enterprise systems and take control of its data, vendor relations mature as well. When working with vendors, only those vendors that can support integration to the enterprise system are considered. Since not all vendors offer SORs that can integrate externally, vendor options become limited and may have higher licensing costs. However, in the long run, it offers potential savings since the organization is not locked into a single vendor and can alternate to keep costs manageable.

Another challenge is data model management. While reducing dependency on SORs, organizations take charge of their own data modeling and data mapping between SORs. With this power to have control, comes overhead, and data architecture maintenance.

With the right team and resources allocated, an organization can not only tame their enterprise systems to work well, but it can also open up possibilities to make the user experience better. Better user experience, better adoption, better results. Win. Win.