Citizen IT

In addition to the number of IT governed applications, the new employee will need to learn the locations of several databases, excel spreadsheets etc… this information is known as parasite IT.

The organic growth of systems and rough data sources are geometric in nature, as they follow a network of knowledge.

Imagine an organisation with a single specialism, then the organisation will have one or two specialist systems and a handful of collaboration tools.

As the organisation grows in complexity, adding specialist systems will require the exchange of information between teams. This requirement grows and often it is easier to build business relationships than to couple systems.

From the perspective of each team manager, it is hardly a problem to consider that there is an email exchange of a particular list, for instance a price book that is done every quarter. Not one group will be immediately impacted by the lack of Information Architecture and design oversight of these information flows.

However, transformations, mergers, growth and even people movements are inevitable. The cumulative effect of the lack of governance will expand the ungoverned information exchange beyond the ability of information architecture to catch-up.

Add to the mix the speed at which a business changes, and the dynamic nature of the requirements changing with it, then it is inevitable that the proliferation of Parasite IT will expand. The problem compounds when we factor the impact of time, as data becomes stale, parasite systems lose relevance with restructures or product evolution.

The clearest way to think of this problem, is by thinking of specialist groups as nodes of a network and the sharing of information as a links in a mesh.

The Opportunity

Create an Intentional Architecture to support Citizen IT with broad access to governed data and tools, ensuring that oversight is always applied.

The objective is to reduce the time spent on invaluable tasks and increase the speed at which valuable tasks can be achieved. The level of integration will still grow, and this will always be the case, however, the engagement required to obtain the required information should not grow exponentially.

To achieve a single pane of glass for all Knowledge Workers, we need to govern the sharing of information and swap the Information Mesh with an Information Hub. An effective pattern to enable the business SME (Subject Matter Experts) to co-develop the Information Hub we need two key capabilities:

1. Digital Workflows
2. Data Services

Digital Workflows are the enabler for a digital enterprise. This technology is ideal for organising cross-group processes and information exchange. Such technologies can extend beyond the process and integrate with document generation as well as seamless integration with data services.

Data Services is where both the Information Architecture and Citizen IT can flourish. The Information Architecture domain is rarely well implemented in organisations, whist there is a hunger for data as an asset and related benefits, I personally have not seen a broad implementation of this concept.

By swapping the Information Mesh (ungoverned) with an Information Hub (governed), we can infer the effort expended in the secondary screen to improve, assuming 20% of work is chore, then optimising any secondary process with timely and integral data as well as orchestrated request-response flows, then it is reasonable to expect a productivity gain of a factor of 4-5 times, reducing the time spent in chores to 4-5%.

Additionally, with a single optimised second screen we can eliminate the suboptimal distribution of licences and accelerate the consolidation of duplicate systems.

The knowledge worker

To illustrate the proposed architecture, I define the knowledge worker as one that has a specialist skill, given the right inputs will work towards an outcome and produce the right output. In the digital age, the knowledge worker will have a specialist system that is highly tuned to the value outcome.

The skill of the knowledge worker is intimately reflected in the systems, processes and procedures that will be applied to the inputs and well defined in this system.

The scope of work of the knowledge worker, however, extends beyond the specialist system and the worker will need to perform administrative tasks as well as data sourcing and presentation tasks that are not taping into the value adding knowledge they bring to the organisation. These tasks in Agile are referred to Toil, I call them chores.

Work types:

1. Specialist work – Value Added
2. Chores

The cumulative effect of growth in companies comes with two costs: increased complexity and increased specialism. The effect of these natural changes means that there is a proliferation of additional applications that additional specialist teams need to use, there is an additional integration burden and the complexities of the interdependencies between teams also increases the demand for secondary use of specialist systems.

The problem

We can set a principle that we would like our specialist work to be performed in a specialist system. This will mean that development of the system will be aligned with the specialism best practices and this development can be decoupled from any other system in our architecture.

As discussed, when the enterprises increase, services in the dimensions of complexity in variety of services, scale in increased segmentation of groups and complexity of products combine to increase the need for inter-group collaboration, and data sharing. Additionally, Mergers and Acquisitions and organic proliferation of systems cause the well-known issues we encounter in IT of duplication of applications, uncontrolled dissemination of data, corruption of data along with other sub-optimal practices.

We can then conclude that the worker will have multiple other screens… Typically in an organisation there is a proliferation of applications that tend to grow organically as with the number of integration points and duplication.

The proliferation of applications can be thought of as a Nx problem i.e., one that grows geometrically. Extending the model where accessing information from a foreign system, the knowledge worker has an increasing level of swivel chair activities.

The business information hub

The ideal architecture from the perspective of the Knowledge Worker is that their tool of choice works as expected. Workers can move between companies and know how to use industry standard systems. By ensuring that the Specialist application is designed to preserve its native features with minimal customisation (i.e. true to function), we enable knowledge workers to be effective by simply leveraging their inherit knowledge of the tool of choice in their fields.

The onboarding of Knowledge Workers into the organisation will consist of guiding them through the interactions with other business groups.

Taking this expectation and extrapolating into a top view of systems, we come to a pattern where all capabilities will have an Operational Data Store (ODS), which will serve to host business data for other systems and gather business data for the domain.

The ODS can integrate with digital workflow and serve information to other groups as well as to request complementary data from other groups to be stored into the ODS and presented to the Knowledge Worker or integrated directly into his tool of choice.

Whilst the architecture may look complex, it follows a few simple rules:

3. Each domain will have its own Operational Data Store.
4. Each ODS will master Enterprise Data for the respective domain.
5. Reference data will always flow from the ODS primary to other areas of the business.
6. Services are built from master data only.
7. Data updates are always updated in the primary first.
8. Enterprise Data Warehouse (EDW) is sourced from the ODS (i.e., indirectly).

Data as an asset

The concept of the ODS is to give form to Data as an Asset. This concept whist desired, is difficult to materialise.

The key focus of Business Reporting is on the companies EDW, where key business decisions are made. Whist it is inherently obvious that the value of decisions made at C level are critically high, and must remain a focus, with the knowledge that we should empower decisions to be made further down the hierarchy, we must also ensure that data is available at ground level.

With Agile, decisions are made by cross functional teams, and access to quality data will improve the quality of decisions made in the same way as with the top level.

It is also anticipated that many eyes on the data at the ground level will improve the fidelity of the data at the top level, enabling more accurate decision-making at C level as with all other levels of the organisation.

Digital Enterprise and Citizen IT

Quality data and access to data are different things. Having set up the right data stores in the ODS, there needs to a be contextualised mechanism to authorise access to datasets relevant to a broader audience than the specialist teams. The specialist teams will have broad access to its own datasets, and the authority to disseminate them when appropriate.

Data Sets should be designed and governed by Data Architects and Domain SMEs (Subject Matter Experts). Domain SMEs will assume the role of data owners and manage how these can be shared to other groups and under what conditions.

In this instance, we enable sharing of datasets by building data services through the digital workflow solutions. For example, we can wrap REST (Representational State Transfer) APIs and expose it as webforms with a process to fetch and update data to a record and/or to orchestrate a request-response between groups. By using the process workflows, the context in which the dataset is to be used, can be controlled and the dataset can be governed by specialist teams that own them as well as periodically reviewed by an Information Architect.

Having access to data services and process design will enable groups to collaborate and share data in the context of the processes that they design. Teams will have full access to data in their ODS and are able to share data from the ODS through process integration.

Automation and Enrichment

The scaling and proliferation of citizen IT is expected to mature in the organisation by providing a richer shared vision of “The Art of the Possible.” The seismic shift in expectation will fuel requirements for further integration and automation. The ever-growing appetite will be the perfect opportunity to fuel the relationship between the business and IT Architecture, coupled with a framework such as Open Agile Architecture, will lead to a healthy balance between intentional architectures and emergent architectures as well described in OAA standard. (https://publications.opengroup.org/standards/enterprise-architecture/c208).

Evolution and innovation

The additional layer can be seen as the enterprise playground for innovation, the datasets from ODS can be augmented with external reference data, market data, public data and blended. The Layer itself can be treated as the source for specialist systems for Business Process Re-Engineering, Data Science, Data Architecture, and other experimental capabilities.

The layer can be organised in the following Capabilities

• Information Architecture
• Citizen IT and Process Re-Engineering
• Integration and Automation Services
• Experimental Service

The capabilities will make use of different technologies and different principals to ensure that the architecture is maintained, and services are developed in a governed tested and productised.

Consider the Information Architecture, Operational data is managed and governed by teams, then the collection of ODS feed into the Enterprise Data Warehouse. A Master Data Management solution can augment this capability and govern the principles discussed. A workflow automation technology, including a low code platform can reference data directly from the ODS in confidence that has been governed by the Information Architecture and is true to master. Utilising this data with process workflow, we enable citizen IT to flourish. The services must be configured to only expose information which has been curated, furnishing the platform with integral data whist ensuring that the principles herein a built into the system. By Design, the architecture removes the risk of data corruption, duplication and becoming stale.

Additional Data Sets for data science and experimental capabilities can be implemented as their own data store and governed to exposed to the citizen IT only when their desired maturity has been reached.

Pattern for closed systems

It is not always possible to create read replicas of productions data, for closed systems where the data is hosted externally such as the case of Software as a Service or where the data is locked in a monolith for instance – it is still valuable to have the Operational Data Store within the same layer. One will need to observe quotas and other constraints and be very deliberate in how and when to use direct integration and when to pre-load data.

These patterns can overcome some of the common issues in a closed data system.

Suggested pattern for Monolith

Legacy applications with closed data systems can be wrapped to adapt them to the preferred citizen IT technology.

As an illustration, a keystroke system can be used in conjunction with a Robotic Process Automation tool to enable a drag-and-drop update. In this contrived example, the data from reference systems (ODS from another specialist system) can be listed in a form, where the user can select the appropriate value and the Robotic Process will key in the full menu path, search, and update the intended value.

Suggested pattern for SaaS (Software as a Service)

Similarly, SaaS solutions do not readily allow access to the database and API services are often limited to a set number of records and metered.

Creating a base set of data that can be accessed within the control of the organisation is useful and will reduce the need to add read only users to reference commonly shared information.

By being intentional about the Information Architecture, the organisation will be self-sufficient. As a side benefit, the organisation will also reduce the risk of vendor lock-in.

Conclusion

More is less.

Add to each capability a dedicated Operational Data Store and intentionally build a Workflow Capability as the primary means to govern data exchange between workgroups - i.e., creating a Data Hub and as a direct consequence, we enable Information Architecture to flourish in the ODS and enable Citizen IT to flourish through the workflow Capability.

By extending the Data Hub with additional shared or specialised capabilities, we can also enable external reference data, data science, data analytics, ML (Machine Learning), and robotic process automation to name a few.

The proposed architecture and patterns have the intent to maximise the customisation of supporting services, specific to the enterprise and preserve the specialist systems to be true to their intended specialism. For instance, a CRM (Customer Relationship Management) system will be implemented as stock standard, and minimal customisation will be required. Information that is required to propagate to delivery teams can be customised in the ODS and the instructions configured in the Workflow. If there is a need to change CRM, this integration may need to be adjusted, but not rebuilt.

The proposed architecture also creates the opportunity to centralise the implementation of Information Architecture and the creation of a Data Hub. The pattern enables the sharing of information in a process-driven approach so that we maintain the concept of data primary ship whilst maximising the proliferation of Citizen IT.