Nathan E. Myers - Self-Service Data Analytics and Governance for Managers

There are any number of relevant and overlapping frameworks that cover portions of IT governance and even portions of data analytics governance in the finance and accounting environment. However, no single framework exists that is fit for the universe of self-service data analytics builds. We will draw from mature system governance, model governance, data governance, process governance, SOX 404, COSO IC (internal control framework for the financial reporting process), COSO ERM, and COBIT 2019 (ISACA) frameworks, and even the AICPA's Statements on Auditing Standards – to sketch a foundational governance model that your organization can implement and build upon as necessary. This must be done early and determinedly, so it is in place and can play a formative role in safeguarding your organization, as it embarks on its inevitable digital journey.

Let's look at the environment from the perspective of the employee.

Generically, these operators are analysts, though very often, actual analysis is only a sliver of their day, compared to the time spent on the raw processing steps they are expected to perform, prior to generating output for evaluative analysis. Such processing steps likely include capturing information from a number of sources, enriching the data to assemble suitably rich datasets, before completing further processing steps and transformation steps to yield final outputs in the form of information and reports. It is really only at this point that the operator can embark on true analysis in earnest.

Such outputs are often validated against prior periods to attempt to identify any abnormalities or errors. There may be key ratios that are calculated, observed, and compared to get comfort that the output is correct. There may be other sanity checks and detective controls performed to ensure process effectiveness and the integrity of deliverables. We will refer to these broadly as analytical review procedures, and we will assume that these procedures are partially about quality control and error detection, but also partially about understanding the business better, so that value can be added as a true business partner. It is these latter analytical processes that lead to actualization – ensuring high-quality outputs, owning your numbers and outputs, and gaining insights into the business through analysis.

If an organization is large enough to be layered, a pecking order emerges. More junior, if not entry-level, staff will be buried in the assembly of information and information processing. Over time, if they are good, they will strive to get faster, better, and more efficient at assembling and processing data. Those who are able to get their head above water enough to truly understand what the processing outputs are telling them, and those who can glean critical insights surrounding the business, may gain visibility and be recognized as true process owners. With luck, they may graduate to being a reviewer, supervisor, or manager – sampling the butter with a critical and experienced eye, instead of churning all day. This is the aspirational path to advancement for many in the analyst ranks, whether across finance or accounting functions, operations functions, or any of the business analytics or reporting roles that pepper the ranks of large organizations.

In times of great flux from business growth or volatility, departmental reorganizations, regulatory demands, or other external pressures, operators may find themselves quickening the pace to get their heads above water, only to be rewarded with more of the same work to drown them anew. Take a deep breath, because you are about to be pushed right back under the surf, until you bed things down yet again at a higher plateau of utilization, with even less time to perform actual analysis, to learn, and to add value to your business . Just when the operator begins to get to a point where they have learned enough about what their own deliverables and end-products are telling them about the business to begin to add value, they may be asked to cover another 20 accounts, or take on 20 more processes, or to produce five more weekly deliverables.

Very often, processing pain points to the need for additional system features and functionality. Organizationally astute analysts will articulate the need and route the demand item to the core technology demand queue, with the hope that it will introduce time savings, when it is eventually delivered. Such hopes can be dashed with the slash of a red marker, as requests are buried at the bottom of an interminable wish list, when higher priority initiatives take precedence, or when requestors lack the influence to argue for the relative priority of their requests. This is the reality of the environment in which many operators find themselves each day. Later, we will propose that self-service data analytics is one of the few tactical levers that can be pulled to lock down the data preparation, transformation, and processing steps in a stable, controlled manner, and to capture efficiency in the form of time savings. For now, let's look at the same environment from the perspective of managers.

Some managers have come up the ranks in the career progression outlined above; they may have started as an analyst and sharpened their technical skills at a faster rate than others, such that they were able to successfully execute their workload, but even more, they learned from their outputs, demonstrated value to internal clients, and ultimately moved up. Others may have been hired externally and brought into the organization, and may be less aware of the processing steps and rigors that their teams undergo each day. Similarly, existing managers within the organization may have been asked to assume ownership of a function, and may again be less familiar with the processes required to generate departmental deliverables. Irrespective of which of these profiles is most applicable, managers will be expected to deliver an increasing number of accurate and conforming deliverables, these days without the free hand to hire additional resources to meet increasingly stringent demands.

Most managers are focused on minimizing process variance to ensure consistent quality of outputs. In the mature systems-based environment, they insist that as much processing as possible is performed within systems, and that system outputs require little manipulation, in order to generate deliverables. Deliverables requiring complex, multi-step, and unstructured Excel-based operations introduce significant risks. Accordingly, astute managers track the progress of the technology backlog, ensure that they weigh in on the prioritization queue, and shepherd their must-haves through project stages to a scheduled release. In this way, they can ensure that the systems environment supports their processing needs. They would prefer to use well-documented, prescribed and controlled system features and functionality to perform the lion's share of processing, rather than relying on unstructured manual processing steps. The goal is to extract output from systems that is as close as possible to final form for departmental deliverables.

However, often the core technology systems have a lengthy backlog of competing priorities that may have been built up over years, that can be difficult to navigate, and which can result in significant delays in the delivery of needed features and functionality. Many readers will have felt the disappointment when they learned that a promised sprint or release has been postponed, or when they learned that the all-important and long-awaited Phase 2 of a large-scale strategic technology delivery is below-the-line for the year, left unfunded on the shelf. Does that mean that teams must continue to work in an inefficient and unstructured way, until such time as the technology investment is revisited in the next investment cycle? Perhaps not. In the section Arguments for Self-Service Data Analytics Tooling, presented later in this chapter, we will provide a preview of self-service data analytics options and introduce an approach that managers can take to structuring work with analytics-assisted tooling while they await the needed system enhancements.