The Joy of Business Reporting
It’s a great way to grow and learn!
Background
In my last article, I shared that there are 2 commonly perceived “necessary evils” in data analytics work — campaign management and business reporting (sometimes shorthanded as MIS in the financial services industry). Necessary evils because it’s the kind of work that data analysts generally don’t enjoy doing and there’s not much “glory” to be gained from it. However, in that said last article, I posited a different perspective on how campaign management can be inverted to become a harvesting process for knowledge and competitive intelligence. In this week’s article, I want to similarly propose an alternative viewpoint on the other “evil” — business reporting.
Business reporting falls under the broader practice domain of business intelligence. The focus of this class of activities is generally about efficiency — the ability to support real-time reporting, automation, self-service, interactive visuals, etc. Data analysts and data scientists tend to think of business reporting as “delivery” work wherein the business stakeholders define the requirements and the data analysts write the codes to get the job done. Not much thinking required other than to check the availability of data. But this perspective is self-inflicted; it need not be the case.
And so I dedicate my 44th weekly article to providing an alternative perspective on the role of business reporting in the context of the broader discipline of data analytics / data science.
(I write a weekly series of articles where I call out bad thinking and bad practices in data analytics / data science which you can find here.)
Great Way to Learn
Now truth be told, I actually enjoy defining and creating new business reports. Maybe there are others who also do, but I suspect it’s the minority view. To me, it’s a great way to learn. Particularly if you are a junior data analyst, but even if you are a senior one. The activity of business reporting allows you to acquire the following knowledge and competencies:
- Deep domain-specific knowledge
- Generating information signals
- Communicating with data
Let’s unpack each of them.
Domain-Specific Knowledge
Creating a business report is the fastest way to learn about the business / domain. It’s a fantastic knowledge ingestion opportunity. Most data analysts simply execute blindly the requirements from the business stakeholders without challenging or refining it. A good business report should describe the business unit from its strategy to its execution to its outcome.
When a stakeholder request comes in to develop a business report, do you have the ability to gauge if the required metrics are lopsided? For example, if the requirements only focus on activities (i.e. execution) and lack sufficient outcomes. Do you ask the stakeholders about their business or domain strategic objectives, and then review to see if the requested business metrics align with them? Can you recognise alignment / mis-alignment? Are you aware if there are other definition variations of the requested business metrics that other units may already be producing and utilising, and if so, do you know what drives these variations? Perhaps there are already standard definitions (both industry-wide and enterprise-wide) that should be utilised. These simple acts of “curiosity” can pay significant dividends over time, and you’ll be surprised how much domain knowledge you can acquire.
Information Signal Generation
Business metrics are analogous to information signals. They mean something; they have specific interpretations and decisioning utility. The creation of business metrics is the art and science of extracting and amplifying information signals from data. Business metrics are not raw data points. Rather, they are constructed and composite data. Sometimes you see the term transformed variable used instead. In the equivalent world of data science / predictive modelling, you can think of these transformed variables as feature engineering.
You obviously need strong domain knowledge to create the right information signals. When a stakeholder request on business reporting comes through, do you spend time thinking how the various business metrics (that is being requested) triangulate to amplify a desired information signal? Do you spend time thinking whether the desired information signal could have multiple interpretations (i.e. equivocal) thus leading to confusion?
Most business reports are time series oriented, capturing trends and changes over time. Time series data are a treasure trove for understanding leading versus lagging indicators of business outcomes. Do you spend time thinking whether applying smoothing techniques such as “rolling 3 months average” can reduce noise? Do you spend time constructing various ratios to catch outlying exceptions or to better reflect momentum and velocity?
Data Communication
And because you have to assemble the metrics together, business reporting presents a great opportunity to learn about and practice data storytelling. Most business reports are tabular in nature, and very likely in Excel or Excel-like formats. As information density increases, there is a need to present some of this information in chart forms. I’ve written here before about data visualisation and dashboarding. It is a thoughtful exercise. All that good work on creating appropriate information signals can go to waste simply because we fail to consider how the audience will perceive, interact and digest it.
To start off, do you consider if there is a natural hierarchy to how the metrics should be presented? Which comes first and why? Do you consider if an executive summary section is needed, and if so, should it be presented solely as visual charts or complemented by it?
Conclusion
All organisations produce more business reports than they can consume. But they struggle to reduce their business reports for fear of missing out. And this is one of the reasons why many data analysts gripe about the mundanity of producing business reports — it feels like a thankless job. But the same data analyst can make a choice to view business reporting as an extraordinary opportunity to learn and practice. Rather than just blindly scripting and publishing those reports, they can help their stakeholders focus their attention on changes in the information signals that can represent emerging risks or emerging opportunities. This is one also why I’m an anti- advocate for offshoring data analytics work. The lack of understanding of the domain coupled with the lack of engagement with the information only fuels the perception that business reporting is low-value work. It need not be.
