After several webinars, articles and polls on business analysis techniques, I’ve collected many responses from fellow BA’s about their favourites. I like to share information visually – and this time a word cloud seemed appropriate:
While process modelling and asking questions is often front and center when we talk business analysis, it was encouraging to see data analysis mentioned more frequently.
What is data analysis? The definitions will vary depending on the context and the field (statistics, scientific research, business analytics, management analysis etc.). Let’s start with these two:
“Data analysis is the process of transforming raw data into usable information” (OECD)
How can data analysis be useful for business analysis activities?
Understanding business data
A business analyst needs to understand the business, the context of the business problem, essential business processes and applications that support them. Part of this parcel is understanding business data: what are the key data entities, what information does the company generate, use and share? Exploring business data can be an inexpensive way to discover more information and build up the domain knowledge. What are possible sources for data exploration?
- Business reports and operational dashboards – including drilling into detail records. What are the attributes of the records? Values, codes, ranges? Anything surprising, unexpected? What can you learn about day-to-day business from the reports?
- Key Performance Indicators – what is measured? What is reported to shareholders and market analysts? Review your company’s annual report to see what is important.
- Data statistics and profiling – means, averages, histograms for key attributes and metrics – this may be available through your company’s BI solutions.
- Data dictionary or glossary – will provide definitions, types, examples, and cross-references to establish relationships between data entities.
Conceptual data modelling
The next step after looking at separate business entities is understanding the relationships among them. Not all business analysts know how to create simple conceptual data models, or even accept that it is relevant to their job – but those analysts who know how to do it will have a definite advantage. Any business analyst who is interested in developing their career in the architecture direction should be able to create structural models, and a conceptual data model is a good practice.
Targeted scenario analysis
When a business analyst needs very specific information, data analysis may answer some questions more efficiently than any amount of interviewing and elicitation. If you want to confirm whether a particular exceptional scenario is possible, what the range of a certain variable is, or what business events to expect in special circumstances, actual data is your friend. The subject matter experts may disagree, may not have a complete answer, or only see one side of the issue, but the data, if used smartly, can tell you a complete story.
When I worked with my stakeholders on decision matrices, business rules, or possible flow variations, I was usually able to get the few remaining infrequent scenarios from data. Had I simply listened to the expert opinions, we would have missed them, since we are all human and tend to forget obscure facts and dismiss exceptions as “flukes”.
Data quality issue detection
Many companies struggle with bad data due to lack of data governance, cumbersome or non-existent legacy integrations, and years of skimping on proper data architecture. Business analysts often find themselves on the receiving end – discovering that they need to capture a lot of requirements just to deal with bad data. From handling data exceptions to rejected batch uploads to rolling back month-end reports and regulatory filings, bad data can cause a ripple effect through many business processes.
Applying the business analyst mindset, it’s usually a good idea to look for the root cause of the data issues, rather than coming up with more band-aid requirements. Investigating where and how bad data is generated can be a better use of analyst’s (and developer’s, and tester’s) time. Analyzing reports, queries, data exploration, test results and defect logs can help detect the source of the issues – and lead to upstream changes to minimize bad data creation.
A business analyst who can run a basic query or two, or explore data in the BI “sandbox”, or just be curious enough to get someone else to help with that, would be on the right track to help improve data quality in their company.
Business process improvements
Most projects involve some business process changes or automation of tasks. Experienced business analysts know the importance of improving processes before automating them – automating a suboptimal process is a waste of resources.
Identifying opportunities for process improvements often requires data. Analyzing process performance statistics at specific points allows to detect bottlenecks, delays, excessive exceptions, and scenarios that require additional processing. While not every company can afford the tools for automated process mining, a business analysts with some data analysis skills can help their client identify opportunities for improving process effectiveness.
Project cost-benefit analysis
Smart organizations engage their senior business analysts in cost-benefit analysis to determine feasibility of changes. This requires facts, metrics and analytics. What is the average time required to process a case in current state? How many events of each type are processed every month? When are the peak times and what is the peak load? How much revenue is lost when defective products are sold, or services rejected by customers? How many messages are sent and received that could be eliminated with automation? What is the average wait call time while a customer service representative is manually searching for customer order information? I’m sure you can add many more examples.
These are relatively simple analytics to gather when data is available. A hard look at these metrics can save companies a lot of money by prioritizing changes based on the expected ROI (return on investment), instead of a gut feeling.
These scenarios by no means cover all possible applications of data analysis in business analysis activities. Consider the advantages you can get from data in various project situations and test it out – and then add your examples to the list.
Not everything can be resolved with data. However, a typical analyst is more likely to under-use the opportunities afforded by smart data analysis than use it where not appropriate.