Two scenes.
Scene one. A woman walks her dog early on a Tuesday morning. She passes a row of closed shops, an empty car park, and two men standing outside a café. Nothing about that scene is unusual.
Scene two. It is 3am on a Friday night. The same woman, the same dog, the same empty car park - except now the shops are shut with shutters down, and the two men are the only other people visible. The data is almost identical. The context is completely different.
The same numbers, different meaning. This is the central challenge of working with data - and it applies just as directly to financial reporting and tax compliance as it does to a late-night street.
Data beyond face value
Finance and tax professionals are trained to work with numbers. Balance sheet totals, tax rates, effective rates, variances against prior year. These numbers feel objective.
They are not. Every number in a financial statement is the output of decisions: accounting policies, allocation methods, cut-off judgements, adjustments, estimates. The number is real. But what it represents - and whether it is the right representation - depends entirely on the context in which it was produced.
Understanding data beyond its face value means asking: where did this come from? What decisions produced it? What would change if those decisions were made differently?
The data journey: from origin to insight
A trial balance number has a history. It started as a transaction - a payment, a receipt, an accrual. That transaction was coded to an account, by someone, using a chart of accounts that reflects decisions made (often years ago) about how to categorise the business’s activity. The transaction was processed through a period close, subjected to adjustments, and emerged as a line in a TB that was then extracted, mapped, and loaded into a tax platform.
By the time a tax analyst sees that number, it has been through multiple transformation steps - each of which introduced potential for error, inconsistency, or misrepresentation.
This is not an argument for distrust. It is an argument for understanding provenance. Data you understand is data you can rely on. Data whose journey you cannot reconstruct is data whose reliability you cannot assess.
The risks of bias
Data can be manipulated - deliberately or unconsciously - to support a preferred conclusion. The most common manipulations are not dramatic; they are subtle choices about presentation that shift the reader’s perception without technically misrepresenting the numbers.
Consider a Bitcoin price chart. Show the price from January 2021 to July 2021 - a dramatic rise followed by a sharp fall. The story is volatility and risk. Now show only March to May 2021 - the same data, a different excerpt - and the story is a period of strong growth.
Neither chart contains false numbers. Both are designed to produce a specific impression.
In financial reporting, similar effects arise from:
- Axis manipulation - starting a chart axis at a value other than zero, making small changes look dramatic
- Period selection - choosing start and end dates that flatter the trend
- Metric choice - selecting the measure that presents results most favourably from among several legitimate alternatives
None of these are necessarily dishonest. But they are reasons to interrogate the data you receive, and to be thoughtful about the data you present.
The importance of live data
There is a difference between data that describes a situation and data that describes a situation as it is now.
An apple bought last week and an apple bought this morning are the same fruit - but one is worth eating and one is not. The data on the label is identical. The underlying reality is not.
In tax compliance, the equivalent is a trial balance that was finalised two months ago versus one that reflects yesterday’s close. A provision based on an outdated TB is a provision built on stale data. The numbers may look clean; the picture they paint may not be current.
Working with live, up-to-date data requires systems that can produce it efficiently - which is one of the strongest arguments for automation and real-time data connections between ERP, tax platform, and reporting tool. The less time it takes to extract and process data, the closer to current your analysis can be.
The data story
The best financial analysis does not just report numbers. It tells a coherent story about what those numbers mean, where they came from, what they compare to, and what they imply.
That story requires context: prior-year comparatives, budget versus actual, entity-level drill-down, macro-level narrative. Without it, a reader is left with numbers that could mean many things - and will likely interpret them through their own assumptions.
This is as true for a note in a statutory account as it is for a tax provision review. The number is the beginning of the communication, not the end.
Conclusion
Data literacy is not about technical skill. It is about asking the right questions: Where did this come from? What decisions shaped it? What would change if I looked at it differently? What am I being invited to conclude, and is that the right conclusion?
These questions are as relevant in a year-end review meeting as they are in a data analytics context. The numbers in front of you are a representation of something - and understanding that representation more deeply is one of the most valuable skills in finance and tax.
If you would like to discuss how better data flows and more transparent processes could improve the quality of information in your tax and finance function, get in touch.
