Excel revolutionised accounting – but that revolution is now decades old. In the age of AI and big data, a new tool has entered the scene: DuckDB. For digitally fluent accountants, DuckDB is not just a lightweight database – it is what Excel should have evolved into. For those straddling the worlds of finance and data, this is the moment we have been waiting for.
When I first started out in Finance, there was one undisputed hero: Excel. We pushed it to its limits, macros, array formulas, custom-built VBA scripts, and then some. We used it for financial modelling, data cleansing, forecasting, even as a poor man’s database. We demanded more power, more rows, more reliability. Eventually, Microsoft responded.
Power Query and Power Pivot were born-not as replacements for Excel, but as bolt-ons to meet the growing hunger for more robust data processing within the familiar spreadsheet interface. We wanted Excel to do more, and Microsoft delivered. But while these tools were revolutionary, they were still built around the same core idea: a spreadsheet interface.
Now, years later, we are in the age of big data and AI. And for me, the tool that feels like Excel’s true successor is not even from Microsoft. It is DuckDB.
Why DuckDB Is the Next Logical Step
DuckDB is a fast, in-process SQL OLAP database. It reads Parquet and CSV files directly, supports complex SQL queries, and integrates with tools like Python and R-but what is game-changing is that it runs entirely on your local machine with no setup, no cloud dependency, and no IT gatekeeping.
If DuckDB had existed back during my early finance days, I probably would not have jumped ship to IT and gone deep into data engineering. That leap happened mostly because I outgrew the tools available to me as a finance professional. I needed versioning, scale, traceability, and the ability to work with more than a million rows something Excel simply could not provide.
A Foot in Both Worlds – Until Now
These days, I am fully immersed in the tech world. I don’t straddle accounting anymore. But I look back and see finance professionals today doing what data analysts are doing, data wrangling, reconciliations, KPI tracking-all centred around large, complex datasets. The only difference? Finance folks are still doing it in Excel. Imagine reconciling intercompany transactions from two 5-million-row CSVs, a task that would crash Excel but is a simple join query in DuckDB.
And honestly, that no longer makes sense.
In a big data world, relying on Excel’s row limits or multi-tab nightmares is like using a flip phone in the age of smartphones. You can do it, but you are making life harder for yourself.
Power Query Is Great-But It’s Not a Database
I have enormous respect for Power Query. It is one of the best data ingestion and transformation tools available, especially for non-coders. But – and this is key – it is not a database. It does not store data. It does not manage data in a structured, ACID-compliant way. You cannot write data back to it in a controlled manner. For finance professionals who want to *touch* the data-not just view it, but append it, update it, reshape it – this is a limitation.
DuckDB changes that. It is not just read-only. You can write data back. You can insert, update, delete, and store data in tables, just like a real database. It supports transactions, joins, window functions, and even local Delta Lake-style workflows using Parquet files. That is not just a technical win – it is an emotional one. Accountants like to own the numbers. DuckDB lets them do that, at scale.
Why DuckDB Beats Microsoft Access for Today’s Needs
Some might ask – what about Microsoft Access? Wasn’t it supposed to be the “easy” database for the non-technical user?
Yes, Access had its moment, particularly with form-building and reporting on top of internal data. But its limitations – scalability, poor support for large files or concurrent users, and dated architecture – make it feel out of place in today’s data-driven world.
DuckDB, on the other hand, is open source, cross-platform, and stunningly lightweight. Its `.db` file is small enough to email – just like an Excel file. And unlike Access, DuckDB supports modern analytics workloads with dozens of powerful built-in functions and native support for open data formats like Parquet.
It’s Simple Enough to Start Today
Here is what working with DuckDB can look like:
“`sql
SELECT CustomerID, SUM(InvoiceAmount)
FROM ‘invoices.parquet’
GROUP BY CustomerID;
“`
That is it. No VBA. No Power Query GUI. Just clean, understandable SQL.
And even if you have never written SQL before, AI copilots like ChatGPT or GitHub Copilot in VS Code can help you write queries just by describing what you want. The learning curve has never been lower. Getting started is as simple as visiting the DuckDB website and following their 5-minute install guide for Python or the command line. Or even better: ask IT to create an empty lightweight DuckDB .db file for you (which you can save and reuse as a template), install an IDE like DBeaver, and that is it you’re ready to start ‘database-ing’.
Safe by Design: Why IT Should Love DuckDB Too
One of DuckDB’s biggest hidden advantages is its isolation. It does not touch production systems. It does not connect to the company-wide SQL Server. And that is exactly why it is safe.
By running locally and only interacting with the user’s own `.db` file or flat files, DuckDB avoids the typical “oops” moments that make IT nervous – no accidental deletes from a shared database, no broken relationships, no unintended orphan records. Your changes affect *your* file, and nothing more.
This design does not just protect enterprise data – it empowers users to experiment freely without waiting on database permissions or fearing mistakes.
And since DuckDB files are tiny-just like Excel workbooks – versioning is simple. Tools like SharePoint or OneDrive can track every saved version of your `.db` file. Need to roll back to yesterday’s state? Just click “Restore previous version.” It is audit-ready without needing DevOps.
Separating Logic from Data: SQL’s Hidden Strength
One of the most overlooked benefits of working with SQL instead of Excel is the clear separation between logic and data. In SQL, your transformation and aggregation logic lives in a separate `.sql` file. The data lives in a database or flat file. That makes audits much simpler, you can just send someone the SQL script and the data source. If the results differ, it is a data issue, not a logic mystery.
This avoids the classic Excel scenario: ‘What the heck is wrong with this formula?’ Only to discover someone accidentally deleted a row, broke a reference, or altered cell A1. SQL removes those risks. If you are querying read-only data from a data warehouse, you know no one is altering the source, and your logic is consistent every time.
Even better, versioning becomes effortless. Need to roll back to a previous calculation method? Just open the old `.sql` file. These files are tiny, just a few kilobytes, so no more waiting for massive Excel files to load just to check a logic tweak.
The Rise of the Digital Accountant
Let’s be clear: in 2025, digital skills are no longer a “nice-to-have” for accountants. They are essential. AI can now generate reconciliations, explain SQL code, and automate tedious processes. The differentiator is not who can use Excel faster. It is who can work with data at scale, and who can automate tedious workflows.
DuckDB gives accountants a serious seat at the data table-without needing to become full-fledged engineers. It lowers the barrier to entry into code-driven analytics while providing the robustness that modern finance work demands.
And best of all, the skills you gain in DuckDB do not just live inside one tool – they transfer. To PostgreSQL. To Snowflake. To SQL Server. To Databricks. You are not learning in isolation – you are building a foundation for long-term growth.
Final Thought
If Excel is like driving a car with manual gears, DuckDB is like piloting a Tesla on autopilot. You are still in control – but now you can go further, faster, and safer.
So, if you are an accountant tired of stretching Excel to its breaking point, give DuckDB a try. You might just fall in love with data again.
About the Author
Maha Kepakisan, CPA, is a former accountant turned data engineering leader with over a decade of experience delivering enterprise-scale data solutions across industries including finance, telecom, insurance, and marketing technology. Currently Analytics and Engineering Manager at Ritchies Transport, Maha has led large cloud migrations, built scalable architectures, and developed analytics platforms used by some of New Zealand’s most recognisable organisations. He holds multiple Microsoft certifications and is passionate about helping finance and business teams modernise their data workflows through automation, strong governance, and a citizen-led approach.
