Where it started
I didn’t set out to build a framework. I was just trying to solve a problem I kept running into — over and over again.
NetSuite reporting is actually pretty good.
It can produce a clean P&L, it’s flexible enough, and SuiteAnalytics is powerful.
But the moment you try to run a business on it — not just report on it — things start to break down.
The problem wasn’t reporting. It was alignment.
What I kept seeing:
- The business runs weekly, but reporting is monthly
- Teams use custom segments, but budgets can’t align to them properly
- Finance produces a P&L, then reshapes it in Excel to make it usable
- The “real” numbers live outside the system
At that point, NetSuite becomes a source of data — not the source of truth.
Accounting periods exist for good reasons. They’re how the books are closed, audited and reported, and NetSuite handles them well. But most of the businesses I work with don’t run month to month. Manufacturers, FMCG businesses, distributors and ecommerce businesses often plan, trade and review performance week by week. So the questions management asks are weekly ones. How did this week go? How does it compare with the same week last year? Are we on track against budget?
Answering those properly takes more than a date on each transaction. It takes a financial calendar the whole business agrees on: weeks that roll up cleanly into periods, structures such as 4-4-5, and comparisons that line up the same fiscal weeks year on year rather than the same calendar dates. Without that shared calendar, each weekly report can end up defining its own week, and the numbers stop agreeing with each other.
The second difficulty is dimension and frequency together. Management rarely wants just a single P&L. They want sales, COGS and margin by site, customer group, channel or product, and they want it weekly, not after the month has closed.
NetSuite holds all of that detail, and SuiteAnalytics can slice it. The hard part is producing it so that it reconciles. The weekly margin by customer group has to add back to the P&L, and the budget has to sit at the same grain as the actuals it’s compared with. In my experience, that’s where the effort goes. Every new cut of the numbers needs the same classifications and business rules applied again, and each one is another chance for two reports to disagree.
“Which sites are driving the result?” sounds like a simple question. Answering it consistently, every week, at the right level of detail and tied back to the ledger, is not.
Why reporting escapes NetSuite
When the system can’t produce the shape management needs, the work doesn’t stop. It moves. Finance exports the P&L and reshapes it in Excel. Someone builds a workbook that maps accounts to the categories the board actually uses. In time, a separate BI environment appears, with its own copy of the data and its own version of the rules.
None of that is unreasonable. Each step solves a real problem for someone working to a deadline. But the business logic ends up scattered across reports, saved searches, workbooks, spreadsheets and people’s heads. Definitions drift. Part of each month-end goes into reconciling the reports with each other rather than understanding the business.
That’s the pattern I kept running into, and it’s the gap the rest of this page is about.
So why not just use Power BI, Snowflake or a warehouse?
So why build reporting capability inside NetSuite at all, when there are excellent analytics tools available?
Power BI is very good at visualisation and self-service analysis. Snowflake and other cloud data warehouses are built to bring together data from many systems at scale. Oracle’s own tools, from SuiteAnalytics to AFP&B and its analytics warehouse, each do their job well. AI is opening up genuinely new ways to ask questions of data. I’m not arguing against any of them, and I’m not suggesting that all financial modelling belongs inside NetSuite.
My point is about what those tools are given to work with. If weeks aren’t defined consistently, budgets don’t sit at the same grain as actuals, and Gross Margin means something different in every workbook, even a very capable tool will reproduce the disagreement rather than resolve it.
ClearSuite’s job is to establish reliable financial structure close to the transactional source: a consistent financial calendar, budgets aligned with actuals, and clearly defined, reconciled financial data. BI tools, warehouses and AI can then consume that data as a dependable input, rather than each rebuilding the business logic for themselves. As I wrote after one project, those tools “become more valuable when they are sitting on top of a consistent business reporting model.”
So I started solving it piece by piece
Not as a product. Just as a way to make things work properly.
1. The Financial Calendar (CSR Weekly Calendar)
First problem:
NetSuite doesn’t really know what a “week” is in a meaningful way.
So we teach it.
A proper financial calendar:
- aligns transactions to weeks, not just periods
- supports 4-4-5 structures
- gives you a consistent base for reporting and comparison
Once everything is aligned to a week, things start to click.
2. CSR Budget Records (CSRBT)
Next problem:
Budgets don’t behave the way operational reporting needs them to.
They don’t:
- handle custom segments well
- align cleanly to weekly views
- reflect how the business actually plans
So instead of fighting native budgets, I introduced a simple idea:
Treat budgets like transactions.
CSRBT lets you:
- model budgets and forecasts at weekly grain
- include any segment or dimension
- align them directly with actuals
Now budget vs actual stops being a workaround — it just works.
3. SpeedLedger
Final problem:
Even when the data is right, it’s still slow and awkward to work with.
SuiteAnalytics is great — if the data is already in the right shape.
So SpeedLedger is just that:
- a curated, flattened financial layer
- structured for speed and clarity
- designed specifically for board-level reporting
It’s not replacing NetSuite.
It’s making NetSuite usable at the level people expect.
What this all leads to
Once those three pieces are in place:
- Weekly reporting becomes natural
- Board packs don’t require Excel reshaping
- Budgets, actuals, and forecasts actually line up
- You can see the business as it runs — not after the fact
This isn’t about adding complexity
It’s actually the opposite.
It’s about removing all the small workarounds that build up over time:
- spreadsheets
- manual reshaping
- inconsistent logic
And replacing them with something consistent and repeatable.
One version of the numbers
Calendars and budgets get the structure right. The next step is agreeing on what the numbers mean.
In one engagement with a food manufacturer, the business had all the usual ingredients of a modern NetSuite reporting stack: SuiteAnalytics, custom financial reports, saved searches, budgets, spreadsheets and board packs. Even so, questions like “Why is EBITDA down?” were surprisingly hard to answer consistently. The data was there. The business logic was scattered.
Two pieces of ClearSuite’s reporting architecture now work together at that site, both inside NetSuite and with no external data warehouse. The CSR Semantic Map is a set of NetSuite records that defines, once, how accounts and classifications map to business concepts such as Net Sales, Gross Margin, Employment Costs, CODB and EBITDA. CSR SpeedLedger uses that mapping as a management ledger, so those concepts mean the same thing in every workbook and dashboard. The result wasn’t more reports. It was fewer arguments about the numbers.
Persisted reporting facts are separate again, and they’re about speed. Take two weeks of sales and COGS by fiscal week, site and customer group, against last year and budget. Once that question has been answered, keeping the resulting reporting facts means the next person doesn’t have to ask the ledger the same question again.
Keeping the meaning with the numbers
Behind every set of management accounts sits a layer of interpretation that rarely gets written down. Which margin are we talking about: invoice margin, manufactured margin, or whatever management has decided to call achieved gross margin? Are sales claims and rebates in or out? Which week is “this week”, and is “down” measured against last week, budget or the same week last year?
Part of that can be captured as structure, and that’s what the CSR Semantic Map does today. At the food manufacturer, it’s implemented as a set of NetSuite records that describe rules such as “this account belongs to Net Sales” once, so reports read the rules instead of each re-implementing them. It’s operational, and it underpins the CSR SpeedLedger implementation there. It’s a foundational part of ClearSuite’s reporting architecture rather than a product in its own right.
The rest is harder to capture. The commentary, explanations and decisions behind the numbers (why a result moved, which exception was agreed and why) tend to live in saved searches, workbook formulas, spreadsheets, scripts written years ago by someone who has since left, and the heads of a few experienced people. That knowledge is how an organisation interprets itself. It’s valuable, and it’s fragile. Keeping that knowledge alongside the numbers, so it survives staff changes and is applied consistently, is a longer-term direction for ClearSuite, not something it offers today.
Where AI fits
Looking further ahead, this is where I think AI-assisted financial analysis could become genuinely useful. I like what NetSuite is doing here. Prompt Studio, Text Enhance and Ask Oracle put AI inside the application, close to the transaction.
But giving AI access to ERP data doesn’t give it access to what the business means by that data. Ask why gross margin is down this week, and the hard part isn’t the reasoning. It’s knowing which margin, which week and which comparison. Without that context, a wrong answer can sound just as convincing as a right one.
What I’d want underneath it is a proper foundation: a shared financial calendar, budgets at the same grain as actuals, governed financial definitions, and reconciled facts that have already been calculated. In that picture, the prompt provides the instruction, the governed definitions provide the meaning, the curated facts provide the evidence, and the language model provides the reasoning and the language.
To be clear, this is a longer-term direction, not a finished product. AI doesn’t just need a ladder into your data. It needs scaffolding.