Datama Webinar: Break Down Your Margin Live in Excel with Waterfall & KPI Tree

No export, no manual rebuild, no tool switch: Datama turns your Excel range into a board-ready margin variance analysis.

That's the topic of Datama's latest webinar, hosted by Guillaume, Datama co-founder and a former strategy consultant at Bain & Company, where he built "margin analyses by the dozen in Excel." The full replay is embedded below.

Watch video

The problem: margin variance is a tedious exercise in Excel

The scenario is familiar to any controller, CFO or business unit lead: a pivot table wired to the transactional database, with volumes, revenue, costs and margin, sliced by year, business unit, country, category and sub-category.

To understand why margin moved between two periods, the analyst drags and drops dimension after dimension (split by country, then by business unit, then by category), trying each one until something tells a story. Then comes isolating the price effect, the cost effect, the volume effect, and formatting all of it for the executive committee. The result: sprawling Excel files that turn into gas plants, rebuilt from scratch at every closing.

How Datama automates the analysis, right inside Excel

Datama is an add-in that installs in one click across most of your data and AI tools: BI tools (Tableau, Power BI, Qlik, Looker Studio), spreadsheets (Excel, Google Sheets) and LLMs (Claude, ChatGPT). In Excel, the add-in is called Waterfall and KPI Tree and lives on the Microsoft AppSource marketplace. The install is Microsoft-certified: your data never leaves your Excel environment, your own machine runs the calculation.

Once the add-in is added, you paste a license key and select the source data range. Datama then builds the analysis automatically.

The market equation: decomposing your margin into Volume × (Price − Cost)

The first step is building what's called a market equation: the decomposition of a macro-indicator into sub-indicators. For a margin, that looks like:

Margin = Volume × (Price − Unit cost)

Where average price and average unit cost are themselves computed by dividing revenue and costs by quantities. This equation can be as simple or as granular as you need: you could split unit cost into variable and fixed cost per unit, or group several steps into one using sub-steps. Datama also lets you add icons to each step, turning the equation into an actual decision tree that's easier to share with internal stakeholders.

A dynamic waterfall, automatically scored by dimension

This is where Datama goes beyond a manually built pivot table. Once the equation is set, you just indicate what to compare (year N vs N-1, or actuals vs a forecast or a target) to get a waterfall, a KPI tree or a tabular view.

Datama then screens every available dimension and computes an interest score for each one, telling you which is actually worth digging into. In the webinar demo, total margin grows 54% (from €685K to €1M):

• Splitting by category is a dead end: home and food grow almost identically (+53% and +55%), simply reflecting their weight in the mix.
• Splitting by business unit is the right move: nearly all the growth comes from affiliates.

The waterfall stays clickable at every step: clicking on the "affiliates" segment re-runs the same scoring inside that segment, and this time reveals that volume, not country, is the driver (every country grows comparably among affiliates), instantly pointing the business conversation at the right question.

The mix effect, finally separated from the price effect

Mix effects are one of the most time-consuming, and most poorly handled, parts of a manual margin analysis. They're often invisible, hidden behind the price effect. Datama separates them automatically: behind your price effect sits a mix effect and a performance effect.

In the webinar example, the affiliates' share of the mix moves from 34% to 57% (+16 points). Since affiliates sell at a higher average price than the overall average, that mix shift alone mechanically pulls the average price up, without any actual price change. Some call this Simpson's paradox: selling more of a pricier product raises the average price, even if you actually lowered prices. Without this decomposition, a finance team could wrongly conclude there was a real pricing increase.

From dynamic Excel to boardroom-ready slides

Every waterfall Datama generates ships with a title, a chart and auto-written commentary (scoped only to the dimensions worth mentioning), ready to present to the board. And the analysis doesn't stay locked in Excel: the HTML Viewer by Datama extension for PowerPoint lets you import the file generated in Excel straight into a slide.

The result stays dynamic in PowerPoint, in edit mode or presentation mode. You can switch the comparison live in front of your executive committee (say, from "vs N-1" to "vs Target") and watch the chart and commentary recompute instantly, without going back to Excel. Datama can even build double comparisons within a single variation.

On formatting, nothing is locked: colors, where commentary sits relative to the chart, rounded bars; everything is configurable, in Excel and in PowerPoint alike, and stays reactive to the underlying data. Filter down to a single country, for instance, and Datama fully recalculates the analysis (drivers included) on that scope.

Test your hypotheses live (what-if)

Because everything stays inside Excel, you keep the full logic of a normal spreadsheet. You can build a simulation cell (say "+10% on furniture margin"), reference it in your source data, and watch Datama's waterfall and commentary recompute live as you tune the assumption.

The same analysis engine is available as a Power BI extension: same methodology, same waterfall rendering, directly inside your reporting dashboards.

What about just asking an LLM?

The obvious instinct today is to ask an AI assistant the same question. The webinar tests exactly that with the Claude for Excel add-in (Opus): the answer is verbose, and wrong. The AI attributes the increase to France 100%, corrects itself toward Spain once challenged, then eventually lands on the right driver (affiliate volume). Ask it for a waterfall chart and it produces one, but focused on the wrong axis (countries instead of affiliates), not dynamic, and impossible to render natively inside Excel.

This isn't a model problem, it's a method problem: an LLM generates text, it doesn't compute deterministically. Datama stays complementary to AI tools rather than competing with them. The Datama skill for Claude and ChatGPT lets your AI assistant lean on the same reliable calculation engine instead of reconstructing the analysis its own way. And just like in Excel, your data stays in your environment.

Who is this for?

• Financial controllers preparing the monthly margin report
• FP&A and finance business partner teams who need to explain a variance to the business
• Anyone still presenting their waterfalls as hand-rebuilt PowerPoint slides at every closing

What's coming next

Several features are coming to the Excel add-in: right-click what-if simulations, annotations showing the differences between two charts, "nested" market equations grouping several steps into one (for example price and unit cost merged into a margin rate), and automated export of Datama's calculation tables into a dedicated tab from a simple cell range. Deeper AI integration is also planned, so assistants like Claude can tap directly into the Datama engine from inside Excel.

How to get started

The Waterfall and KPI Tree add-in is free on the Microsoft AppSource marketplace, no account required, with Datama branding and no ability to save your use cases. A 15-day free trial then unlocks saved analyses and removes the branding, with Datama's team available to support a POC.

→ Install the add-in: Datama on Microsoft AppSource → Learn more: datama.io

About Datama

Datama is a variance decomposition platform used by finance, product and data teams in retail, travel and financial services. The Excel add-in, the PowerPoint extension and the Power BI extension bring the same deterministic methodology to every tool your team already uses.