Waterfall, Lollipop, and Treemap Charts: When the Unusual Chart Is the Right One (2026)

Dashboards & Reports
Tutorial
Updated Sep 02, 2026

Bar charts and line charts handle most of what a survey dashboard needs to communicate, and that's exactly why the less common chart types get overlooked even in the specific situations they're actually built for. A waterfall chart isn't a stylistic alternative to a bar chart - it answers a genuinely different question (how did we get from this number to that one) that a standard bar chart can't answer at all. The same is true of a lollipop chart and a treemap: each one exists because a specific, common reporting need doesn't fit comfortably into the standard chart vocabulary, not because it looks more interesting on a slide.

Table of Contents

  1. Waterfall: Showing the Journey Between Two Numbers
  2. Lollipop: A Long Comparison Without the Visual Weight
  3. Treemap: Size and Hierarchy at Once
  4. Polar Area: A Rounder Alternative to a Bar Comparison
  5. A Note on Exporting These Chart Types
  6. A Worked Example
  7. FAQ

Waterfall: Showing the Journey Between Two Numbers

A standard bar chart comparing a starting satisfaction score to an ending one shows exactly two facts: where you started and where you ended. A waterfall chart shows the same two endpoints, plus every meaningful step in between - detractors who left, passives who converted to promoters, new promoters who joined - each rendered as a small rising or falling bar stacked between the start and end totals. This is the right chart specifically when the path between two numbers is the actual finding, not just the two numbers themselves. A quarter-over-quarter NPS bridge, a churn-and-expansion revenue walk, or a response-funnel drop-off from invited to completed are all classic waterfall use cases, precisely because in each of them, understanding which specific factors contributed to the net change matters more than the net change alone.

Lollipop: A Long Comparison Without the Visual Weight

A lollipop chart is, functionally, a bar chart with the solid bar replaced by a thin line and a single dot at the end - the same comparison, communicated with a fraction of the visual ink. This matters most for a long list of categories being compared on one axis: fifteen or twenty solid bars sitting side by side start to visually compete with each other and read as a dense, heavy block, while the same fifteen or twenty thin lines with end-point dots stay visually light and let a reader's eye track the pattern of dot heights (or positions, for a horizontal version) without the chart feeling cluttered. It's a particularly good fit for a long ranked list - departments by engagement score, product features by adoption rate - where a standard bar chart's solid fills would otherwise dominate the page.

Treemap: Size and Hierarchy at Once

A treemap divides a rectangle into nested, proportionally-sized blocks, where each block's area represents its share of the whole, and blocks can be grouped into larger regions representing a parent category. This solves a specific problem neither a bar chart nor a pie chart handles well: showing both the relative size of many categories and their place within a hierarchy in a single view. A treemap breaking down support tickets by product area, and within each product area by specific issue type, shows at a glance both which product area generates the most volume and which specific issues within that area are driving it - a combination a flat bar chart would need two separate charts to convey. Treemaps work best with a moderate number of categories at each hierarchy level; too many thin slivers at the bottom level become as hard to read as an overcrowded pie chart, for the same underlying reason.

Polar Area: A Rounder Alternative to a Bar Comparison

A polar area chart arranges categories radially around a center point, with each category's value represented by how far its wedge extends outward - visually closer to a pie chart than a bar chart, but unlike a pie chart, the angle of each wedge stays equal while the length varies with the value, which keeps the more perceptually reliable length comparison intact rather than relying on angle judgment. It suits a moderate number of categories (rarely more than eight or so, for the same crowding reasons a pie chart runs into) where a circular, less clinical presentation fits the context better than a straight bar comparison - a brand or design-feedback summary presented alongside other more visual, less spreadsheet-like content, for instance.

A Note on Exporting These Chart Types

Worth knowing before building a dashboard around these chart types specifically: when exporting a dashboard to PowerPoint, waterfall, treemap, polar area, and lollipop charts export as a picture of the chart rather than as a fully editable, native PowerPoint chart object, unlike standard bar, column, line, scatter, doughnut, and pie charts, which export as editable objects. This doesn't affect how the chart looks or reads in the exported file, but it does mean a colleague working from the exported PowerPoint won't be able to click into one of these specific chart types afterward and adjust its data or formatting the way they could with a standard chart - worth factoring in if the exported deck is likely to be edited further downstream by someone other than the original builder.

A Worked Example

A subscription business wants to explain a quarter's net revenue change to its board - not just that revenue grew 8%, but why. A waterfall chart starting at last quarter's total, with individual bars for new customer revenue, expansion revenue from existing customers, and revenue lost to churn and downgrades, shows the board that gross new revenue was actually strong, but a meaningful churn bar is eating into most of the net gain - a story a single "revenue went from X to Y" bar chart would have completely flattened. The same deck also includes a lollipop chart ranking twenty individual product features by adoption rate, chosen specifically because a bar chart of the same twenty features felt visually heavy and made the pattern harder to scan, and a treemap showing support ticket volume by product area and issue type, letting the team see at a glance which specific issue was driving the largest single ticket category without needing two separate charts to show it.

FAQ

When should I use a waterfall chart instead of just showing before-and-after bars?
Whenever the individual contributing factors between the two numbers are part of the story you're telling - if only the net change matters and nothing about the path there, a simple before-and-after comparison is simpler and clearer.

Is a lollipop chart just a stylistic choice, or does it actually communicate differently than a bar chart?
Functionally the same comparison, so it's largely about reducing visual weight for long lists - the more categories you're comparing, the more a lollipop chart's lighter visual footprint helps versus a solid bar chart, which can start to feel heavy and dense past ten or so categories.

How many levels of hierarchy can a treemap reasonably show?
Two levels (a parent category and its sub-categories) is the practical sweet spot for most dashboard use - beyond that, the nested blocks get too small to label or read clearly, and a treemap starts losing its main advantage over a simpler chart.

Does the PowerPoint image-export limitation mean I should avoid these chart types?
Not necessarily - it only matters if the exported file specifically needs to be edited further by someone else after export. For a chart that's being presented as-is, the image-based export looks identical to a live chart and communicates the data just as effectively.


For more on choosing the right chart type generally, see How to Choose the Right Chart for Your Dashboard.

waterfall chart survey data lollipop chart dashboard treemap chart use case unusual chart types

Related Articles

Getting a Dashboard Presentation-Ready: What PowerPoint Export Does and Doesn't Handle (2026)

Exporting a dashboard straight to PowerPoint saves the hours normally lost to rebuilding every chart by hand - as long as you know which charts come out as fully editable objects and which come out as flat images, and design with that distinction in mind from the start rather than discovering it after the export is already in front of a client. This guide covers what actually happens during export, and how to structure a dashboard so the exported version needs the least possible cleanup.

What a Dashboard Can't Tell You (2026)

A dashboard is very good at telling you that something happened - a score moved, a segment underperforms, a category grew. It's structurally bad at telling you why, and it's easy to mistake a dashboard full of confident-looking charts for a complete picture when it's really only ever answering half the question. This guide covers the specific kinds of understanding a quantitative dashboard can't produce on its own, no matter how well it's built.

Why the Same Dashboard Tells Two People Different Stories (2026)

Two people can look at the exact same chart, built from the exact same data, and walk away with genuinely different conclusions - not because either one is reading it carelessly, but because where their eye landed first, what number they happened to compare it against, and what they already expected to see all quietly shaped the interpretation before any conscious analysis started. This guide covers the specific, well-documented ways a dashboard's layout and framing bend how it gets read, independent of the data itself.

From Static Report to Live Dashboard: What Changes in How You Design It (2026)

A PDF report is finished the moment it's sent - a fixed artifact, true at one point in time, never quietly out of date because it never changes at all. A live dashboard is never finished in that sense, and designing one like a frozen document is a common, avoidable mistake. This guide covers what actually needs to change in how you write headlines, structure navigation, and think about an exported snapshot once the thing you're building keeps moving after you publish it.

Comparing Survey Waves on a Dashboard Built for One Survey at a Time (2026)

A dashboard is naturally built around a single survey, and a tracked metric doesn't live in a single survey - it lives across several, one per wave. That mismatch is easy to miss until the second wave comes in and there's no obvious built-in place for a wave-over-wave line to live. This guide covers practical, honest ways to compare waves anyway: what to build manually, what to note by hand, and where to draw the line before manual workarounds cost more time than they save.

We value your privacy

We use cookies and similar technologies to improve your experience, analyze site traffic, and personalize content. Learn more