Shopify Analytics Adds Scatter Plots and Radar Charts to Custom Reports — How Merchants Should Use Them

Table of Contents

1. Key Highlights
2. Introduction
3. What the new chart types add to Shopify Analytics
4. How to use scatter plots effectively in Shopify custom reports
5. How radar charts work and when to use them
6. Practical use cases for merchants, marketers, and operations teams
7. Design and data-preparation best practices
8. Integrating visuals into decision workflows
9. Common pitfalls and how to avoid them
10. Implementation walkthrough: building scatter and radar charts in Shopify
11. Limitations and what to watch for
12. How analysts should validate visual insights before acting
13. Examples and templates you can copy
14. Measuring impact and continuous improvement
15. FAQ

Key Highlights

• Shopify’s custom reports now include scatter plots and radar charts in the visualization picker, enabling direct, interactive pattern detection and multi-metric comparisons without exporting data.
• Scatter plots reveal relationships between two metrics (for example, revenue vs. units sold or sessions vs. conversion rate); radar charts compare several metrics across categories on a single axis after appropriate normalization.
• Applying these visuals effectively requires attention to scale, sample size, labeling, and complementary analyses to avoid misinterpretation and to turn visual patterns into operational decisions.

Introduction

Shopify has expanded its custom-reporting toolkit by adding scatter plots and radar charts to the visualization options available when creating or editing reports. These visuals move analysis out of spreadsheets and into the analytics interface, where merchants, analysts, and marketers can spot clusters, outliers, and multi-dimensional trade-offs at a glance. Used correctly, they accelerate decisions about product assortment, promotions, traffic acquisition, and inventory management. Used carelessly, they can suggest misleading relationships or obscure important context.

This article explains what each chart type reveals, shows concrete examples merchants can replicate, outlines practical design and data-preparation rules, and explains how to combine these graphics with other analyses. The goal is to make these new tools immediately useful for everyday decisions: identifying low-margin heavy movers, comparing category health across multiple KPIs, and prioritizing optimization efforts where they will move the needle.

What the new chart types add to Shopify Analytics

Custom reports have long supported tables, bar charts, and line charts for trends and categorical comparisons. Scatter plots and radar charts add two different analytical capabilities.

• Scatter plots map the relationship between two quantitative variables for each item or group. They show correlation, clusters, and outliers. That helps answer questions such as: which products sell many units but bring little revenue? Which traffic sources deliver high sessions but low conversions?
• Radar charts place multiple metrics on a circular layout, letting you compare a small number of categories across those metrics simultaneously. They excel at profiling categories (product collections, channels, regions) on a compact visual axis that highlights relative strengths and weaknesses.

Both types are not replacements for standard charts; they are complementary. Scatter plots are diagnostic tools for relationships and segmentation. Radar charts condense a handful of performance indicators into a single profile that makes relative trade-offs visible.

How to use scatter plots effectively in Shopify custom reports

Scatter plots pair two numeric metrics (X and Y axes) for each data point. In Shopify, a point can represent a product, a traffic source, a campaign, or a customer cohort. When configured with meaningful metrics they turn raw numbers into decisions.

Practical steps for setting up a scatter plot

1. Choose the entity to plot as a point. Common choices are product, SKU, collection, traffic source, or UTM campaign.
2. Pick two metrics for the axes. Typical X–Y pairs: units sold vs. gross revenue, sessions vs. conversion rate, average order value vs. return rate.
3. Add a size or color encoding where available. Use point size for a third metric (for instance, profit margin or inventory on hand) and color for groups (collections or supplier).
4. Apply filters to remove low-sample noise: require a minimum number of sessions, orders, or transactions before plotting a point.
5. Label or hover-enable key points so stakeholders can identify products and channels without scanning raw tables.

Example: Spotting low-margin heavy movers A mid-size apparel retailer plotted units sold (X axis) against gross profit per unit (Y axis) with point size set to total revenue. Clusters emerged:

• Top-left cluster: many units, low margin. These items drove sales volume but compressed profits. They were candidates for a pricing review and margin-focused promotions.
• Bottom-right cluster: few units, high margin. These represented niche items with strong margins where targeted advertising could scale sales without sacrificing profitability.
• High top-right outliers: best-of-both worlds — high units and high margin — marked as priorities for inventory protection and promotional cross-selling.

On-the-ground steps the retailer took after spotting the clusters:

• Increased minimum advertised price and reviewed supplier terms for low-margin high-volume SKUs.
• Temporarily reduced marketing spend on one low-margin campaign and redirected budget to high-margin, low-volume items to test uplift.
• Flagged high-performing outliers as candidates for merchandising prominence on the homepage.

Interpreting correlation and causation A trend line on a scatter plot shows correlation but does not prove causation. If sessions and conversion rate show a negative correlation, investigate campaign targeting, landing-page mismatch, or sample-size differences rather than assuming traffic quality is the sole driver.

Techniques to improve interpretability

• Add a trend line (linear or local) to summarize the relationship.
• Compute Pearson or Spearman correlation coefficients in a supporting table to quantify association strength.
• Use jitter or small transparency settings to avoid overplotting when many points overlap.
• Apply logarithmic or percent scales for metrics that span orders of magnitude, such as revenue or sessions.

Common analytic traps with scatter plots

• Plotting conversion rate for items with very few sessions. A product with two visits and one order shows 50% conversion — that’s noisy and misleading. Apply minimum session/order thresholds.
• Using raw revenue with products of different price ranges without normalizing for price or unit; plotting revenue and margin per unit together may conflate volume and price.
• Ignoring temporal dynamics. A scatter built on lifetime data may hide recent shifts. If you need to act now, filter to a relevant period (last 30 or 90 days).

Advanced uses: multi-dimensional scatter plots Scatter plots become more powerful when you use a third numeric variable as point size or a categorical variable as color. Examples:

• X: sessions, Y: conversion rate, Size: average order value, Color: traffic channel. That configuration helps prioritize channels that not only convert but deliver higher AOV.
• X: returns rate, Y: gross margin per sale, Color: supplier. That points to suppliers causing costly returns that erode margins.

A/B test postmortem After a pricing experiment, plot items by percent change in units sold (X) versus percent change in revenue (Y). Points falling into quadrants quickly reveal outcomes:

• Upper-right: both units and revenue increased — net wins.
• Lower-right: units up but revenue down — likely margin loss.
• Upper-left: revenue up but units down — possibly higher AOV from price increases.

How radar charts work and when to use them

Radar charts place multiple axes radiating out from a center, each axis representing a metric. A polygon connects the values across axes for a category. Overlaid polygons make comparisons immediate: which categories have balanced performance, which excel on specific metrics, and which lag across the board.

When radar charts produce useful insight

• You want a compact profile of a small number of categories across several metrics. For instance, comparing three or four product categories against metrics like average order value (AOV), conversion rate, return rate, and lifetime value (LTV).
• Relative performance matters more than absolute numbers. Radar charts emphasize shape: a category with a "spiky" polygon is uneven; a broader, fuller polygon indicates balanced strength across metrics.

When to avoid radar charts

• When you must compare many categories (more than five) or many metrics (more than seven). Too many spokes make the chart unreadable.
• When absolute differences are the primary concern. Small differences in distances on a radar can be hard to interpret precisely; bar charts or tables may convey magnitudes better.
• When metrics are on different scales and you have not normalized them. Radar charts require consistent scales across axes or normalization so the shapes truly reflect relative performance.

Normalization and scaling: a must for fair comparison A radar cheat that produces misleading shapes: plotting AOV on a dollar scale alongside conversion rate as a percentage. The AOV axis will dwarf the conversion-rate axis unless both are rescaled. Two common approaches:

• Min-max normalization per metric: scale each metric to a 0–100 range based on observed min and max. This preserves relative position across the dataset.
• Z-score standardization: convert each metric to a standardized score based on mean and standard deviation. Useful when you want to detect which categories are above or below the mean, adjusted for metric variability.

Concrete example: Category health radar A retailer compares three categories—Activewear, Loungewear, and Outerwear—across four metrics: conversion rate, average order value, return rate (inverted so higher is better), and gross margin percentage. After normalizing each metric to 0–100, the radar reveals:

• Activewear: high conversion, moderate AOV, moderate margin, low returns — a balanced polygon that suggests scaling marketing makes sense.
• Loungewear: low conversion, high AOV, low margin, very low returns. The uneven shape shows Loungewear could benefit from targeted funnel improvements to raise conversion.
• Outerwear: moderate conversion, highest AOV, highest margin, high returns. Returns make Outerwear riskier to scale despite healthy revenue.

Handling the inverted metrics For negative metrics (return rate, refund rate, defect rate), invert the axis so higher values on the radar always represent better outcomes. This simplification prevents misreading a strong-looking polygon that actually reflects top-performing negative metrics.

Comparing many categories: small multiples over overlays Overlaying many categories on a single radar chart creates clutter. Two cleaner alternatives:

• Select the top 3–4 categories by revenue or traffic and overlay their polygons for direct comparison.
• Create small-multiples: a grid of identical radar charts, one per category. This retains uniform scaling and enables side-by-side inspection without overlapping shapes.

Color and labeling choices

• Use distinct, moderately saturated colors with transparent fills so overlaps remain visible.
• Label axes clearly with the metric name and units or normalization method (e.g., "AOV (scaled 0–100)").
• Display numeric values on hover for precise comparisons rather than relying on polygon area alone.

Practical use cases for merchants, marketers, and operations teams

Scatter plots and radar charts unlock decisions across the merchant function. Here are actionable scenarios and a short playbook for each.

Merchandising: assortments and promotions

• Use scatter plots to identify SKUs that have high units sold but low profit margin. Move these into clearance or adjust supplier and pricing terms.
• Use radar charts to compare collections across AOV, conversion, return rate, and inventory turnover. Prioritize replenishment and promotional focus on categories that show balanced high scores.

Traffic acquisition and paid channels

• Plot sessions (X) vs. conversion rate (Y) for each channel. Channels with high sessions and low conversion are candidates for landing-page optimization or targeting refinement.
• Use color or point size to add cost per acquisition (CPA) so channels that convert cheaply are readily visible.

Customer segmentation and lifetime value

• Scatter plot cohorts by average order value (X) versus repeat purchase rate (Y). High-AOV, high-repeat cohorts deserve loyalty programs and premium upsells.
• Radar charts to profile segments across LTV, return rate, AOV, and engagement. That helps tailor lifecycle marketing.

Inventory and supplier management

• Map returns rate (Y) against gross margin (X) by supplier to flag relationships where poor product quality eats into margins.
• Radar profiles for each supplier across lead time, defect rate, return rate, and average cost; suppliers with lopsided shapes require contract renegotiation or quality audits.

Customer support and product quality

• Scatter product defects reported (X) vs. return rate (Y) to find products needing design fixes or updated product descriptions.
• Use radar charts for product families to compare defect reports, return rates, warranty claims, and average resolution time.

A/B testing and experimentation

• After an experiment, plot percent change in sessions vs. percent change in conversion by page or campaign. This clarifies whether gains come from traffic shifts or improved conversion.
• Radar charts can profile variants across multiple business metrics for a consolidated view of experiment impact.

Real-world example: optimizing ad spend with a scatter-driven playbook A direct-to-consumer electronics brand used a scatter plot to evaluate paid channels, plotting sessions on the X axis, conversion rate on the Y axis, and bubble size representing average order value. Findings:

• Channel A: very high sessions, low conversion, small AOV (lower-right). The team shifted budget to test new creatives and landing pages for this channel.
• Channel B: modest sessions, high conversion, high AOV (upper-middle). They increased spend on Channel B immediately.
• Channel C: moderate sessions, moderate conversion, but high variance month to month (wide dispersion). The team applied a minimum sample filter and paused budget while further attribution modeling was performed.

Outcome within two months: average CPA fell by 18% and revenue-per-click rose 12% after reallocating spend from unoptimized high-traffic channels to high-conversion, high-AOV channels.

Design and data-preparation best practices

Good visuals start with good data. These guidelines ensure the new chart types deliver reliable, actionable insight.

Set minimum-sample thresholds Low-activity items produce noisy metrics. Require a minimum count (for example, at least 50 sessions or 10 orders) before plotting a point. Shopify’s filtering options let you impose those constraints when building the report.

Normalize appropriately for multi-metric charts For radar charts, normalize metrics to a consistent range. Document the method used (min-max, z-score) in the chart subtitle or axis label so readers understand the scale. If you switch normalization rules across reports, the shapes will not be comparable.

Choose the right time window Match the time window to the decision you want to make. Tactical inventory moves need recent data (14–30 days). Strategic assortment choices can rely on longer windows (90–365 days). Use filtered views to compare trends over different windows.

Label axes and annotate outliers Axes without units confuse readers. Always include units (sessions, %, $) and, where applicable, the normalization method. Add annotations for notable outliers — e.g., “Product X: one-off bulk sale on March 3” — so consumers of the chart know whether to act on the point.

Avoid overplotting and improve clarity

• Use jitter, transparency, or smaller marker sizes when many points overlap.
• Consider binning or aggregating at a higher level (category instead of SKU) when there are hundreds of points.
• Provide hover tooltips with identifying details: product name, SKU, sample counts, and the raw metric values.

Make charts interactive and shareable Leverage Shopify’s interactivity: hover for tooltips, click to filter or drill down, and save the report to a dashboard. Add notes or a summary when sharing with stakeholders to reduce misinterpretation.

Document assumptions and definitions Record metric definitions in a visible place: what counts as conversion, how refunds are treated, and whether returns reduce revenue in the metrics displayed. Different teams interpret metrics differently; a shared glossary avoids disputes.

Test perceptual choices Area perception can mislead: larger polygon area in a radar does not always equal better performance across all metrics. Test charts with stakeholders and adjust labeling and legends to ensure accurate reading.

Integrating visuals into decision workflows

Visuals drive decisions when they are embedded in routine workflows, not one-off analyses. Here are ways to operationalize scatter plots and radar charts.

Weekly merchandising reviews Include a scatter plot on the merchandising dashboard that tracks units vs. margin. Require owners of items in the “high volume, low margin” quadrant to deliver remediation plans within two weeks.

Monthly channel performance meeting Start the channel meeting with a scatter plot of sessions vs. conversion rate colored by CPA. Use it to set optimization experiments for the next month and to reallocate the budget swiftly.

Quarterly product-category strategy Use normalized radar charts to compare categories across profitability, growth, returns, and inventory turnover. Make category owners present a 90-day plan based on the chart profile.

Alerting and automation Pair visuals with alerts: when a product moves into a risky quadrant (e.g., returns above X% and margin below Y%), automatically create a ticket in the product-quality tracker.

Document decisions with the chart snapshot Embed a saved snapshot of a chart in the ticket or strategic memo that records the rationale and the specific metrics that triggered the action. This creates an audit trail connecting insight to action.

Common pitfalls and how to avoid them

These pitfalls recur when teams adopt new visual tools. Recognize and prevent them.

Mistaking correlation for causation A cluster of products with low conversion may coincide with a merchandising change. Hypothesize drivers, then test with experiments or segmented analysis. Use attribution models to check channel effects.

Misinterpreting radar area People often equate larger polygon area with superior performance. Use axis-by-axis labels and, when precise comparisons matter, accompany radar charts with a small table that gives raw or normalized numbers.

Plotting noisy metrics without filters Conversion rates on items with fewer than 10 sessions are dominated by noise. Always impose thresholds or aggregate to a higher level.

Overloading charts with variables Scatter plots should usually be limited to three dimensions (X, Y, and size/color). Radar charts should present no more than 5–7 axes and 2–4 polygons. Otherwise, the reader spends time untangling the chart instead of acting on it.

Using inconsistent normalization across reports If a radar chart uses min-max scaling based on a 90-day window while another uses a 365-day baseline, the shapes are not comparable. Standardize normalization in your reporting playbook.

Failing to consider seasonality A scatter built on one-time promotional spikes can misrepresent typical behavior. Use seasonally adjusted windows or compare to the same period in prior years when seasonality is strong.

Implementation walkthrough: building scatter and radar charts in Shopify

Shopify makes both chart types available in the visualization picker when you create or edit a custom report with compatible data. Here is a step-by-step guide that works as a checklist.

Creating a scatter plot

1. Open Shopify Admin > Analytics > Reports.
2. Click Create report or open an existing custom report.
3. Select the data entity (Products, Channels, Collections, etc.) and choose compatible metrics.
4. In the visualization picker, select Scatter plot.
5. Assign the X and Y axes. Choose a third metric as point size if useful and set a grouping color for categories.
6. Apply filters for minimum sample (orders or sessions), date range, or channel.
7. Add labels and a subtitle describing any normalization or filters.
8. Save and pin the report to a dashboard. Share with stakeholders and add a short note describing the recommended action.

Creating a radar chart

1. Open or create a custom report based on the entity you want to profile (Collections, Categories, Regions).
2. Pick the metrics to compare; pick 3–7 metrics maximum.
3. Normalize metrics where necessary (Shopify may not normalize automatically; perform calculations or create derived metrics if available).
4. In the visualization picker, select Radar chart.
5. Select which categories or groups to render as polygons (top 3–4 by a chosen metric).
6. Set color, transparency, and legend settings.
7. Label axes with units and normalization method.
8. Save, document the normalization method, and share.

If your organization needs consistent presentation, build a templated report and clone it when new periodic reports are needed. Add a short “how-to-read” note inside the report description so cross-functional teams interpret the visuals consistently.

Limitations and what to watch for

The new visuals are valuable, but they have constraints.

Data compatibility Charts work only with compatible data. Some derived metrics or highly customized data models may not be directly plottable and will require pre-aggregation or formula-based fields.

Interpretability for non-analysts Some stakeholders may misread scatter and radar charts. Pair visuals with guided commentary or short annotations that point to the intended meaning and recommended actions.

Chart count and dashboard clutter Resist the urge to put every new chart on the dashboard. Choose a few high-value visuals that answer recurring questions. Too many charts dilute focus and increase cognitive load.

Export and downstream analysis If deeper statistical analysis is required — regression, time-series modeling, or advanced segmentation — export the data to a dedicated analytics environment. Use the Shopify visualization to spot hypotheses and the exported dataset to test them rigorously.

Plan-level differences and product changes Shopify periodically updates features and plan allocations. If a chart type or metric is not available to your account, check plan documentation and Help Center guidance. The visualization picker and its available metrics can evolve; treat saved reports as living artifacts and review them when Shopify updates the analytics interface.

How analysts should validate visual insights before acting

Every visual insight should trigger brief validation steps before operational action.

1. Check sample sizes. Ensure thresholds for sessions or orders are met.
2. Recompute metrics in a supporting table. Numbers in hover tooltips are useful, but an actual table confirms calculations.
3. Segment by relevant dimensions. A poor-performing channel for one campaign may perform well in another.
4. Test with an experiment. If the scatter suggests a pricing change, run a controlled experiment on a subset of SKUs.
5. Monitor after action. Add the affected SKUs or channels to a short dashboard to observe the impact and confirm the hypothesis.

A simple validation framework: ALIVE

• A: Adequate sample size
• L: Look for confounders (promotions, holidays)
• I: Isolate the segment
• V: Verify with a table or export
• E: Experiment or monitor after action

Examples and templates you can copy

Below are repeatable pairings and configurations to try in Shopify. Use them as templates and adapt thresholds to your business size.

Template 1 — Product prioritization

• Entity: Products
• X: Units sold (last 90 days)
• Y: Gross margin per unit
• Size: Total revenue
• Filter: At least 50 sessions and 10 orders
• Use: Identify high-volume, low-margin SKUs to adjust pricing or supplier strategy.

Template 2 — Channel efficiency map

• Entity: Traffic sources
• X: Sessions (last 30 days)
• Y: Conversion rate
• Color: Campaign type (Paid, Organic, Email)
• Size: Average order value
• Filter: Channels with at least 100 sessions
• Use: Reallocate media budget to high-conversion, high-AOV channels.

Template 3 — Category profile radar

• Entity: Collections (choose top 3 by revenue)
• Metrics: Conversion rate, Average order value, Return rate (inverted), Gross margin percentage
• Normalization: Min-max scaling across the selected collection set
• Use: Decide which categories to promote, discount, or optimize for returns.

Template 4 — Post-experiment scatter

• Entity: Product pages
• X: % change in sessions vs. control
• Y: % change in conversion vs. control
• Color: Variant
• Use: Quickly classify experimental outcomes into win/loss quadrants.

Measuring impact and continuous improvement

Adopt a cadence to measure whether decisions triggered by these visuals produce expected business outcomes.

1. Baseline: Before acting, capture the metric baseline for affected SKUs or channels.
2. Action: Implement the change driven by the visual insight.
3. Monitor: Observe the impact over a pre-defined window (14–90 days depending on the action).
4. Evaluate: Compare actual change to expected uplift. Use statistical tests if sample sizes permit.
5. Iterate: Scale successful tactics and rollback or revise unsuccessful ones.

Keep a living log of actions, the visual evidence that motivated them, and their outcomes. This helps refine which visuals drive reliable outcomes and which produce false starts.

FAQ

Q: Where exactly are these new chart types available in Shopify? A: Scatter plots and radar charts appear in the visualization picker when creating or editing a custom report that has compatible numeric metrics. Use Analytics > Reports, then Create report or Edit an existing custom report to access the picker.

Q: What kinds of metrics work best on scatter plots? A: Two quantitative metrics that are logically related work best. Examples include sessions vs. conversion rate, units sold vs. revenue, and return rate vs. gross margin. Add a third metric as point size or use color to categorize groups.

Q: When should I choose a radar chart over a bar or line chart? A: Use radar charts when the objective is to compare a small set of categories across several metrics to reveal strengths and weaknesses in a single compact view. Avoid them for large numbers of categories or when exact magnitudes matter more than relative profiles.

Q: How do I avoid misleading radar charts when metrics have different units? A: Normalize metrics to a common scale before plotting. Min-max scaling (0–100) and z-score standardization are two common approaches. Document the normalization method on the chart for clarity.

Q: Can I export the underlying data for deeper analysis? A: Yes. If you need more advanced statistical modeling or manipulation, export the dataset from the report to a spreadsheet or analytics environment for regression, cohort analysis, or time-series decomposition.

Q: How do I handle low-sample noise in scatter plots? A: Apply minimum thresholds for sessions or orders to filter out noisy points. Consider aggregating data to higher levels (e.g., category instead of SKU) when sample sizes are small.

Q: What are the visual best practices I should enforce across reports? A: Use clear axis labels with units, normalize consistently for multi-metric visuals, limit the number of variables plotted, annotate outliers, and include a short explanation of definitions and filters used.

Q: Will these charts work for historical trend analysis? A: Scatter plots and radar charts are primarily cross-sectional tools. For trends, use line charts or small multiples that show how metrics evolve over time. You can, however, build scatter plots for multiple time windows and compare results.

Q: Where can I learn more about using these charts in Shopify? A: Shopify’s Help Center contains updated documentation on custom reports and visualization types. Consult the Help Center and your internal analytics team for company-specific metrics and reporting standards.

Q: Any final advice for teams adopting these visuals? A: Treat these charts as hypothesis generators. Use them to spot patterns quickly, but validate with supporting tables, filters, and experiments before making permanent operational changes. Keep visual conventions and normalization consistent so reports remain comparable over time.