The Results page for each Shelf Test question shows where respondents clicked on the shelf, how they shopped overall, and how each product performed. Use these views to compare products, understand shopping behavior, and export the data for deeper analysis.
Want a guided walkthrough? Take the course in Lighthouse Academy.
In This Article
- Viewing the Heatmap
- Analyzing Shelf Summary Statistics
- Analyzing Product Metrics
- Who Each Metric Is Based On
- Exporting the Results
Viewing the Heatmap
The heatmap shows where respondents clicked on the shelf. A click is any click on a screen location, whether the respondent was selecting a product to buy or just looking at it.
- Click the Shelf Metrics dropdown and select a view.
- To download the image, click the hamburger menu and select PNG, PDF, or EPS.
| View | What it shows |
|---|---|
| Shelf Heatmap | Click density overlaid on the shelf as originally programmed. It's drawn with a 2D kernel density estimate in nine layers—the bottom layer covers the area where 99% of clicks fall, and the top layer covers 50%. |
| Product Heatmap | Total clicks per product on a color scale from the minimum (green) to the maximum (red). |
| Clicks | The exact coordinates of each individual click, unaggregated. |
| Shelf view | The shelf as originally programmed, with no respondent data. |
⚠️ Note: The shelf image behind the heatmap is captured when the survey launches and isn't retaken. On older surveys with randomization enabled, respondents may have seen a different layout than the one shown in Results. Randomization is no longer available on shelf-based stores; it's offered only in the Ecommerce store.
Analyzing Shelf Summary Statistics
Summary statistics for the whole shelf appear below the shelf visualization. The Calculation column shows how each one is computed.
Time-based statistics use the geometric mean instead of the arithmetic mean. Shopping behavior varies widely, and some respondents spend far longer at the shelf than others. The geometric mean keeps those respondents in the data as real shopping behavior without letting them skew the result. That's why, for example, Average time at shelf uses the geometric mean.
| Statistic | Calculation | What it measures |
|---|---|---|
| Average cart spend | Arithmetic mean | How much money each respondent spent. |
| Average cart size | Arithmetic mean | How many products each respondent purchased. |
| Average number of unique products in cart | Arithmetic mean | How many different products each respondent purchased. |
| Total products purchased | Count | All products purchased by all respondents. |
| Total products picked up | Count | All products picked up from the shelf by all respondents. |
| Total product spend | Sum | All money spent on products by all respondents. |
| Average time at shelf | Geometric mean | Time from entering the exercise to exiting it. |
| Average time to first click | Geometric mean | Time from entering the exercise to first clicking a product to examine it closer. |
| Average attention span per product | Geometric mean, then arithmetic mean | How long respondents viewed product details. Each respondent's viewing time per product is averaged with the geometric mean, then those averages are combined with the arithmetic mean. |
⚠️ Note: Average cart spend and Purchase amount were previously calculated using the geometric mean. To reproduce the old values, export Response level data and apply the Excel formula
=GEOMEANto the amounts spent.
Analyzing Product Metrics
Product metrics break results down by individual product so you can compare them side by side.
- Click the Product Metrics dropdown below the shelf summary statistics and select a metric.
- Hover over a bar to see a further breakdown for that product. When Purchase selection is selected, hovering also shows live significance testing.
| Metric | Calculation | What it measures |
|---|---|---|
| Purchase selection | Percentage | The share of respondents who bought the product. |
| Purchase quantity | Arithmetic mean | How many units of the product each buyer purchased. |
| Purchase amount | Arithmetic mean | How much each buyer spent on the product. |
| Purchased, picked up only | Percentage | Splits respondents into three groups: those who picked up the product and bought it, those who picked it up but didn't buy it (Picked up only), and those who did neither. |
| Attention span | Geometric mean | Time spent in the product's detail view. |
| Time to cart | Geometric mean | Time from starting the task to adding the product to cart. |
| Time to first click | Geometric mean | Time from entering the exercise to first clicking the product to examine it closer. |
Who Each Metric Is Based On
Most metrics are calculated only among respondents who took a specific action with a product. If no respondent qualifies, the metric reads N/A rather than zero.
| Metric | Based on |
|---|---|
| Purchase selection | All respondents who saw the question |
| Purchase quantity, Purchase amount | Respondents who purchased that product |
| Attention span | Respondents who opened that product's detail view |
| Time to cart | Respondents who added that product to their cart at any point |
| Time to first click | Respondents who clicked that product |
| Average cart spend, Average cart size, Average number of unique products in cart | Respondents who purchased at least one product |
Display-only products
A product set to Display only in its Product Card can't be added to a cart, but it stays in the charts rather than disappearing. Its clicks appear on the heatmap, and its Purchase selection reads 0%. Because no one can buy it, metrics based on buyers—Purchase quantity, Purchase amount, and Time to cart—read N/A.
Exporting the Results
Two exports include Shelf Test results. Use Summary data to get the numbers you see on the Results page. Use Response level data to get respondent-level data along with per-product Shelf Test metrics.
| Export | What you get for a Shelf Test question |
|---|---|
| Summary data (Excel) | A copy of the Results page: the nine shelf summary statistics, plus the per-product table for the Product Metrics view you currently have selected. |
| Response level data — Excel | Respondent-level data, plus a separate Q<n> Shelf-Test question sheet for each Shelf Test question with the per-product metrics. You get this sheet with either data type. |
| Response level data — JMP | Respondent-level data plus the Shelf Test metrics. |
| Response level data — CSV or SPSS | Respondent-level data only. These formats don't include the Shelf Test metrics. |
Exporting response level data
- Click the Export icon in the left-hand navigation.
- In the export-type dropdown, select Response level data.
- Under File types, select Excel, SPSS, CSV, or JMP.
- If you selected Excel, choose a Data type: Numerically coded with labels or Answer text.
- Optionally, select Hide aytm Logic or Exclude OE Code Columns.
- Click Download file.
Exporting summary data
- On the Results page, select the Product Metrics view you want to export.
- Click the Export icon in the left-hand navigation.
- In the export-type dropdown, select Summary data. Excel is the only file type for this export.
- Click Download file