> For the complete documentation index, see [llms.txt](https://www.analytics-docs.laplacetec.com/guide/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://www.analytics-docs.laplacetec.com/guide/en/customdashboard/custom/datasettings/cross_selling.md).

# Cross-purchase analysis

{% hint style="info" %}
This analysis finds **which products sell together with which other products**. It counts the combinations placed in the same cart, which you can use as the basis for bundles, recommendations, and joint promotions.
{% endhint %}

## Choose the analytics mode <a href="#analysistype" id="analysistype"></a>

In the chart editor, open the **Data** tab in the settings panel on the right, then click the **Cross-purchase analysis** icon under **Analytics mode**. Hover over an icon to see its name.

| Item         | Value                                       |
| ------------ | ------------------------------------------- |
| Data source  | Order data · Base product order data        |
| Period unit  | Total · Day · Week · Month · Quarter · Year |
| Default unit | Day                                         |

{% hint style="warning" %}
**Minute and hour units are not available.** Cross-purchase data is aggregated by store and date, so you can't view it at a finer grain than a day. If the unit is set to minute or hour when you switch to this analysis, it is automatically changed to **Day**.
{% endhint %}

## Breakdown <a href="#dimensions" id="dimensions"></a>

This step decides which combinations you look at. Click **Breakdown** on the **Data** tab and choose a basis such as product, option, base product, or category; A ↔ B combinations are then built on that basis. You can add more breakdowns to narrow the combinations further.

## Which view to use <a href="#charttype" id="charttype"></a>

Choose one of four options under **Chart type** on the **Chart** tab of the settings panel.

| Chart type                 | When to use                                                   |
| -------------------------- | ------------------------------------------------------------- |
| Affinity heatmap (default) | Scan the A × B grid by color to find strong combinations      |
| Cross-purchase network     | See as a relationship map how products cluster when they sell |
| Cross-purchase trend       | See whether a specific combination is growing over time       |
| Product pair table         | Sort combinations by value and pull a top list                |

Use **Products shown** on the **Chart** tab to control how many products are displayed. The number of combinations grows with the square of the product count, so a large value makes the view crowded.

## Reading the measures <a href="#measures" id="measures"></a>

Cross-purchase analysis has measures used only in this analysis. Here are the most common ones.

| Measure                           | Meaning                                                                                                                                                                                                                                 |
| --------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Lift (cross-purchase association) | How much more often items actually sell together compared with the probability of selling together by chance. **1.0 means chance level; above 1 means the items attract each other**; below 1 means they actually push each other away. |
| Cross-sell payments               | The number of payments where A and B were in the same cart. This is **the absolute volume sold together**.                                                                                                                              |
| A CSR (payments)                  | The share of orders containing A that also contained B. Read it as **the probability that a customer who bought A also buys B**.                                                                                                        |
| A payments · B payments           | The number of times each product sold on its own                                                                                                                                                                                        |
| A OR B payments                   | The number of times A or B sold                                                                                                                                                                                                         |
| Cross-sell paying customers       | Counted as **number of people** instead of number of payments. There is a separate set of customer-count measures.                                                                                                                      |

{% hint style="info" %}
**Look at absolute volume and association together.**\
However high the Lift, if the pair sold together only a few times it may be chance. Conversely, even with many payments, if both products are bestsellers anyway, Lift will be near 1 and there's no reason to bundle them. It's safest to look at **Cross-sell payments (absolute volume) + Lift (association)** together.
{% endhint %}

### Cross-selling opportunity threshold (Lift) <a href="#opportunity" id="opportunity"></a>

**Cross-selling opportunity threshold (Lift)** sets the Lift multiple at or above which a combination is highlighted as an "opportunity". The default is **1.5x**. Loosen it if you have little data and few combinations are caught; tighten it if too many combinations are caught.

* 1.2x (loose) · 1.5x (default) · 2x · 2.5x · 3x (strict)

## Ask your dedicated manager <a href="#contact" id="contact"></a>

Having trouble using cross-purchase analysis?

Contact us through the inquiry page or by email below. Your manager will be happy to help.

> 💬 [Contact us on our website](https://www.analytics.laplacetec.com/contact)

**Email:** [**product@laplacetec.com**](mailto:product@laplacetec.com)