> 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/cohort_analysis.md).

# Cohort analysis

{% hint style="info" %}
A cohort is a group of customers who took a specific action at the same reference point. Cohort analysis in Laplace custom dashboards groups customers by their first purchase date and tracks their repurchase behavior over time.
{% endhint %}

Cohort analysis helps you understand:

* **How often customers acquired in a given period come back to buy again**
* **When the actual customer value (LTV/LTR) pays back your ad spend**
* **How customer quality compares across marketing channels and campaigns**
* **Repurchase cycles and when customers churn**

{% hint style="info" %}
💡 Example uses of cohort analysis

• "Of the customers who first purchased in February, what % repurchased within D+7?"

• "When will this month's new-customer CAC (customer acquisition cost) be paid back by contribution margin?"

• "How does the repurchase rate differ between the first-purchase cohort and the 2nd-purchase cohort?"
{% endhint %}

## Data settings <a href="#datasettings" id="datasettings"></a>

In the chart editor, open the **Data** tab in the settings panel on the right, then click the **Cohort Analysis** icon under **Analytics mode**. All of the settings below are on the **Data** tab.

### Data source <a href="#datasource" id="datasource"></a>

The **data source** step is where you choose the source data for the analysis. Depending on your goal, you can choose from the data you have connected, such as orders, ads, base product data, and Cafe24 Analytics.

<figure><img src="https://616205588-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fqr1veGyUqCNwQBBmk3yt%2Fuploads%2Fgit-blob-1a555b9988234c7eb08ef8388abcc72569f7ca48%2Fcustomdashboard__custom__datasettings__cohort_analysis__01.png?alt=media" alt="Cohort analysis data source and measures" width="436"><figcaption></figcaption></figure>

1. Choose order data or base product order data.
2. For measures, pick one value from data source A to serve as the baseline, then pick the measures to visualize from data source B.

### Cohort settings <a href="#conversionsettings" id="conversionsettings"></a>

This step sets what defines a cohort. Under **Time-based**, choose the purchase count that defines the cohort: first-purchase cohort, 2nd, 5th purchase cohort, or purchase cohort. The default is **First-purchase cohort**. Under **Behavior-based**, you can split cohorts by customer purchase behavior, such as **First purchase amount range** · **Average order value range (AOV)** · **Purchase frequency** · **Average purchase interval** · **Time since last purchase**.

<figure><img src="https://616205588-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fqr1veGyUqCNwQBBmk3yt%2Fuploads%2Fgit-blob-9c175263e4f965d279d99aaf55ec572afcc95a66%2Fcustomdashboard__custom__datasettings__cohort_analysis__02.png?alt=media" alt="Cohort settings" width="432"><figcaption></figcaption></figure>

<table><thead><tr><th>Setting</th><th width="416">Meaning</th><th></th></tr></thead><tbody><tr><td>First-purchase cohort</td><td>Sets the first purchase as D+0. Use it to analyze repurchase patterns after acquiring new customers.</td><td>Most common</td></tr><tr><td>2nd, 5th purchase cohort</td><td>Uses the 2nd, 3rd, etc. purchase as the reference point. Use it to track behavior after a specific purchase stage.</td><td>Advanced</td></tr><tr><td>Purchase cohort</td><td>Uses every purchase as a reference point. Use it to track post-purchase behavior regardless of purchase count.</td><td></td></tr></tbody></table>

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

<figure><img src="https://616205588-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fqr1veGyUqCNwQBBmk3yt%2Fuploads%2Fgit-blob-e8bda874c57ed3eda0c02c51c2a9ccfef0c1e442%2Fcustomdashboard__custom__datasettings__cohort_analysis__03.png?alt=media" alt="Cohort breakdown" width="313"><figcaption></figcaption></figure>

Set the breakdown used to visualize the chart; the chart is drawn according to the breakdown you choose.

* As with basic chart analysis, you can add multiple breakdowns and view them together in one chart.
* Click the default **'Store'** breakdown to change it to the one you want.

### Refunded / canceled orders <a href="#refund" id="refund"></a>

<figure><img src="https://616205588-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fqr1veGyUqCNwQBBmk3yt%2Fuploads%2Fgit-blob-aa766719c35998acebc49c48aa22245440260da7%2Fcustomdashboard__custom__datasettings__cohort_analysis__04.png?alt=media" alt="Include refunded / canceled orders" width="430"><figcaption></figcaption></figure>

The **Include refunded / canceled orders** switch sets whether refunded or canceled orders are included in the results.\
Turn it **on** to include refunded and canceled orders in the analysis; turn it **off** to get results based on data that excludes them.

## Chart settings <a href="#chartsettings" id="chartsettings"></a>

On the **Chart** tab of the settings panel, you can change how the chart is visualized. Understanding how each option changes the actual numbers helps you tailor the analysis to your goal.

### Chart type <a href="#charttype" id="charttype"></a>

<table><thead><tr><th width="156">Type</th><th width="424">Description</th><th>Best for</th></tr></thead><tbody><tr><td><strong>Line</strong></td><td>Shows how each cohort's repurchase rate changes as a line. Lets you compare retention curves over time at a glance. Useful for seeing differences between cohorts visually.</td><td>Recommended: comparing trends</td></tr><tr><td><strong>Line + table</strong></td><td>Shows the line on top and the cohort table below. Both also appear on the dashboard card.</td><td>Recommended: trend and numbers together</td></tr><tr><td><strong>Cohort table</strong></td><td>A heatmap table with cohort dates as rows and D+N as columns. Read repurchase rates from each cell's value and color intensity. Useful for checking exact numbers and sharing.</td><td>Recommended: checking numbers</td></tr></tbody></table>

### Retention method

On the **Chart** tab › **Retention method**, choose **Calendar** · **Rolling** · **Frequency**.

![Retention method on the Chart tab](https://616205588-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fqr1veGyUqCNwQBBmk3yt%2Fuploads%2Fgit-blob-f111c500f23580eb0c1e644eed1a8d2a06034fa5%2Fcustomdashboard__custom__datasettings__cohort_analysis__05.png?alt=media)

This is the most important setting and the easiest to confuse. The same data can produce completely different numbers depending on the method.

{% hint style="info" %}
Key concept: what is D+N?

D+0 = cohort reference date (first purchase date)

D+1 = 1 day after the reference date

D+7 = 7 days after the reference date

→ If the period unit is "day", D+1 = the next day; if it is "month", D+1 = the next month
{% endhint %}

#### Calendar (fixed-period basis)

Measures the share of customers who purchased in the Nth period (calendar-based) from the cohort reference date.

| ✅ When to use                                                                                                                                                             | ⚠️ Caveats                                                                                                        |
| ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------- |
| When measuring behavior at a specific point, such as "purchased exactly in month 1 (D+1) after the first purchase". Suited to subscription/recurring-delivery businesses. | Customers who purchased only between D+0 and D+1 may not be counted in D+1. Each period is counted independently. |

Example: period unit "month" + Calendar → D+1 counts only customers who purchased in "the month after their first purchase"

#### Rolling (cumulative basis)

Measures the share of customers who purchased at any time from the cohort reference date through D+N. (Cumulative)

| ✅ When to use                                                                                                                                                          | 👉 Characteristics                                                                                                        |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------- |
| When tracking behavior within a time window, such as the share of "customers who repurchased at least once within D+7". Suited to e-commerce with irregular purchases. | Values increase (or stay flat) monotonically from D+0 → D+N. Generally higher than Calendar. This is the Laplace default. |

Example: period unit "day" + Rolling → D+7 = share of "customers who repurchased within 7 days of their first purchase"

#### Frequency (frequency basis)

Analyzes based on the number of purchases (frequency) customers made in the D+N period.

| ✅ When to use                                                                                                                                                          | ⚠️ Caveats                                                                                                                                 |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------ |
| Analyses where purchase frequency matters, such as "How many times on average did customers buy during D+30?". Suited to frequently purchased consumer goods and F\&B. | Counts the number of purchases, not the repurchase rate (share). Recommended to use with the "Cohort payments (cohort\_pay\_cnt)" measure. |

{% hint style="info" %}
🔑 Choosing a method

• General e-commerce (irregular purchases): Rolling recommended → easy to read as "repurchase rate within D+30"

• Subscription/recurring delivery: Calendar recommended → checks whether customers purchased in exactly the Nth cycle

• Purchase count analysis: Frequency → see "how many times customers bought per month on average"
{% endhint %}

### Accumulation <a href="#tablesettings" id="tablesettings"></a>

On the **Chart** tab › **Accumulation**, decide how D+N values are accumulated. This setting is the key to LTV/LTR analysis.

<table><thead><tr><th width="166">Setting</th><th>Meaning</th><th>Main use</th></tr></thead><tbody><tr><td><strong>General</strong> (no accumulation)</td><td>Values counted independently for each D+N period. A high D+1 can still be followed by a low D+2.</td><td>See the repurchase-rate trend per period. "When do most customers churn?"</td></tr><tr><td><strong>Cumulative (on or before)</strong></td><td>Each D+N value includes the sum of D+0 through D+(N-1). D+3 = D+0 + D+1 + D+2 + D+3.</td><td>Track cumulative LTV/LTR. Check "How much has one customer spent by D+30?"</td></tr><tr><td><strong>Cumulative (on or after)</strong></td><td>Each D+N value includes the sum from D+(N+1) onward. Sums toward the future.</td><td>Estimate remaining value after a given point. Analyze additional profit after CAC payback.</td></tr></tbody></table>

{% hint style="info" %}
💡 Example uses of accumulation

• Goal: "Does new customers' cumulative LTV at D+90 exceed CAC?"

→ Settings: B measure = LTV + Cumulative (on or before)

→ Result: The cumulative LTV from D+0 to D+90 rises as a line

<br>

• Goal: "On which day is the repurchase rate highest?"

→ Settings: Repurchase rate + General (no accumulation)

→ Result: See at which D+N churn spikes
{% endhint %}

## Settings by goal <a href="#usecases" id="usecases"></a>

The measures and breakdowns for cohort analysis vary widely depending on what you want to analyze. Below are examples of settings commonly used by commerce businesses.

### Scenario 1: Track new-customer repurchase rate <a href="#node" id="node"></a>

See how quickly and how often new customers buy again after their first purchase.

<table><thead><tr><th width="217">Setting</th><th>Recommended value</th></tr></thead><tbody><tr><td>A (cohort baseline)</td><td>Base product order data cohort → Cohort ARPPU</td></tr><tr><td>B (tracked measure)</td><td>Base product order data → Repurchase rate</td></tr><tr><td>Cohort basis</td><td>First-purchase cohort</td></tr><tr><td>Period unit</td><td>Day or Week</td></tr><tr><td>Retention method</td><td>Rolling</td></tr><tr><td>Accumulation</td><td>General</td></tr></tbody></table>

![Scenario 1 result](https://616205588-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fqr1veGyUqCNwQBBmk3yt%2Fuploads%2Fgit-blob-35ca30624f1e3ff1ecb6dab281539d14da42fa02%2Fcustomdashboard__custom__datasettings__cohort_analysis__06.png?alt=media)

→ How to read it: As in the image (D+0 = 1.91%, D+1 = 1.09%, D+7 = 0.13%), the repurchase rate declines over time, forming a retention curve. This shows how effective it is to drive quick repurchases within the first 7 days.

### Scenario 2: CAC payback timing (LTV analysis)

<table><thead><tr><th width="217">Setting</th><th>Recommended value</th></tr></thead><tbody><tr><td>A (cohort baseline)</td><td>Base product order data cohort → Cohort ARPPU</td></tr><tr><td>B (tracked measure)</td><td>Base product order data → Cumulative contribution margin or LTV</td></tr><tr><td>Retention method</td><td>Rolling</td></tr><tr><td>Accumulation</td><td>Cumulative (on or before)</td></tr><tr><td>Cohort basis</td><td>First-purchase cohort</td></tr></tbody></table>

![Scenario 2 result](https://616205588-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fqr1veGyUqCNwQBBmk3yt%2Fuploads%2Fgit-blob-cef316e9b5cae3c82cc84c5d7851091c651f257b%2Fcustomdashboard__custom__datasettings__cohort_analysis__07.png?alt=media)

→ How to read it: The D+N point where cumulative contribution margin exceeds ad spend (CAC) is the "CAC payback point". It should match the point where the LTV/CAC ratio exceeds 1.0.

### Scenario 3: Compare quality across cohorts

<table><thead><tr><th width="217">Setting</th><th>Recommended value</th></tr></thead><tbody><tr><td>Chart type</td><td>Line</td></tr><tr><td>B (tracked measure)</td><td>Repurchase rate or LTV</td></tr><tr><td>Period unit</td><td>Month</td></tr><tr><td>Retention method</td><td>Rolling</td></tr><tr><td>Data filter</td><td>Add a filter by channel</td></tr></tbody></table>

![Scenario 3 result](https://616205588-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fqr1veGyUqCNwQBBmk3yt%2Fuploads%2Fgit-blob-aadb2e04a61f8df954e0bab92acf923d2851a672%2Fcustomdashboard__custom__datasettings__cohort_analysis__08.png?alt=media)

→ How to read it: Overlaying monthly cohort lines shows whether customers acquired in a particular month are higher quality than in other months. If customers acquired during a promotion have a low long-term repurchase rate, you may want to revisit your promotion strategy.

## FAQ

<details>

<summary>Q. Why is D+0 higher than D+1?</summary>

A. D+0 is the value on the cohort reference date (first purchase date) itself. Because additional purchases on the day of the first purchase, or the total ARPPU for that date, are counted, it is normal for D+0 to be higher than later D+N values. In the image, D+0 = 1.91% is the highest, followed by a declining pattern.

</details>

<details>

<summary>Q. Rolling and Calendar give very different numbers. Which one is right?</summary>

A. Both are accurate; they simply measure different things. Rolling counts "purchased at least once by D+N", so values are higher; Calendar counts "purchased only in exactly the D+N period", so values are lower. Rolling is more intuitive for general e-commerce, Calendar for subscriptions.

</details>

<details>

<summary>Q. How much does including refunds/cancellations affect LTV?</summary>

A. The impact is large in categories with high return rates (fashion, electronics, etc.). When using "Actual Sales (actual\_sales)" or "Actual contribution margin (actual\_contribution\_margin)", we recommend turning **Include refunded / canceled orders** off. In categories with low return rates, the difference is small.

</details>

<details>

<summary>Q. What is the difference between "Cohort users" and "Repurchase user count"?</summary>

A. Cohort users (cohort\_user\_cnt) is the total number of new users in the cohort as of D+0. Repurchase user count (repurchase\_user\_cnt) is the number of those users who actually repurchased. Repurchase rate = Repurchase user count ÷ Cohort users × 100 (%).

</details>

<details>

<summary>Q. Do values keep rising when I use "Cumulative (on or before)"?</summary>

A. If the measure is one where cumulative sums are meaningful, such as revenue or LTV, it increases monotonically. For ratio measures such as repurchase rate, however, applying Cumulative (on or before) can make the results harder to interpret. We recommend applying Cumulative (on or before) to amount-based measures such as LTV or contribution margin.

</details>

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

Having trouble using cohort 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)