> For the complete documentation index, see [llms.txt](https://docs.shopperos.stackline.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.shopperos.stackline.com/concierge/insights.md).

# Insights

**Concierge → Insights** shows how shoppers use your concierge. You see conversations, the sales they lead to, support results, conversation length, and shopper sentiment. One period and one set of filters apply to every tab.

<figure><img src="https://489913050-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FblcXcPwOUg5osGsERflQ%2Fuploads%2Fgit-blob-ccfed3a152b93b1ab2892e71993e1289703b2087%2Fconcierge-insights.jpg?alt=media" alt="Concierge → Insights → Summary"><figcaption><p>The Summary tab. The period and Filter at the top apply to every tab. Click a metric to open its tab.</p></figcaption></figure>

## Six tabs answer six questions

Each tab answers one question about your concierge.

<table data-view="cards"><thead><tr><th></th><th></th></tr></thead><tbody><tr><td><strong>Summary</strong></td><td>All 20 key metrics in one place, and every conversation.</td></tr><tr><td><strong>Conversations</strong></td><td>How many conversations shoppers have, and how they start them.</td></tr><tr><td><strong>Shopping</strong></td><td>The sales the concierge drives.</td></tr><tr><td><strong>Support</strong></td><td>How well the concierge handles customer support.</td></tr><tr><td><strong>Duration</strong></td><td>How long conversations last.</td></tr><tr><td><strong>Sentiment</strong></td><td>How shoppers feel about their conversations.</td></tr></tbody></table>

## Choose the period and filters

The period and the filters at the top apply to every tab.

**Period**: choose **YTD**, **52W**, **26W**, **13W**, **4W**, or **Custom**. **52W** is the default. Each section subtitle names the period, for example "during the last 52 week period".

To filter the figures:

{% stepper %}
{% step %}

### Open the filters

At the top, click **Filter**.
{% endstep %}

{% step %}

### Choose values

Choose the values to include.
{% endstep %}

{% step %}

### Apply

Click **Apply**. A dot on **Filter** shows that filters are applied.
{% endstep %}
{% endstepper %}

Done. Every tab shows only the filtered conversations. To remove all filters, click **Clear**.

| Filter                               | Values                                                                                                                                                                                  |
| ------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Channel**                          | **Chat**, **Voice**, **Email**, **SMS**                                                                                                                                                 |
| **Conversation Type**                | **Shopping**, **Product Support**, **Returns Support**, **Order Support**, **Payments Support**, **Rewards Support**, **Shipping Support**, **Account Support**, **Reviews**, **Other** |
| **Sentiment**                        | **Positive**, **Negative**, **Neutral**                                                                                                                                                 |
| **App Format**                       | **Web App**, **Mobile App**                                                                                                                                                             |
| **App Owner**                        | **Brandclub**, **Client**, **ChatGPT**                                                                                                                                                  |
| **Classification**                   | **Brand**, **Category**, **Subcategory**, **Product**                                                                                                                                   |
| **Order Format** (Shopping tab only) | **Checkout**, **External**                                                                                                                                                              |

On the **Support** tab, **Conversation Type** lists only the support types.

## Summary shows every key metric

**Key Metrics** shows 20 metrics in four columns: **Conversations**, **Shopping**, **Value**, and **Experience**. Click a metric to open the tab with its details.

| Metric                    | What it measures                                                                   |
| ------------------------- | ---------------------------------------------------------------------------------- |
| **CONVERSATIONS**         | The number of conversations.                                                       |
| **CONVERSATION RATE**     | The share of shoppers who saw a concierge element and started a conversation.      |
| **CONVERSATION DURATION** | The average length of a conversation.                                              |
| **SHOPPING RATE**         | The share of conversations about shopping.                                         |
| **SUPPORT RATE**          | The share of conversations about customer support.                                 |
| **PRODUCT IMPRESSIONS**   | How many times the concierge showed products.                                      |
| **PRODUCT CLICKS**        | How many times shoppers clicked those products.                                    |
| **PRODUCT CTR**           | Product clicks as a share of product impressions.                                  |
| **CONVERSION RATE**       | Orders divided by shopping conversations.                                          |
| **ORDERS**                | The orders that came from concierge conversations.                                 |
| **UNITS SOLD**            | The units in those orders.                                                         |
| **RETAIL SALES**          | The sales value of those orders.                                                   |
| **AVG ORDER VALUE**       | Retail sales divided by orders.                                                    |
| **COST SAVINGS**          | An estimate of the support labor cost saved, based on the number of conversations. |
| **TOTAL ECONOMIC VALUE**  | Retail sales plus cost savings.                                                    |
| **POSITIVE SENTIMENT**    | The share of conversations with positive shopper feedback.                         |
| **NEUTRAL SENTIMENT**     | The share of conversations with no shopper feedback.                               |
| **NEGATIVE SENTIMENT**    | The share of conversations with negative shopper feedback.                         |
| **DEFLECTION RATE**       | The share of support conversations that did not need a human agent.                |
| **RESOLUTION RATE**       | The share of support conversations with tickets that the concierge resolved.       |

Below the metrics, **Data Table by Conversation** lists each conversation. See [Read a conversation](#read-a-conversation).

### Learn more opens a metric's definition

Some chart subtitles end with **Learn more**. Click it to open a dialog with the metric's definition. Close it with **×**.

For example, the **Conversation Rate** chart on the **Conversations** tab has this definition: the percentage of consumers who saw a concierge element and chose to engage in a conversation. It equals unique conversations divided by unique viewers.

## Conversations shows how shoppers start chats

This tab shows how many conversations shoppers have, and how they start them.

| Chart                             | What it shows                                                      |
| --------------------------------- | ------------------------------------------------------------------ |
| **Conversations**                 | Conversations over time. Choose the time step in the subtitle.     |
| **Conversations by Type**         | Conversations by conversation type.                                |
| **Conversations by Category**     | Conversations by category. You can switch to subcategory or brand. |
| **Conversations by Access Point** | Where shoppers opened the concierge.                               |
| **Conversations by App Format**   | Web or mobile app. You can switch to **App Owner**.                |
| **Conversations by Channel**      | Chat, voice, email, or SMS.                                        |
| **Conversation Rate**             | The conversation rate by day.                                      |
| **Shopper Inputs by Format**      | Shopper messages by format. You can switch to **Language**.        |
| **Concierge Outputs by Format**   | Concierge replies by format. You can switch to **Language**.       |

On breakdown charts, switch between the count and **% of** the total.

## Shopping shows the sales the concierge drives

**Key Metrics** has three columns:

| Column         | Metrics                                                      |
| -------------- | ------------------------------------------------------------ |
| **Engagement** | **Product Impressions**, **Product Clicks**, **Product CTR** |
| **Conversion** | **Conversion Rate**, **Orders**, **Units Sold**              |
| **Sales**      | **Retail Sales**, **Avg Order Value**, **Avg Sell Price**    |

**Avg Sell Price** is retail sales divided by units sold.

The tab then shows **Orders** over time, and orders by category, access point, app format, and channel.

**Data Table by Product** lists **Impressions**, **Clicks**, **Conversion**, and **Retail Sales** for each product. Change the view to **Brand**, **Category**, or **Subcategory**.

## Support shows deflections and resolutions

This tab shows how well the concierge handles customer support. **Key Metrics** shows three rates:

| Metric                    | What it measures                                                        |
| ------------------------- | ----------------------------------------------------------------------- |
| **Customer Support Rate** | How many of all conversations were about support.                       |
| **Deflection Rate**       | Support conversations that the concierge handled without a human agent. |
| **Resolution Rate**       | Support conversations with tickets that the concierge resolved.         |

The charts show **Deflections** and **Resolutions** over time, and by type, category, access point, app format, and channel.

**Data Table by** lists **Support Conversation**, **Deflections**, and **Deflection Rate**. Change the view to **Product**, **Brand**, **Category**, **Subcategory**, or **Conversation**. The **Conversation** view also shows **Resolution Type** and **Resolutions**.

## Duration shows how long conversations last

| Metric                        | What it measures                                           |
| ----------------------------- | ---------------------------------------------------------- |
| **AVG Conversation Duration** | The average length of a conversation.                      |
| **AVG Shopper Inputs**        | The average number of shopper messages in a conversation.  |
| **AVG Concierge Outputs**     | The average number of concierge replies in a conversation. |

The charts show these metrics over time, and by type, category, access point, app format, and channel.

## Sentiment shows how shoppers feel

**Key Metrics** shows the share of **Positive**, **Neutral**, and **Negative** conversations.

Positive and negative come from shopper feedback on a conversation. A conversation with no feedback counts as neutral.

The charts show sentiment by type, category, access point, app format, and channel. Switch between the count and **Sentiment %**.

## Read a conversation

In **Data Table by Conversation**, click a **Conversation Name**. The full conversation opens in a new tab.

The table shows each conversation's **Date**, **Conversation Type**, **Duration**, **Sentiment**, **Orders**, and **Deflections**.

## Export the conversation table

You can download the conversation table as a file.

{% stepper %}
{% step %}

### Start the export

In **Data Table by Conversation**, click the export icon.
{% endstep %}

{% step %}

### Choose the fields

In the **Export** dialog, tick any additional fields you want in the file.
{% endstep %}

{% step %}

### Export

Click **Export**. You see **Your download is processing**, and then **Your download is ready**.
{% endstep %}

{% step %}

### Get the file

Click **View downloads**.
{% endstep %}
{% endstepper %}

Done. The file is in your downloads. Next, use what you learn to [improve the concierge's knowledge](/concierge/knowledge.md).

On the **Shopping** data table, and on the **Support** data table in every view except **Conversation**, the export icon starts the export at once.

## Troubleshooting

| Message                                                                                    | What to do                                          |
| ------------------------------------------------------------------------------------------ | --------------------------------------------------- |
| **Something went wrong with your export. Please try again.**                               | Try the export again.                               |
| **Your export is taking longer than expected. It will keep processing in the background.** | Wait. The file is ready later under your downloads. |
| **Failed to open conversation preview. Please try again.**                                 | Click the conversation name again.                  |

## Next step

Improve the answers behind these figures.

{% content-ref url="/pages/4J0oYqLxOSsYs0aGRCpF" %}
[Knowledge and guardrails](/concierge/knowledge.md)
{% endcontent-ref %}