---
title: "Data source best practices - Chat Thing Documentation"
canonical_url: "https://chatthing.ai/docs/data-source-best-practices"
last_updated: "2026-07-20T15:03:03.239Z"
meta:
  description: "Get better results from your bot with these tips"
  "og:description": "Get better results from your bot with these tips"
  "og:title": "Data source best practices"
  "twitter:description": "Get better results from your bot with these tips"
  "twitter:title": "Data source best practices"
---

**Data sources**

# Data source best practices

Get better results from your bot with these tips.

## Structuring large websites

For sites with 100+ pages, consider splitting into multiple data sources:

- **By section:** Docs, Blog, Product pages as separate data sources
- **By update frequency:** Static pages vs frequently updated content
- **By importance:** Core pages synced daily, archives synced weekly

This lets you:

- Sync only what changed (saves storage tokens)
- Use different CSS selectors per section
- Troubleshoot issues more easily

## CSS selectors for common platforms

The right CSS selector captures your main content and excludes headers, footers, and navigation.

| Platform | Recommended selector |
| --- | --- |
| **WordPress** | `article` or `.entry-content` |
| **Webflow** | `.main-content` or `main` |
| **Squarespace** | `.main-content` |
| **Shopify** | `main` or `.product-description` |
| **GitBook** | `.markdown-body` |
| **Notion (public)** | `.notion-page-content` |
| **Generic** | `main` or `article` |

💡 **Tip:** Inspect your page in browser DevTools to find the best selector. Test on one page before syncing your whole site.

## Crawl vs Sitemap vs Manual URLs

| Method | Best for |
| --- | --- |
| **Crawl** | Discovering all pages on a site (up to 600) |
| **Sitemap** | Large sites with existing sitemaps |
| **Manual URLs** | Specific pages only, or when crawl misses pages |
| **Bulk URLs** | Adding a known list of pages quickly |

## Sync frequency recommendations

| Content type | Recommended schedule |
| --- | --- |
| News/blog (daily posts) | Daily |
| Documentation | Daily or weekly |
| Marketing pages | Weekly |
| Product pages | Daily (if prices/stock change) |
| Static content | Weekly or manual |

## Reducing storage token usage

- Use CSS selectors to exclude unnecessary content
- Split large data sources so you only re-sync what changed
- Remove pages you don't need from the data source
- Use auto-sync with diffs (only syncs changed content)

## Notion tips

- Add top-level pages only - child pages are included automatically
- Don't add the same page multiple times
- If permissions change, disconnect and reconnect Notion in account settings

## YouTube tips

- Transcripts must be enabled on videos
- Add playlists to bulk-add related videos
- Auto-generated captions work, but manual captions are more accurate