Linux 软件免费装
Banner图

MarketBasketAnalysis for WooCommerce

开发者 brian48x
更新时间 2026年10月2日 21:15
PHP版本: 8.1 及以上
WordPress版本: 7.1
版权: GPLv2 or later
版权网址: 版权信息

标签

woocommerce recommendations cross-sell bundles frequently bought together

下载

0.17.0

详情介绍:

Market Basket Analysis turns your WooCommerce order history into ranked bundle and cross-sell opportunities, backed by real co-purchase patterns. Not "people also viewed" guesses. Not random product carousels. Proper association-rule mining (FP-Growth + SQL pairs), the same algorithms big retailers use, shipped as a WordPress plugin that runs entirely on your server. Every rule is auditable: support, confidence, and lift are visible on the Opportunities grid, so you can see why the engine picked what it picked. What's included MBA Plus (separate plugin) Everything in this plugin works with no license and no account. A separate build, MBA Plus, is distributed from marketbasketanalysis.com and adds the revenue-weighted HUI mining engine and hosted-compute features (hosted checkout assembly, the hosted MCP surface for AI agents, and high-quota API access). It is a different plugin, not an unlock: nothing in this one is withheld or limited while it is not installed. AI features here (the ai_catalog cold-start engine, AI bundle copy, and catalog enrichment) use your own Anthropic or OpenAI API key and the provider bills you directly. We never sit between your store and the LLM. How it's different Most WooCommerce recommendation plugins fall into two buckets: rule-based ("show 4 random products from the same category") or "people also bought" lookups that ignore co-occurrence patterns. MBA implements proper association-rule mining, the same family of algorithms used by Amazon's "Frequently Bought Together" feature and Adobe Sensei recommendations, just as a WordPress plugin that runs on your server with no SaaS sign-up. Compatibility

安装:

  1. Upload the plugin folder to /wp-content/plugins/. The in-admin install via Plugins, Add New, Search "MarketBasketAnalysis" applies once the plugin is live in the WordPress.org directory; until the listing is approved, the upload path is the way in.
  2. Activate the plugin through the Plugins menu in WordPress.
  3. Make sure WooCommerce is installed and active.
  4. Navigate to WooCommerce, MBA: Dashboard to run your first mining job.
  5. (Optional) Add the FBT widget to a product page via the Gutenberg block or the [mbawoo_fbt] shortcode.
  6. (Optional) The separate MBA Plus plugin, with the revenue-weighted HUI engine and hosted compute, is available at marketbasketanalysis.com.
For a detailed quick-start, see the installation guide.

屏幕截图:

  • Jobs page: kick off a mining job, see status and rule counts for past runs.
  • Tabbed Settings page: General, Mining, Bundles, Notifications, AI Catalog Pairing, Cost & Scoring.
  • Storefront Frequently Bought Together widget on a single-product page.

常见问题:

Does this work with a fresh store with no orders yet?

The data-driven engines (sql_pairs, fp_growth) require order history. Around 200 completed orders in the analysis window is the sensible floor below which patterns are too sparse to mine. For brand-new stores with no order history, the free ai_catalog engine produces recommendations from catalog metadata alone, using your own Anthropic or OpenAI API key (the provider bills you directly).

Will mining slow down my site?

No. Jobs run in the background via WP-Cron or are kicked off manually from the admin. The storefront widget hits precomputed rules from a fast database table; lookups are sub-millisecond and never a real-time computation at page load.

Does this support High-Performance Order Storage (HPOS)?

Yes. The miner checks the site's HPOS setting and queries the right tables automatically. No configuration needed.

Where does my data go?

Mining runs locally. The data-driven mining engines make no outbound API calls, and individual orders and customer records never leave your server. AI features call your own Anthropic or OpenAI account directly using a key you paste into Settings. They send product catalog data and aggregate statistics (for example, how many baskets contain a product pair), never customer names, email addresses, or individual orders. The External services section below lists exactly what each optional service receives and when. We never see the catalog payload or the LLM response. Every AI feature (the ai_catalog engine, AI bundle copy, and catalog enrichment) needs only your provider key, and the provider bills you directly. You set a monthly token budget for the AI assistant and the storefront rationales under Settings, AI Catalog Pairing, so a busy month cannot surprise you; 0 means no limit.

How does the FBT widget choose recommendations?

For a given product, the widget looks up the top-scored opportunities where that product appears in the antecedent (the "if" side of an association rule). Ties are broken by lift, then confidence.

Can I customize the widget styling?

Yes. Every block exposes its CSS classes (.mbawoo-fbt, .mbawoo-fbt__item, etc.). Override in your theme's stylesheet or use the Gutenberg block's color/layout controls.

How is AI billed?

AI features that call an LLM use YOUR Anthropic or OpenAI API key, and the provider bills you directly. Nothing about them is metered by us. You can cap the monthly token spend for the AI assistant and the storefront rationales under Settings, AI Catalog Pairing.

Is the source code available?

Yes, this plugin is GPLv2+ licensed and you can read and modify any of its code.

Where do I get support?

Use the WordPress.org plugin support forum once the directory listing is approved; until then, marketbasketanalysis.com/contact.

更新日志:

0.17.0 0.16.1 0.16.0 0.15.1 0.15.0 0.14.0 0.13.0 0.12.0 0.11.0