正值第一季度最后一个星期，一家 GMV 为 4000 万美元的 WooCommerce 家居用品店的留存经理打开了仪表板。首次购买量同比增长 12%。获客团队已经完成了他们的工作。复购率，这个决定今年剩余时间是否能实现复利增长的唯一数字，在过去十八个月里一直停滞在 22%。她向下滚动到自动化堆栈的 WooCommerce 购后通知部分。只有一项。订单确认电子邮件。第零天。每笔订单后的前两周内的其他十七天都是沉默的。
这种沉默就是差距。大多数 WooCommerce 购后通知堆栈在结账时发送一封交易邮件，三周后发送一封评论请求，然后就认为工作完成了。工作并未完成。在这两次触达之间，是首次购买者决定是否再次购买、重复购买者决定是否升级、高客单价购买者决定是否推荐的整个窗口期。购后工作流是唯一直接作用于该窗口期的自动化。本文介绍了 14 天、6 次触达的工作流，该工作流在两个季度内将同一家商店的复购率从 22% 提升到 28%。
购后是提升复购率的工作，不是挽回损失的工作
大多数留存文章将购后触达归类于“放弃和挽回”。购物车放弃、浏览放弃、召回、购后。四项内容，一个部分。这种分类是错误的。这四项中有三项是挽回工作流。它们在当前旅程停滞时触发，工作流的职责是使其恢复。购后工作流不挽回任何东西。订阅者已经付款。旅程并未停滞。
购后工作流构建下一次购买。这是不同类别的留存工作。它在不同的层面实现复利增长。购物车放弃可挽回 8% 至 14% 的收入，但仅限于那些首先放弃购物车的订阅者。浏览放弃可提升浏览到购物车的转化率，但仅限于访问过产品页面的订阅者。购后工作流提升复购率，这适用于商店拥有的每一位付费客户。对整个客户群的 5% 的改进产生的增量收入，比对一小部分客户的 20% 的改进产生的增量收入更多。
这就是定义购后工作流在堆栈中位置的留存数学。它是唯一一个复购自动化。其他一切都是挽回工作。
一家中型 WooCommerce 商店，70% 的收入来自重复客户，如果将购后工作流视为一次性的事后想法，那么它最大的收入杠杆就被接错了。结账和下一次购买信号之间的 14 天窗口期是客户生命周期中最具操作性的时期。拥有这个窗口期的复购自动化是留存团队本季度可以实现的成本最低的收入增长，因为客户已经付费。
购后工作流的 6 次触达
Six touchpoints, fourteen days, one workflow. The vocabulary the rest of this article uses (START, WAIT, DECISION, SPLIT_PATH, ACTION, END) is the standard PushEngage Workflows builder grammar, covered in depth on the PushEngage Workflows builder announcement post. The single trigger here is the order-placed event, mapped from the WooCommerce woocommerce_payment_complete hook into a PushEngage custom event with the order ID, the order value, and the product list in the payload.
The workflow runs on Multiple Sequential mode. One active instance per customer at a time, but the next order fires a new instance after the current one completes. The workflow never overlaps on itself.
Touchpoint 1 — Order confirmation push, immediate
The first touch fires within seconds of woocommerce_payment_complete. This is transactional, but the tone matters because it is the first message a new buyer receives from the brand after handing over money. Include the order summary and the shipping ETA. Push notifications outperform the order-confirmation email here because the email opens hours later, and the push lands while the buyer is still on the confirmation page.
The order confirmation push is the lightest touchpoint to ship but the easiest one to get wrong. A flat “thanks for your order” with no useful detail wastes the impression. The right pattern is a one-line confirmation with the order number and an explicit ship-by date.
Copy example. Title: Your order is in, Maya. Shipping by Friday. Body: Order #18432, $147.00. We will text the tracking link when it ships.
Touchpoint 2 — Shipping update, day 1 to 3
The second touch fires off woocommerce_order_status_changed to processing or completed. The carrier handoff is the natural moment for the second push. The retention move on this touchpoint is the “track your order” link, which routes back to the store, not to the carrier. A buyer who returns to the store on day two to track shipping is two clicks from another browse session.
The link should land on a tracking page hosted on the store domain that surfaces the carrier status inline, with a “while you wait” section showing two complementary products from the same category. This is one of the highest-return real-estate moves in the entire workflow.
Touchpoint 3 — Delivery confirmation, day 3 to 7
The third touch fires off the carrier delivery webhook or, if the carrier webhook is unreliable, off the estimated delivery date. The subscriber is now a holder of the product. This is the first moment in the workflow where the product itself is in the buyer’s hands. The retention move is “the first 24 hours with [product]” content, a short link to setup tips, care instructions, or a getting-started guide depending on the product category.
For a home-goods store, this might be assembly instructions or a styling guide. For a coffee store, brewing tips. For a supplements brand, the daily-routine recommendation. The content is the retention move; the push is the delivery vehicle.
Touchpoint 4 — Usage check-in, day 5 to 9
The fourth touch is the first proactive check-in. The subscriber has had the product for 48 to 72 hours by the time this fires. The retention move depends on the product category. A coffee grinder triggers a “first brew tips” push. A skincare set triggers a “your day 5 routine” push. A piece of furniture triggers a styling-tip push. The check-in is positioned as care, not as upsell, but the deep link routes to a category-relevant cross-sell page.
This touchpoint also serves as the natural place to cross-sell consumables and accessories. The coffee grinder buyer needs beans. The skincare-set buyer needs the next-tier serum. The furniture buyer needs the matching side table. The cross-sell is contextual, tied to the product purchased, not a generic “you might also like” carousel.
Touchpoint 5 — Review request, day 7 to 14, with SPLIT_PATH on send time
The fifth touch is the canonical post-purchase touch and the only one with a built-in A/B test. The review request notifications fire 7 to 14 days post-delivery, depending on product category. The SPLIT_PATH node sends 50% of subscribers a morning push at 9 AM subscriber-local time and 50% an evening push at 7 PM subscriber-local time. Run the split for two weeks, then promote the winning send time via the workflow’s winner_edge_id.
PushEngage customer data shows morning sends win for coffee, breakfast, and supplement brands. Evening sends win for entertainment, home goods, and lifestyle brands. The split runs on a per-store basis; do not assume the winner from another store’s data.
The deeplink target depends on the review platform. Native WooCommerce reviews route to the product page review section. Trustpilot and Yotpo integrations route to the hosted review page. Confirm the integration before lifting this touchpoint.
Copy example. Title: How is your French press treating you, Maya? Body: Drop a quick review. We read every one.
Touchpoint 6 — Replenishment or cross-sell, day 10 to 14, with DECISION on product category
The sixth touch is the one that directly moves repeat-purchase rate. The DECISION node routes the subscriber based on whether the purchased product is a consumable (coffee beans, supplements, candles, skincare, pet food) or a durable (furniture, electronics, apparel, decor). Consumables route to a replenishment push with a category-specific cadence: beans at day 14, supplements at day 20, candles at day 30. Durables route to a cross-sell push for a complementary category.
This is the touchpoint that pays for the workflow. A 5% bump in replenishment-order rate is meaningful revenue on a customer base of 50,000 buyers. The cross-sell variant produces a smaller but compound lift on AOV for the second order.
Copy example, replenishment. Title: Running low on beans, Maya? Body: 12oz refill in stock. One-tap reorder available.
The DECISION nodes that make this workflow contextual, not formulaic
The same workflow runs every paying customer through six touchpoints, but it does not run them through the same six touchpoints. Three DECISION nodes route subscribers down different branches of the same workflow. This is what separates a generic post-purchase workflow from a contextual one.
The first DECISION fires after Touchpoint 3 (delivery confirmation): is this a first-time customer or a repeat buyer? First-time customers route to a stronger usage check-in at Touchpoint 4, framed as onboarding. Repeat buyers route to a lighter usage check-in framed as appreciation. The same touchpoint, two different tones, one DECISION.
The second DECISION fires before Touchpoint 6 (replenishment vs. cross-sell): is the purchased product a consumable or a durable? The category is read from the WooCommerce product taxonomy at workflow start time and stored on the subscriber attribute last_order_category. Consumables route to the replenishment branch. Durables route to the cross-sell branch.
The third DECISION fires before Touchpoint 5 (review request): is the order value above or below the store’s average order value? High-AOV orders get a slightly delayed review request (day 10 instead of day 7) because the product is more considered and the buyer needs more time to form an opinion. Below-AOV orders get the standard day-7 review request.
Three DECISION nodes, one workflow, many tailored journeys. The same post-purchase workflow handles a $30 candle order and a $400 dining chair order without producing the same touchpoints for both.
The SPLIT_PATH on send time for the review request
Review request notifications are the only touchpoint in the workflow with a built-in A/B test, and the test is composed as a workflow node, not bolted on as a separate config. The SPLIT_PATH node allocates 50% to a morning send and 50% to an evening send, with load balancing handling the per-subscriber routing so that small-sample noise does not skew the split.
The metric is not click-through rate. The metric is review-submission rate within 72 hours of the send. CTR favors evening sends because subscribers are more likely to tap a notification at home, but submission rate often favors morning sends because subscribers actually complete the review when they have a few minutes with their coffee, not when they are about to start a Netflix episode. Optimize against the right metric.
After two weeks at meaningful volume (a minimum of 800 subscribers per branch is the rough rule of thumb), promote the winner via winner_edge_id. The workflow then routes 100% of subsequent subscribers to the winning send time, and the SPLIT_PATH disappears from active use without being deleted from the workflow.
Exit criteria and why this workflow is different
大多数工作流在达成目标时退出。购物车工作流在 purchase 时退出。挽回客户工作流在 last_active < 7 days 时退出。购后工作流则不同。它不会在大多数信号时退出。它会在第 14 天自然完成。
原因是工作流的职责。恢复工作流具有二元成功条件：订阅者已恢复或未恢复。工作流的职责是推动订阅者越过终点线。订阅者一旦越过，工作流的职责就完成了，然后触发退出。购后工作流没有二元成功条件。没有“订阅者完成了购后旅程”这样的说法。每个触点都是其自身的留存策略，而只打开触点 3 和 5 的订阅者也从工作流中获得了价值。
确实触发的唯一退出信号是成功退出：目标 next_purchase_within_60d。在首次购买后 60 天内下第二次订单的订阅者，即完成了工作流的实际职责。工作流会优雅地退出，并让下一个购后实例（针对新订单）接管。多次顺序运行类型确保了没有重叠。
订阅者退订会根据标准的 PushEngage 工作流终止规则终止工作流，但降级或流失信号则不会。此工作流每个订单一次性运行，其职责是在任何自然结束之前最大化该付费关系的价值。
静默时间与时区尊重
中端 WooCommerce 商店通常有国际订阅者。晚上 7 点的网站时间评论请求，对于亚太地区的买家来说，相当于当地凌晨 4 点。购后工作流设置静默时间 start_at: "22:00" 和 end_at: "08:00"，并带有 reschedule 回退。本应在静默时间内发送的通知将被重新安排到第二天早上 8:01 发送，以订阅者的时区为准，然后回退到网站时区，最后是 UTC。
skip 回退是此工作流的错误选择。skip 会静默丢弃通知并从分析中移除。通知永远不会送达，工作流会完全错过一个触点。对于一个其全部职责是提高复购率的工作流，每个触点都必须送达。reschedule 可确保这一点。
侧边栏：通过您的 AI 助手管理购后工作流
WordPress 插件 4.2.4 及更高版本通过 Abilities API 注册了两个 WooCommerce 限制的功能：list-push-automation-campaigns 和 update-push-automation-campaign。运行支持 MCP 的助手（Claude、ChatGPT 桌面版、Cursor）的留存经理无需打开 WordPress 管理界面，即可列出活动的购后工作流、启用或禁用特定工作流，以及更新管理员或客户配置。工作流本身保留在 PushEngage 中；这些功能将配置界面暴露给团队日常运营所使用的任何助手。有关完整的功能列表，请参阅 PushEngage AI 助手公告 和 WP 插件 4.2.4 发行说明。
每次工作流的分析：阅读复购率漏斗
当分析显示正确的数字时，WooCommerce 推送通知在损益表上很容易证明其价值。工作流级别的分析在每个节点显示三个数字：排队中、已完成和已退出。对于购后工作流，每个触点的表格在 14 天窗口内对 10,000 订单队列的显示如下：
| 触点 | 排队中 | 已完成 | 已退出 | 备注 |
|---|---|---|---|---|
| 开始（订单已放置） | 0 | 10,000 | 0 | 所有条目继续 |
| 触点 1（订单确认） | 0 | 9,920 | 80 | 80 个在结账时退订 |
| 触点 2（发货更新） | 220 | 9,640 | 60 | 60 个退货或取消 |
| 触点 3（送达确认） | 380 | 9,180 | 80 | 80 个标记为送达问题 |
| 触点 4（使用情况签到） | 410 | 8,820 | 350 | 350 个退出目标 next_purchase_within_60d |
| 触点 5（评论请求） | 290 | 8,180 | 640 | 640 个提前再次购买 |
| 触点 6（补货/交叉销售） | 0 | 7,890 | 290 | 290 个窗口中期退订 |
| 结束 | 不适用 | 7,890 | 不适用 | 自然完成 |
此表中最重要的数字是 1,280 名订阅者（350 + 640 + 290），他们在工作流完成前进行了重复购买。这相当于 12.8% 的工作流内复购率，在 60 天的测量窗口内，该比率会与 22% 的基线复合增长。
队列级别的比较是工作流旨在提高的回访 KPI。在第一个季度进行 A/B 测试。一半的新订单将通过购后工作流。另一半则不通过。在首次购买后的 60 天，测量两个队列的重复购买率。
| 队列 | 60 天时的重复购买率 |
|---|---|
| 对照组（无购后工作流） | 22% |
| 工作流队列 | 28% |
每月 10,000 笔订单，平均订单价值 (AOV) 为 145 美元，增加 6 个百分点的提升，每月可带来 87,000 美元的额外重复购买收入。该工作流在运行的第一周就能收回成本。在一个季度内，该项目是留存团队能获得的最便宜的收入。这就是正确衡量的重复购买自动化：不是每次通知的点击率，而是证明工作流旨在提高的指标的提升的队列比较。
在 PushEngage 中构建
PushEngage 工作流构建器提供了一个购后模板，该模板映射到上述六个触点。安装 PushEngage WooCommerce 集成插件（版本 4.2.4 或更高版本支持 Abilities API），将 woocommerce_payment_complete 映射到 order_placed 自定义事件，然后应用该模板。静默时间段和 SPLIT_PATH 节点已预先配置。DECISION 节点需要产品类别分类映射，该映射由集成插件自动从 WooCommerce 读取。
对于更广泛的留存堆栈，此工作流与系列中的其他四个 WooCommerce 推送通知工作流并行运行：WooCommerce 购物车放弃工作流、WooCommerce 浏览放弃工作流、WooCommerce 降价提醒和WooCommerce 补货通知。每个工作流都在客户生命周期的不同层面运行。购后工作流是唯一直接提高重复购买率的工作流。其他四个用于挽回停滞的当前旅程；这个工作流用于构建下一个旅程。
要更全面地了解这些工作流如何融入完整的电子商务留存计划，请参阅电子商务推送通知中心，其中涵盖了这些蓝图所实现的广告系列类型。
PushEngage 免费套餐从第一天起就提供 200 个订阅者、所有渠道以及完整的 Workflows 引擎。这足以在请求预算之前，在受控队列上证明购后工作流的效果。从免费套餐开始，并在不到一小时内启用购后模板。将整个客户群的重复购买率提高 3 个百分点就能为平台带来多倍的回报，并在下次损益表审查时显示为已发现的收入。
