Online pizza ordering has a high abandonment rate: commonly somewhere between 40% and 70% of started carts never complete. Unlike retail abandonment, which is often idle browsing, a built pizza cart represents real intent within a narrow time window. The recovery message therefore has to be fast, and the diagnosis behind it is often more valuable than the recovery itself.
A retail abandoned-cart email sent four hours later can still work; someone will buy the shoes tomorrow. Nobody orders Friday dinner on Saturday morning. If the message arrives after the household has eaten, it is not a recovery message, it is an advertisement.
| Delay | Channel | Message |
|---|---|---|
| 20–30 min | Push or SMS (if consented) | "Your order is still here" + one-tap resume. No discount. |
| 60–90 min | Resume link, cart contents, delivery time estimate | |
| Next equivalent daypart | Soft re-offer. Optional small deferred value. | |
| No further contact | : | Two attempts is the limit. A third reads as surveillance. |
The first recovery message should never carry a discount, for the same reason rung one of a win-back should not: a meaningful share of abandonments are interruptions rather than rejections. The phone rang, someone came to the door, the app crashed. Those customers need a link, not a bribe.
Discounting immediately also teaches a lesson you do not want taught. A customer who learns that abandoning a cart produces a discount will abandon every cart, and you will have built a self-inflicted price cut with extra steps.
The recovery campaign is worth running. The abandonment data is worth more, because it points at operational problems that are costing you orders you never see.
Every abandonment report is a bug report written by someone who was trying to give you money.
You can only recover abandonments where you have an identity, which is another argument for capturing contact details early in the flow rather than at the payment step. Asking for a phone number or email before the address step raises identified abandonments substantially and costs very little conversion if the ask is framed as order tracking.
Between 40% and 70% of started carts is typical, with better-optimised direct ordering sitting at the lower end. What matters more than the level is the step where it happens: a high rate concentrated at one step is a fixable problem rather than a fact of life.
Only with prior express consent for marketing messages, and it should be treated as promotional rather than transactional. The 20-minute window makes it the most effective channel for this use, which makes consent capture during the ordering flow genuinely valuable.
You cannot. You have no visibility into them. That asymmetry is one more reason to work on direct-ordering share: an abandoned direct cart is recoverable and an abandoned marketplace cart is invisible.
Keep reading
Cutting friction out of direct ordering, because every point of conversion is worth more than any campaign, and every direct order is margin you keep.
Audience & dataPractical ways to turn anonymous tickets into identified customers (online ordering, loyalty, QR on the box, the till, and wifi) without annoying anyone.
Campaigns & offersThe seven automated journeys every pizza location should have running (welcome, second order, habit, drift, win-back, birthday and post-catering) and the order to build them in.
Campaigns & offersWhy restaurant send-time is different from retail: timing against the meal decision rather than against open rates, and per-customer send windows.
marketing.pizza runs all of it, every night, across every store you have.