StackEvidence Guides · Article index · StackEvidence home
Reader question
Everyone says abandoned-cart emails "recover sales." Before I switch one on, what does it really cost me to run — the tool, the discount I give away, my time, deliverability risk, and support — and how do I work out whether it pays for itself on my numbers?
The short answer
Price abandoned-cart recovery as five costs, not one: (1) tool or subscription, plus per-message charges; (2) discount or free-shipping margin given away on recovered orders — including orders that would have completed anyway; (3) setup and maintenance time; (4) deliverability cost, when extra mail hurts inbox placement because it goes to unengaged or non-consenting addresses; (5) support cost, in recovery-link, discount-code, and "I didn't ask for this" messages.
A programme pays for itself only when the margin on orders it caused — not all recovered orders — exceeds all five costs together. No recovery rate is promised here, because none was verified; treat any vendor-quoted rate as a sales claim, not a forecast for your store.
Step zero: the consent gate
Before costing a recovery series, verify whether the proposed message can lawfully be sent to each recipient in the relevant jurisdiction. Checkout entry alone is not universal marketing permission, and a recovery promotion is not automatically a transactional message. Respect opt-outs, keep the relevant permission records and confirm current requirements with authoritative guidance. This article does not determine legal eligibility or recommend sending any campaign.
Who can you actually email?
Recovery only works for people who entered an email at checkout or were already subscribers; anonymous abandoners cannot be emailed at all. Define your denominator (emailable abandoners) before discussing any result: 12 orders from 100 emailable abandoners is a different claim from 12 from 140 abandoners, most of whom could never be mailed.
The five costs, one by one
1. Tool and per-message charges. Use the current quote for your contact and send volume, including channel charges, overages and any required plan upgrade. Prices can depend on sends, profiles or emailable contacts. No current vendor price is asserted here; the $20 example below is invented.
2. Discount margin given away. Every discount redeemed on a recovered order is margin surrendered — including discounts on orders that would have completed without the series, a cost that can be easy to miss when comparing subscription prices.
3. Your time. Setup once, plus maintenance each month, at an hourly value you choose. A $0-per-hour option is allowed below, but must be labelled "time valued at $0 for this test" — not "time is free."
4. Deliverability. Extra mail to unengaged or non-consenting addresses can hurt inbox placement for everything else you send. It cannot be quantified from the sources used here: monitor unsubscribe rate, spam-complaint count, and whether recovery mail goes only to consenting, emailable addresses, rather than forecasting it as a percentage.
5. Support. Count recovery-related messages — broken links, expired codes, "stop emailing me." A pattern of them signals the series itself is the problem.
The sequence that keeps costs down
Begin with the smallest permission-compliant reminder flow that answers a real checkout problem. Do not assume every abandoned cart needs an incentive or a multi-message series. Set a clear stop rule for purchase, opt-out and repeated non-engagement. Test that the recovery link restores the intended checkout securely without exposing another person's details. Choose timing and channel only after verifying vendor controls and recipient eligibility; no universal cadence or conversion uplift is claimed.
Worked example: separating caused orders from anyway orders (EXAMPLE)
Fictional teaching inputs, not measured results or a forecast. A seller has 100 eligible, emailable abandoned checkouts in a month, a hypothetical $20 tool cost, and 12 later completed orders. For arithmetic only, assume a defensible comparison established that 7 orders were incremental and 5 would have occurred anyway. Email opens, clicks, coupon use or a short attribution window cannot establish that causal split by themselves. In practice use a suitable comparison or holdout with adequate data, or record incrementality UNKNOWN.
Product price is $40 and margin before recovery incentives is $25 per order. Four of the 12 orders use a $4 incentive, so total incentive cost is $16. Monthly work is 2 hours valued at a hypothetical $25 per hour, or $50.
- Incremental margin before incentives: 7 × $25 = $175.
- All recovery incentives, including on anyway orders: $16.
- Tool plus valued time: $20 + $50 = $70.
- Net before any additional support or deliverability cost: $175 − $16 − $70 = +$89.
Count incentives once, including those that subsidise orders that would have happened anyway. The result cannot support a real decision until the incremental-order estimate and all relevant costs are credible. Counting all 12 orders as caused would overstate the benefit.
Your five-cost worksheet (one month, your numbers)
Run this for one real month:
- Tool cost: subscription plus per-message SMS and WhatsApp charges actually incurred.
- Discount cost: discount value redeemed on recovered orders, include incentives on all attributed orders. Codes and link parameters do not establish causality; record the incremental-versus-anyway split UNKNOWN without a defensible comparison.
- Time cost: setup hours once, plus maintenance hours per month, times your hourly value (with the $0-per-hour option labelled as above).
- Deliverability check: unsubscribe rate, spam-complaint count, and confirmation that recovery mail goes only to consenting, emailable addresses; any rise against your normal broadcast baseline is a cost flag.
- Support cost: count of recovery-related support messages that month.
Decision rule (INFERENCE — a decision aid, not a vendor limit): continue only if caused margin exceeds costs 1 + 2 + 3 for two consecutive months, with support and deliverability flags at zero or explained. Otherwise shorten the series, remove the discount, or pause.
The $0 way to start
Inspect any included recovery controls in a platform you already use. If no new subscription is required, message charges and your time may still matter. Manual follow-up is subject to the same permission and opt-out requirements as automation; an existing relationship alone does not settle eligibility. Do not start sending until that gate is established.
When to pause
Pause if the consent basis cannot be confirmed for a jurisdiction you sell into, if the caused-versus-anyway split is UNKNOWN for two months running, if unsubscribe or complaint counts rise against your baseline, or if support messages about the series keep arriving. Pausing may reduce variable costs, but a subscription can continue until separately changed; one running on unconsented addresses or uncounted discounts costs reputation and margin together.
What to read next
Check the underlying delivery and operating responsibilities first: hosted versus WordPress checkout. This costing worksheet does not activate or test any checkout.
Sources
Authoritative permission references checked 26 October 2026; consult the current rule that applies to your recipients:
The five-cost worksheet and all numeric examples are editorial teaching methods. No vendor recovery rate, pricing or causal experiment result is asserted.
Limitations
No recovery campaign or experiment was run. No recovery rate, causal split or quantified deliverability effect is claimed. The example assumes incrementality solely to teach arithmetic. If that input is unknown, real profitability is unknown. Permission is jurisdiction-specific; this is not individualized legal advice.
Published by StackEvidence. Contact · Privacy. Documentation-based education; examples are fictional and no hands-on vendor result is claimed. Hosted on BlogBaser.