Write-up · Home-goods seller
The cart flow was leaking. Two messages fixed most of it.
By Rhea Delgado, Principal SMS Strategist · Published July 26, 2026 · Figures illustrative and anonymized
A home-goods seller on a large marketplace platform had an abandoned-cart SMS flow converting at 2.1%, below Postscript's published 3.97 to 7.84% benchmark band. Nine weeks later the same flow converted at 5.6% and earned $8.40 per message sent, up from $2.10. Nothing exotic happened: timing, copy length, and a hold-back test did the work.
| Metric | Before | After | Window |
|---|---|---|---|
| Flow conversion rate | 2.1% | 5.6% | 9 weeks |
| Earnings per message | $2.10 | $8.40 | 9 weeks |
| Messages in flow | 4 | 2 | rebuild |
| First-message delay | 15 min | 45 min | rebuild |
What was actually broken
Four messages fired over 48 hours, the first at 15 minutes, each over 300 characters with two links. Attentive's May 2026 analysis of 25 billion sends found messages beyond 240 characters convert 8.5% worse, and this flow proved the point. Message three and four generated 9% of the flow's revenue and 61% of its opt-outs. The rebuild cut to two messages: a 45-minute nudge naming the exact item, and a 20-hour follow-up with the single best incentive the margin allowed.
How we knew the lift was real
A 10% hold-back group got no cart messages for the full nine weeks. Flow-attributed revenue per recipient beat the hold-back by a wide enough margin to clear the platform's attribution noise, which is the check Rhea insists on after watching inflated “flow revenue” claims fall apart in audits. The $8.40 earnings per message lands inside Postscript's published $3.52 to $10.95 cart-flow band: strong, not fantastical.
Steal this checklist
- Delay the first cart message past 30 minutes; let people finish checkout on their own.
- Keep every message between 160 and 240 characters, one link, item named.
- Kill any message whose opt-out share outruns its revenue share.
- Run a hold-back before you brag about flow revenue.
Flow rebuilds like this one are the core of SMS program management; the aggregate numbers across all 38 programs sit on the Results page. Benchmark sources: Postscript's 2026 report and Attentive's copy analysis.