The rule in one sentence: SMS owns the moments where minutes matter and one tap finishes the job; email owns everything that needs proof, education, or more than a sentence, and the flows that pair them beat either alone. The channel data is unambiguous about where SMS earns its keep: flow-based SMS clicks near 10% on average with top performers past 16%, and SMS flows account for just 7.6% of sends while driving 45.2% of total SMS revenue, at roughly 8 times the revenue per recipient of SMS campaigns. Automation is where the channel concentrates its value, exactly as it is in email, and the design question is never email or SMS: it is which moment belongs to which.
The division of labor: speed versus depth
SMS is opened in minutes, read in seconds, and resented at length: the channel's power and its danger are the same property. Email tolerates paragraphs, images, and proof stacks, and forgives frequency SMS never will. So the assignment rule across every flow in this library: the time-critical nudge and the one-tap action go to SMS; the objection-killing, the education, and the storytelling stay in email; and consent economics referee the edges, because an SMS opt-in is scarcer and more revocable than an email address, and every send spends some of it.
Flow by flow: the assignments
Back-in-stock: SMS leads, email completes
The clearest SMS win in the account: restocks are races, and Klaviyo's data shows the same restock moment clicking at 5.3% on SMS against 1.9% on email, with SMS converting at 8.5% against email's 7.9%. The waitlist build offers the SMS lane at opt-in and sends both channels on the inventory event; the email carries the fit notes and proof for the hesitant, the text wins the sprint for the decided.
Abandoned checkout: SMS opens, email persuades
The staggered pattern from the checkout builds: a single SMS shortly after abandonment (the reminder, the restored-cart link, one line) followed by the email sequence that does the actual objection work: the proof, the guarantee, the offer discipline. Two texts about one checkout is where recovery becomes harassment; the channel gets one shot, early, and hands off.
Transactional: SMS for the doorstep, email for the record
Delivery-day and out-for-delivery notices are phone-native moments; the receipt, the tracking detail, and the education stack live in email per the owned-stack build. The same primacy rule governs both channels: the operational payload leads, and marketing rides only where it does not obscure.
Dunning: mirror the middle touch
A card fix is a phone-native task, which is why the dunning build mirrors its day-3-to-4 email on SMS where consent exists: one text, the no-shame framing, the hosted update link. The deadline email stays email, because consequences deserve the fuller channel.
Replenishment and winback: email first, SMS as the single closer
The persuasion-heavy flows stay email-led: replenishment's usefulness framing and winback's what-changed argument need room. The one defensible SMS in each: the final-window nudge (your refill window closes today; the one-tap reorder) for customers who click texts but skim email, gated to profiles whose behavior says the channel works on them.
Welcome, browse, and the education arcs: email, almost entirely
The welcome series is proof and story, the exact content SMS cannot carry; the defensible welcome text is a single hello that delivers the offer code for wallet convenience. Browse abandonment on SMS reads as surveillance and stays off the channel entirely, per the service-not-pursuit rule in the browse build.
Consent economics: the budget behind the assignments
Every assignment above assumes the subscriber gave both consents, and the scarcer one gets spent slower. The working rules: collect SMS consent where its value is obvious (the back-in-stock opt-in, the delivery-updates checkbox, the checkout), never as a popup afterthought; cap flow SMS to the moments listed here so the channel's open-in-minutes power survives; and route by demonstrated behavior, because the customer who never taps texts but reads every email is telling you their assignment, and Klaviyo's engagement data lets the flows listen. The revenue concentration statistic is the discipline's reward: nearly two-thirds of SMS flow revenue comes from new buyers, the exact audience whose consent is freshest and whose patience is thinnest.
What to measure
- SMS flow clicks against the near-10% average and 16% top mark, per flow, because a text underperforming email in its own slot is in the wrong slot
- Revenue per recipient by channel per moment: the restock, the checkout nudge, the dunning fix, each judged separately rather than as channel totals
- Opt-out rate per SMS touch, the consent-budget ledger: any flow text whose opt-outs outpace its conversions is spending principal, not interest
- Channel-preference drift: the share of dual-consent profiles engaging only one channel, which should re-route their assignments quarterly
Frequently asked questions
Should abandoned cart recovery be email or SMS?
Both, staggered: one SMS shortly after abandonment for the fast nudge, then the email sequence carrying proof and offer discipline. SMS gets one shot; the persuasion is email's job.
Where does SMS clearly beat email?
Time-critical, one-tap moments: back-in-stock (5.3% vs 1.9% clicks on the same event in Klaviyo's data), delivery-day notices, and the card-fix nudge in dunning. Speed is the channel's entire advantage.
Which flows should never use SMS?
Browse abandonment (it reads as surveillance) and the education-heavy arcs: welcome storytelling, post-purchase onboarding, winback persuasion. If the message needs a second sentence of proof, it needs email.
Why do SMS flows outperform SMS campaigns so heavily?
Same reason as email: triggered timing. SMS flows are 7.6% of sends but 45.2% of SMS revenue at roughly 8 times the revenue per recipient, because the automation fires when the moment is live instead of when the calendar says.
How much flow SMS is too much?
When opt-outs outpace conversions on any touch. The channel's consent is scarcer than email's; the assignments above cap flow SMS at the handful of moments where minutes genuinely matter.