SMS or Email: Which Message Earns Which Channel

TL;DR. The question is never "should we do SMS" but "which message earns an SMS". Email carries content, catalogue and anything that benefits from length; it remains the volume channel and the cheapest per contact. SMS carries urgency, and nothing else. The rule that makes the whole thing viable is non-overlap: never both channels to the same contact on the same day. And the only measurement that settles anything is the holdout.

The SMS-versus-email debate is badly framed because it compares raw rates. SMS shows spectacular metrics, but it addresses a smaller, more recent, more consenting list. Comparing its rates to email's is comparing a VIP segment to a general audience.

The right comparison happens message by message.

The arbitration grid

What belongs to email

Content. Guides, methods, behind the scenes, editorial. None of it fits in a text, and trying produces a link clicked out of curiosity rather than interest.

Catalogue. New arrivals, selections, recommendations. Email lets you show; SMS forces you to describe.

Anything without a deadline. If the message is still valid in three days, it has no business being a text.

Volume. Email has a near-zero marginal cost. It carries the 2 to 5 weekly campaigns we consider a healthy cadence on an engaged base.

What belongs to SMS

Checkout abandonment. Short window, maximum intent.

Back in stock. The contact asked to be told, availability is limited.

The closing hours of a promotion. Once, on a real deadline.

Cart abandonment above threshold. Provided the threshold is calculated : see SMS marketing cost.

The test that settles it

One question does the work: does this message lose its value if the reader opens it tomorrow?

If yes, it is an SMS candidate. If no, it is an email. Messages that clearly pass neither : and there are many, mostly deserve not to be sent.

The non-overlap rule

A contact does not receive both channels on the same day. It is the rule that makes a multichannel programme viable, and the one brands apply least.

The classic scenario appears within weeks: a third-party SMS tool coexists with the email platform, the two hold lists that do not talk to each other, and the same customer receives the same offer twice in a day. Both tools report good numbers, because both claim the sale.

Applying non-overlap requires both channels to share one customer profile. That is the main reason to consolidate SMS and email on a single platform, more than interface convenience : see Klaviyo omnichannel marketing.

Measuring what SMS actually adds

Native attribution systematically over-credits the most recent channel in the journey. A customer who got an email in the morning, a text in the afternoon and buys in the evening will be credited to SMS by most tools.

Attribution is also where the two channels are least comparable. Email opens stopped being a usable signal after iOS 15 pre-fetching, so email gets judged on clicks and revenue; SMS has no open at all, only a click and a delivery receipt. Comparing a channel measured on clicks with one measured on opens produces a verdict about the measurement, not about the channel.

The only reliable measure is the holdout: exclude a random sample from the SMS send and compare its revenue to the exposed group. The difference is real incrementality.

The result often surprises. On high-intent flows, incrementality is high. On broad campaigns it is frequently near zero: the sales credited to SMS would have happened anyway, carried by email. The measurement framework is in email marketing attribution.

The indirect effect on email deliverability

The two channels run on separate infrastructure: a badly received text does not degrade your email sender reputation.

The effect exists nonetheless, by another route. A contact saturated by SMS disengages from the brand as a whole. They stop opening emails, then stop clicking. And click is the only engagement metric that survived iOS 15, and the one your placement depends on : we target 1.5 to 2.5% on an engaged base, and below 1% emails get opened but no longer sell.

In other words: over-sending on SMS costs email revenue, without the link appearing in any report.

This is the case for reading both channels on one dashboard rather than two. A brand whose SMS revenue climbs while total revenue stays flat has not gained a channel: it has moved purchases from one to the other, and it is now paying per message for what it used to get at close to zero marginal cost.

The frequency cap, applied to both

The practical consequence is that you need a global cap, not a per-channel one.

A contact receiving three email campaigns and two texts in a week has received five solicitations, whatever the split. That total is what needs steering, watching two alarm signals: above 0.5% unsubscribes per campaign on average, pressure or targeting need revisiting. On complaints, Google asks senders to stay under 0.30% and recommends holding under 0.10%, a margin that absorbs the spikes of a busy week. That second figure is the one we apply internally, so the margin is still there when Q4 arrives.

Consent does not carry across channels

This is the difference arbitration exercises forget most often, and it is not a technical one.

In France, the CNIL requires prior consent that is free, specific, informed and unambiguous for electronic marketing addressed to a private individual, collected through a positive action: an unticked box that the person ticks. Pre-ticked boxes are excluded.

In practice, an email address collected without any mention of SMS does not entitle you to send a text. The number has to be collected for that use, with the purpose stated at the point of collection. That is what makes two-step collection sturdier than a single form that scoops up everything.

The channel also carries a public risk email does not have: an unsolicited text can be reported to 33700, the French platform for fraudulent and unwanted SMS. A report is not undone by fixing your targeting the following week.

What this means for arbitration: a scenario that stacks up on paper does not launch if the phone numbers were not collected for it. The topic is covered in SMS and GDPR in France.

FAQ

Do we have to choose between SMS and email?

No, but you have to choose for each message.

Does SMS convert better than email?

On high-intent flows yes. On broad campaigns no, and the gap reverses once cost is included.

How do we know SMS is really adding revenue?

By holdout. Without it you credit SMS with sales email would have captured.

Can we send an email and an SMS on the same day?

To the same contact, no.

Can we text people who gave us their email address?

No. In France the CNIL requires consent that is specific to the channel and collected through a positive action, with the purpose stated at the point of collection. An address obtained without any mention of SMS does not cover the phone number, and an unsolicited text can be reported to 33700.

Does SMS damage email deliverability?

Not directly, but a saturated contact stops opening and clicking, which degrades the signal your placement depends on.

Going further

The channel framework is in SMS marketing: definition and examples, and programme scoping on our SMS marketing agency page.

Review your email and CRM stack with Deliver