Email capture gets most of the attention because an email address is safe. It survives a device change, it is easy to segment, and it never expires. A phone number behaves differently. It lives in the same app the visitor uses to talk to family and friends, which is why the permission question deserves more care than a checkbox at the bottom of a form.
That closeness cuts both ways: it makes a text powerful, and it makes a careless ask feel like an intrusion. Most markets also treat phone numbers more strictly than email, so getting it wrong costs you more than a subscriber.
Why bother with a phone number when email is easier?
The two channels fail differently. Email fails by being ignored in a crowded inbox. SMS fails by being intrusive, which is the reason it still gets attention.
- Time sensitive messages: a restock on the size someone is reading, a cart reminder inside the hour, a delivery update.
- People who ignore email: a second route to the same visitor, without paying for the traffic again.
- Service that feels like service: reminders that would be awkward as an email but natural as a text.
Skip it where per message costs are high against a low order value.
What should the popup offer in exchange for a number?
A number feels more personal than an email address, so a vague offer gets refused in a way an email signup rarely is. "Sign up for updates" asks the visitor to guess. A promise tied to the current page does the opposite.
| What the visitor is doing | Offer that fits | Why it holds up |
|---|---|---|
| Reading a sold out size | A text when that product is back | The number earns its keep on a real event |
| Building a cart | A reminder by text within the hour | The value is time boxed by nature |
| Just bought something | Delivery updates, then a review nudge | Service first, marketing later |
| First visit, no intent | Nothing yet, keep showing the email ask | A cold text offer teaches people to ignore you |
SMS capture works best as a second step after intent exists, not as a louder email popup.
How should the opt in be worded?
Consent for phone numbers has to be explicit, single purpose, and documented. Grey small print saying "by continuing you agree to receive marketing messages" is not informed consent, and pre ticked boxes fail the same test.
- A separate, unticked checkbox that names the channel. Email consent should not be reused for texts.
- Say what will be sent and how often: "about twice a month" beats "periodic offers".
- Name the sender, so the message does not arrive from an unknown number.
- State the exit: "reply STOP to stop".
- Store the wording shown, the timestamp, and the page it came from. That record is your defence if a complaint arrives.
When should the ask appear?
Timing carries more weight here than with email, because the interruption is bigger. Never on first load, and never while another overlay is fading out. Use intent instead of seconds: a second view of the same product, a size guide opened, a cart created, or exit intent on a product page.
Treat the ask as one interruption among everything else on the page, with the restraint described in popup timing and triggers. If someone just gave you an email address, do not immediately ask for more.
How should the field and the confirmation work?
Keep the number collection to one field, with a country hint inferred from the page and an input that opens a numeric keypad. Validate only after submission, so a half typed number does not turn red while someone is typing.
Then confirm. A confirmation text that asks for a keyword reply does two jobs: it proves the person owns the number, and it produces a timestamped record of consent. On desktop, keep the ask off the page or hand the visitor to their phone with a QR code.
Which numbers should you watch?
Judge SMS with the same discipline as every other capture surface, and keep the volume low enough to stay welcome.
- Popup capture rate, split by device, trigger, and offer.
- Confirmation rate, the share who reply and complete the double opt in.
- Revenue per contact, compared against the email list rather than against click rate alone.
- Complaint and STOP rate, your early warning before carriers start filtering you.
- Margin per message, because every send carries a price that email does not.
Compare the copy the way you would any other test, with the approach in A/B testing CTAs, and cap the rhythm with the budget thinking in popup frequency capping. Sunset anyone who has not clicked in months instead of paying to text them, with the same restraint that keeps email capture from feeling like a toll gate.
Ready to test an ask that matches the moment? Build the email and SMS flow with heycustomer.byako.dev
FAQ
Q: Is a phone number harder to capture than an email address? A: Usually yes. Expect a lower conversion rate on an SMS ask than on an email ask with the same offer, and more so on desktop. A confirmed number reaches someone in a narrower window, so compare the two on revenue per contact.
Q: Do I need a double opt in for SMS? A: In most markets it is the safest path, and some require a documented confirmation. A confirmation text with a keyword reply proves ownership and gives you a timestamped record.
Q: Can I ask for a phone number on desktop? A: You can, but expect fewer takers. Either keep the ask to mobile widths, or use a QR code so the visitor confirms on the device where the messages will arrive.
Q: Which popup promise works best? A: The one tied to what the visitor is doing at that moment: a restock for the size they are reading, a reminder for the item in their cart, a confirmation for the slot they are booking. Generic "sign up for texts" offers collect numbers that stop engaging after the first send.