The SMS Recovery List

Run an email opt-in invitation at the 19,970 in RCm4nZ

This is the recoverable list. Email is the only legal channel, because the whole definition of the segment is that there is no SMS consent on record. The ask is a single-purpose email whose only job is to get them back onto a phone field, now that the field no longer requires an off-page confirmation.

Send to a 2,000-profile slice first and read the real opt-in rate. These people already declined to finish once, so 19,970 is a ceiling and not a forecast. If the slice converts at the 10 to 13% Klaviyo quotes for phone-only asks aimed at existing email subscribers Klaviyo benchmark, the full pool is worth roughly 2,296 subscribers, or $17,477 against the measured $7.61 per SMS subscriber. I would not put that in a plan until the slice reports.

Prioritise inside the segment: the 14,707 who are existing customers first, since they carry $2,333,310 of historic revenue and a reason to hear from us.

Repoint the tap-to-text keyword, or retire it

Keyword JOIN still deposits opt-ins into SUmgWj “SMS Subscribers”, a list of 224 profiles created on 19 June. I checked all 54 flows in the account: zero live flows trigger on SUmgWj. The SMS welcome flow TWWDQd triggers on VdWm2S alone.

So every person who ever texted us to subscribe got no welcome message, and at the measured $7.61 per SMS subscriber that is about $1,705 of welcome-flow revenue never earned. The tap-to-text button is gone from the popup now, but the keyword is still live on the toll-free number and still reachable by anyone who texts it. Either point it at VdWm2S or add SUmgWj as a second trigger on TWWDQd.

Find out what is still filling the pool now the popup is fixed

The popup has stopped being the main contributor, because typed numbers now complete at 82.7%. But the pool still grew by 405 profiles in the six days to 20 August, so something else is feeding it. The leading candidate is Shopify checkout, and I have not yet proved it. What I can show is the mechanism: form_settings.shopify_visitor_api on the account carries sync_SMS_consent: true, so checkout phone numbers sync into Klaviyo, and a number typed to get a delivery notification lands in exactly the same NEVER_SUBSCRIBED state as an abandoned popup opt-in. Consent-free numbers emit no event, so they cannot be counted from the event stream; this needs a cohort pull on profiles that gained a phone after 18 August.

This matters because it changes what the pool means, and therefore how item 1 should be written. A number someone typed into a marketing popup is a soft yes we failed to close. A number someone typed to get a shipping notification is not a lead at all, and inviting that person to opt in reads as a cold approach. Split the segment by source before sending anything.

Measure the cost of single opt-in before it becomes a deliverability problem

What we traded away on 18 August was typo protection. A mistyped number now enters VdWm2S as a confirmed subscriber and we pay to text it until it hard-fails. The account-wide failure rate today reads lower after the change than before it, but that comparison is worthless because it tracks campaign volume rather than list quality, so I am not reporting it as a win.

The clean read is cohort-based: take profiles that joined VdWm2S after 18 August, wait for a campaign to hit them, and measure their failure rate against the pre-18-August cohort. If the new cohort fails materially worse, the answer is not to restore double opt-in but to add inline phone validation on the field, which costs nothing in completion.

Get the single opt-in posture signed off

Flipping VdWm2S from double to single opt-in is a legal-posture change, not a UX tweak. What makes it defensible is the TCPA disclosure sitting under the phone field, which is present and correctly worded. What we no longer have is a second signal that the person holding the phone agreed. No sign-off is on record for this yet, and it should be, because the list it governs is now taking on about 222 people a day.

Related and separate: in a 350-profile sample of the recovery pool, Klaviyo had already flagged 15 as MESSAGE_BLOCKED and two as KNOWN_LITIGATOR. Scaled across 39,918 that is on the order of two hundred people who litigate over exactly this category of contact. It is another reason the recovery in item 1 runs over email and never over SMS.