How should we segment the list for campaigns?
Our answer
Send most campaigns to subscribers who have lately clicked, visited or bought, and keep buyers and non-buyers apart. Leave recent buyers out of promotions for 7 to 30 days, and segment broadly while your list is under 20K.
- Confidence
- Medium
- Strongest proof
- 2 official rules
- 2 datasets, 3 practitioners, 6 opinions or claims
- Sources
- 8 sources
- 13 statements
- K
DNZDOLCS
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentRecent site visitors drive campaign revenue. In Pilothouse's client data, 75% of campaign revenue came from subscribers who had visited the site within 90 days. Opens are skewed by proxy opens, so segment on clicks, purchases and site behavior (Litmus).
Klaviyo advises sending regular campaigns to engaged segments and keeping the full list for large promotions. It also tells lists under 20K profiles not to segment too narrowly.
Buyers need their own rules. Zettler Digital excludes customers who just bought for 7 or 30 days, and Omnisend keeps VIPs, its top 5 to 10% by lifetime value, out of sitewide discount blasts.
Our judgment
Judge engagement on clicks, site visits and purchases ahead of opens, and keep opens only with Apple privacy opens filtered out. Klaviyo says Apple Mail Privacy Protection inflates opens, Litmus says proxy opens skew them, Zettler Digital finds open and click windows rarely translate to revenue and Pilothouse's data ties campaign revenue to recent site visitors. Leave recent buyers out of promotions for 7 to 30 days, taking Zettler Digital's 7 or 30 day threshold and Omnisend's 30 day definition of a recent buyer. Start with a few broad segments and add recency, frequency and monetary value (RFM) tiers once they run. That combines Klaviyo's advice not to segment too narrowly under 20K profiles with Top Growth Marketing's start of 3 to 5 segments.
Our judgment, built from Klaviyo, Litmus, Zettler Digital, Pilothouse, Omnisend and Top Growth Marketing: our reading of these sources, not a finding from any one of them.
What the sources say
13 statements from 8 sourcesKlaviyo tells senders to mail its engaged segments (opened or clicked in the last 30, 60 or 90 days, or recently joined) and exclude unengaged ones from broad sends to protect deliverability.
“Your engaged segments are the primary groups of people that you should send to.”
Pilothouse, an agency, says in its client email data 75% of campaign revenue came from subscribers who had visited the site within 90 days. It gives no sample size, brands or period.
“75% of campaign revenue came from subscribers who had visited the site within 90 days”
Klaviyo, which bills on sends, advises lists under 20K profiles to send one campaign a week and work toward two. Larger lists should send at least two segmented campaigns a week.
“You should be sending at least two segmented email campaigns per week.”
Klaviyo, which sells the sending it advises, says send most often to your most engaged subscribers. For lists under 20K, don't segment too narrowly and send one conversion email and one traffic email per month to all.
“The general rule to follow is that you should send to your most engaged subscribers most often, and send to your less engaged subscribers less often.”
Selzy, an email vendor, found across 388,431 campaigns that sends to under 100 recipients had a 23.6% open rate versus 5.7% for sends to 100,000+. It says these are patterns, not controlled A/B tests.
“These are real-world patterns, not controlled A/B tests.”
Omnisend, which sells email and SMS, advises keeping buyers from the last 30 days and VIPs (top 5-10% by lifetime value) out of Black Friday discount blasts. No data behind the rules.
“Exclude recent buyers from your discount blasts to avoid messaging fatigue and buyer's remorse”
Litmus, part of deliverability seller Validity, says opens are skewed by proxy opens, so build engagement segments from clicks, purchases and site behavior. Validity's data favors smaller, targeted lists.
“Senders who shifted to smaller, more targeted email lists saw better outcomes.”
Charle, a Shopify agency, says two to four email campaigns a week suits established Shopify brands and one to two for smaller lists, keeping unsubscribes under 0.5% per send.
“The right number is whatever keeps engagement strong and unsubscribes under 0.5 percent per send.”
Top Growth Marketing, a Klaviyo management seller, suggests starting with 3 to 5 high-impact segments, like engaged non-buyers and at-risk customers, and warns brands over-segment without measuring each one.
“Most brands over-segment, end up with overlapping audiences, and never measure performance per segment.”
Ben Zettler, a Klaviyo partner agency founder, says 30, 60 or 90 day click and open segments rarely turn into revenue. He suggests excluding recent buyers and targeting past buyers who have not purchased recently.
“Those segments look good on a dashboard, but they rarely translate to revenue.”
Klaviyo suggests sending by engagement: for a brand sending 3 times a week, 3 a week to people engaged in 60 days, fewer as engagement ages, and a sunset flow for profiles silent past 180 days.
“Sending too often to unengaged profiles will hurt your sender reputation, whereas sending too infrequently to engaged customers is leaving money on the table.”
Klaviyo suggests starting SMS at 2-4 sends a month, ramping to 4-7, up to 1 to 2 a week for highly engaged subscribers, and excluding people with no click in 150 days from most campaigns.
“Do not send SMS every day unless the subscriber knowingly opted in to daily reminders”
Klaviyo defines an engaged segment as opened or clicked in 30 days or subscribed in 15, says sending to unengaged lists hurts deliverability, and shows how to filter out Apple privacy opens.
“Sending to unengaged subscribers puts your sender reputation and deliverability at risk.”
In the full answer
For readers who sign in- What to do this week3 steps
- The numbers behind it3 figures from cited sources
- Where each source standsA chart placing each source on the question
- The strongest disagreementThe best case against the answer, from a cited source
- How sure we areWhat its confidence rests on, and what would change the answer