Black Friday sequencing with overlapping segments
A home-fragrance brand has 150,000 email subscribers and plans three Black Friday campaigns: VIP early access on Tuesday, general launch on Thursday, and last-chance on Sunday. The VIP segment has 8,000 profiles who purchased at least twice in the last 12 months or spent over $300. The engaged segment has 45,000 profiles who clicked in the last 90 days. The past-purchaser segment has 30,000 profiles who bought in the last 180 days. These segments overlap heavily. Last year, the brand sent all three campaigns to the full list; spam complaints reached 0.07%, unsubscribes reached 2.4%, and the Sunday send had the lowest revenue per recipient. This year the CMO wants maximum revenue, but the deliverability consultant wants frequency caps and exclusions. The team can use Klaviyo segments, conditional splits in campaigns, and A/B tests, but cannot add SMS before Black Friday.
1. How would you sequence the three campaigns across the segments to balance revenue and deliverability? Specify inclusions and exclusions.
Read the worked answer
Sequence by value and suppress overlap. Tuesday VIP early access: include VIP segment, exclude profiles who purchased in the last 7 days and anyone who complained or bounced. Thursday general: include engaged 90-day clickers, exclude VIP and anyone who purchased since Tuesday. Sunday last chance: include engaged 90-day clickers who have not purchased since the campaign started, exclude VIP unless they did not purchase, and exclude anyone who received both prior sends if frequency cap is two. This gives VIP first access, avoids mailing the same person three times, and preserves the Sunday send for non-converters rather than the whole list.
2. What A/B test would you run on the Thursday campaign, and how would you avoid contaminating results with the other sends?
Read the worked answer
Run the A/B test only on Thursday, on a random subset of the Thursday audience, with one variable such as subject-line style or send time. Hold out the VIP and Sunday audiences, and do not change the offer between arms. Use revenue per recipient as the primary metric and spam-complaint rate as a guardrail. Send the winner to the remainder only after the test has enough data; otherwise you contaminate the comparison with the VIP early access or Sunday last-chance behavior. This isolates the test from the other campaign sends.
3. The Sunday send has low revenue per recipient. What would you check before canceling it, and what would you change?
Read the worked answer
Before canceling Sunday, check attribution window, whether recipients already purchased, whether paid ads or organic traffic drove the conversions, and whether the Sunday audience was simply the least engaged leftover. Compare revenue per recipient by segment, not total revenue. If Sunday only reaches people who already bought, replace it with a non-purchaser reminder or a smaller engaged segment. Run a holdout to measure incrementality; if the holdout performs the same, the send is not adding revenue and should be cut or changed.