Klaviyo Interview Questions & Guide

A List is a static, manually managed group (people you added or who signed up to it); a Segment is a dynamic, rule-based group that updates automatically as profiles meet or stop meeting conditions. Strong candidates default to segments for targeting and reserve lists for explicit opt-ins, and they note segments re-evaluate continuously.

A profile is the person; metrics are the events Klaviyo tracks about them (Placed Order, Viewed Product, Opened Email), usually synced from the store. The signal: understanding that metrics — with their properties and values — are what you build flows and segments on, not just email engagement.

A Campaign is a one-time, scheduled send to a chosen audience (a newsletter, a promo). A Flow is automated and triggered by an event, profile action, or date, running per-individual. Strong candidates say the bulk of revenue should come from well-built flows, with campaigns layered on top for timely moments.

Look for the revenue-driving set: welcome series, abandoned cart, browse abandonment, post-purchase, and win-back. Strong candidates prioritise abandoned cart and welcome first (highest intent/revenue), and tie each flow to a specific trigger metric rather than listing them generically.

A trigger is the event that starts a person into the flow (e.g. Checkout Started); flow filters and trigger filters control who actually proceeds and prevent re-entry or messaging people who’ve since converted. The senior signal: using filters to avoid emailing someone an abandoned-cart series after they’ve already purchased.

Strong answers: trigger on Started Checkout, add conditional splits and a flow filter that exits anyone who has placed an order, space messages with sensible delays, and cap the series length. Bonus for testing incentive vs no-incentive and measuring incremental revenue, not just opens.

Klaviyo predicts CLV, expected next order date, and churn risk for profiles with enough history. Strong candidates segment on these (e.g. high-value-at-risk for a win-back, predicted-next-order for replenishment) and acknowledge the data thresholds before predictions are reliable.

Dynamic segments combine profile properties and metric/event conditions (‘placed 2+ orders in 90 days’, ‘viewed a product but didn’t buy’) and update automatically. The signal: building behaviour-driven segments off metrics rather than static demographics, and using them as flow filters and campaign audiences.

The integration syncs orders, products, and browsing events as metrics, plus catalog data for product blocks. Strong candidates use that synced data for replenishment, category-based segmentation, and dynamic product feeds — and they sanity-check that the integration is actually passing the events they rely on.

A sunset flow identifies chronically unengaged subscribers, gives them a last-chance message, and then suppresses them. The mature answer ties this to sender reputation: mailing dead addresses drags inbox placement for everyone, so pruning protects deliverability and is worth the list-size hit.

By targeting recent engagers and gradually re-including lapsed ones rather than blasting the whole list, you keep open/click rates and reputation high. Strong candidates describe warming a new account/domain by starting with the most engaged segment and ramping volume, and managing unengaged subscribers separately.

Explicit consent collected compliantly (TCPA/local rules), separate SMS subscription, clear opt-out handling, and respecting quiet hours and frequency. The signal: treating SMS consent as distinct from email consent and never assuming an email opt-in covers SMS.

Klaviyo attributes orders to a message within a conversion window after open/click. Revenue per recipient (RPR) normalises revenue by audience size so you can compare sends of different sizes. Strong candidates lead with RPR and attributed revenue over open rate, and know the attribution window is configurable.

Look for: comparing against Klaviyo’s industry benchmarks to spot weak flows, then A/B testing subject lines and content one variable at a time, reading results on conversion/RPR with adequate sample size. The senior signal is iterating on flows continuously rather than treating them as set-and-forget.