The revenue channel you own outright

Shopify Email is real and fine for straightforward campaigns on a smaller list, and it is priced accessibly. Klaviyo earns its cost when you need segmentation depth, complex flow logic, per-flow revenue attribution, and predictive features, typically once flows drive a meaningful share of revenue. We implement either honestly, and migrating later is routine, so the first choice does not trap you.

In order of typical revenue contribution: welcome, cart abandonment, browse abandonment, post-purchase, and winback. Welcome earns most because it reaches people at peak interest with permission freshly given. For consumable products, a replenishment flow timed to real purchase intervals often outearns everything except welcome. Build these five properly before spending a minute on clever campaign ideas.

Offer something worth an email address, ask at the right moment, and never buy lists under any circumstances. A well-timed popup with a real incentive converts without poisoning the session, and post-purchase opt-in captures buyers at maximum goodwill. Growth that outruns engagement hurts deliverability, so the list you want is engaged and permissioned, not merely large.

Authentication set up correctly, a warmed sending domain, segmentation that stops sending to people who never open, and a sunset policy that removes dead addresses. Deliverability degrades quietly: the dashboard shows sends, the inbox shows spam folder. Most fixes involve sending less volume to better segments, which also happens to raise revenue per send.

For the right moments, yes: back-in-stock, delivery updates, and high-intent cart recovery perform well because texts get read. As a broadcast channel it burns out fast and carries real compliance obligations around consent and quiet hours. We add SMS after email flows are earning, scoped to the messages where immediacy genuinely matters.

Use conservative attribution windows, compare against holdouts where volume allows, and watch revenue per recipient rather than open rates, which inflated further after privacy changes made opens unreliable. The uncomfortable test is switching a flow off briefly: if revenue does not move, the flow was claiming orders, not creating them. We report on numbers that survive that test.