The Welcome Series That Turns Subscribers Into First Orders

A welcome series email sequence has one job: the first order. Someone just handed you their inbox, the most attention they will ever pay you for free, and every send either moves them toward that order or spends that attention on nothing. Below is the series we build: what each send says, the metric it is judged on, where a discount belongs and what it costs, how to split the flow, and how to measure it honestly. Build it from this. You do not need us to run it.

The series, send by send

Five sends, one idea each. The timing below is our default rather than a rule, and the trigger matters more than the clock.

• 1. The delivery. Trigger: Immediately on the list add. Job: Pay the promise, make the case. What it says: Any promised incentive in the first line, then the strongest plain reason to buy: what it does, for whom, why it beats the obvious alternative. Judged on: Placed order rate.
• 2. The proof. Trigger: Once send 1 has had a full cycle in the inbox. Job: Replace your claim with someone else's. What it says: Customers on camera, before and after, the product in use, a review with a name on it. Judged on: Revenue per recipient.
• 3. The objection. Trigger: After send 2, only to people who haven't ordered. Job: Remove the thing that stops your buyers. What it says: The real blocker: shipping cost, sizing, the cheaper competitor, whether it works for their case. Answered head on. Judged on: Placed order rate.
• 4. The context. Trigger: After send 3, still only to non buyers. Job: Make the brand worth choosing over a generic. What it says: Where the product came from, how it's made, what you refuse to do. This send earns full price. Judged on: Revenue per recipient.
• 5. The close. Trigger: Last in the sequence. Job: Ask directly, end cleanly. What it says: The one product to start with, the honest reason now beats later, one link. Any unspent incentive lands here. Judged on: Placed order rate, with complaint rate as the guardrail.

Two rules hold the table together. One idea per send, because a welcome series is a sequence of single arguments and not a brochure in installments. And buyers leave. In Klaviyo that's a profile filter such as "has placed order zero times since starting this flow", rechecked before every message rather than only at entry, so someone who buys after send 2 is skipped for sends 3 to 5 (Klaviyo Help Center). Once the order lands they belong to the post-purchase flow, not to a sequence still pitching the thing in their hallway. The rest of the stack sits in our Klaviyo flows walkthrough.

One platform detail before you build. A list-triggered flow sends to a profile once. If someone finishes the series, unsubscribes, then subscribes again, Klaviyo will not run them through it again (Klaviyo Help Center). Plan it as a one-shot.

Where the incentive belongs, and what it costs

The incentive goes exactly where you promised it. If the popup said fifteen percent off, it's in send 1, first line, above the fold. Making people hunt for a code you promised burns trust at the one moment you had some.

If you promised nothing, most brands still open with a discount they never owed, which teaches the list that the price on the site is a suggestion. The stronger opener is the reason the product is worth what it costs. Hold the incentive for send 5, where it reads as a close rather than a reflex. Then run the math before you pick a number.

Contribution margin per order. CM = (AOV x gross margin rate) minus variable order costs: payment processing, pick and pack, shipping you subsidize, a provision for returns. AOV is average order value.

What the code takes. A discount of d cuts revenue by (d x AOV) and cuts contribution margin by the same dollars, because none of your unit costs move. CM_discounted = CM minus (d x AOV).

The lift the discount has to earn. Breakeven order lift = (CM / CM_discounted) minus 1. That's the share by which orders must rise just to hold contribution flat. On thin-margin categories that bar is high enough that most incentives lose money quietly.

Judge it per subscriber, not per order. Contribution per subscriber = placed order rate x CM per order. A code that lifts placed order rate while cutting CM harder is a worse series with a better-looking conversion chart.

Then judge the series on lifetime gross profit to CAC. CAC is customer acquisition cost, what you pay to buy one new customer. Your list has its own version: subscriber acquisition cost, the spend behind the popup traffic divided by subscribers added. First-order return on ad spend will happily tell you the discount worked while the cohort's gross profit falls. More of this arithmetic sits in eCommerce unit economics.

The question is never whether a discount lifts conversion. It does. The question is whether it lifts contribution per subscriber, and whether the customers it buys come back at full price.

Split the series before you write a word

One welcome series for everyone is the default setting, and it's why most of them underperform. In Klaviyo a conditional split branches on the same conditions the segment builder offers, including profile and custom properties, with multiple ordered paths and an "everyone else" catch-all in current versions (Klaviyo Help Center). The splits worth building, in build order:

• Buyer versus non buyer. The highest-value split and the cheapest to build. Handle it with a profile filter so it's rechecked before each send.
• Signup source. A popup subscriber, a footer subscriber, a giveaway entrant and a quiz finisher have different intent. Stamp the source as a profile property at signup and branch on it. Giveaway entrants deserve a shorter path of their own.
• Product interest. Quiz answers, the collection they browsed when the popup fired, or the popup variant they saw. Send 1 should name the thing they were actually looking at.
• Channel consent. Email plus SMS subscribers should not get the same message twice in an hour. Smart Sending caps messages per profile in a window, defaulting to sixteen hours for email and twenty four for SMS and push, and it can be turned off for individual flow messages (Klaviyo Help Center). Leave it on here.
• Gift versus self purchase. Different objection, different proof, often a different first product.

Do not build all five at once. Build the buyer split, then source, then product interest, and stop when the next branch stops earning its maintenance.

How to measure it honestly

Most welcome series reporting is a screenshot of open rates, and open rates stopped being evidence. Apple's Mail Privacy Protection inflates them, which is why Klaviyo recommends click rate over open rate as the winning metric when you A/B test a flow email (Klaviyo Help Center). Opens are an instrument, not the scoreboard.

• Placed order rate per send. Per send, not per flow. A flow-level number hides the send doing nothing.
• Revenue per recipient per send. Placed order rate rewards a cheap product. Revenue per recipient catches the send producing fewer, larger orders, usually send 4.
• Contribution per subscriber. Placed order rate multiplied by contribution margin, so a discount can't dress up a loss.
• A holdout, for what the flow actually caused. Attributed revenue only tells you what the flow got credit for. Klaviyo's global holdout groups withhold a set percentage of profiles from campaigns and flows across all channels to measure lift, and you can override the holdout for specific flows so welcome messages keep sending (Klaviyo Help Center). Read that override carefully: exempt the welcome flow and the holdout stops measuring the welcome flow.
• Deliverability, read separately. Complaint rate, unsubscribe rate, bounce rate. A series that converts while poisoning the domain is borrowing from your campaigns.

The checklist, before you call a series measured:

• [ ] Every send has one named metric, written down before it ships.
• [ ] Placed order rate is read per send, not only at flow level.
• [ ] Revenue per recipient is read alongside it.
• [ ] The attribution window is written down and identical across every comparison.
• [ ] A holdout exists and you know which flows it covers.
• [ ] Buyers exit the sequence, verified in actual sends and not in the flow diagram.
• [ ] Complaint, unsubscribe and bounce rates sit on the same report as the revenue.
• [ ] The cohort is judged on lifetime gross profit to CAC, and someone can produce that number on request.

A worked example

Take a brand with average order value A and gross margin rate g. Contribution margin per order is (A x g) minus variable order costs. The series converts subscribers at placed order rate p with no discount, so contribution per subscriber is p x CM. Add a fifteen percent code and contribution per order becomes CM minus (0.15 x A). Breakeven lift is (CM / (CM minus 0.15A)) minus 1. Put your real A, g and cost lines in and you get a number. That number is the bar the discounted series has to clear on placed order rate, and on most catalogs it is far higher than people expect.

Then check the half that decides it. Did the discounted cohort come back? Multiply their repeat rate by their repeat contribution margin, add the first order, divide by subscriber acquisition cost. That's lifetime gross profit to CAC for the cohort, and it settles the argument.

What the discipline is worth on real accounts: one retention build on our email site shows "0% → 59% of Revenue from Retention" alongside "+140% Total Revenue YoY" and a "54% Returning Customer Rate" (email.hayesmedia.co). Another shows "+218% Email Revenue YoY" and "39.56% of Revenue from Email & SMS" next to a "59.4% Open Rate" (email.hayesmedia.co). Note which of those three we would never optimize toward. Your figures will differ, because they depend on your margin, catalog and traffic source. The mechanism does not.

Mistakes we see

• Opening with a discount nobody promised. You've trained the list to wait, and every campaign after this competes with the code you gave away.
• Sending the pitch to people who already bought. Usually a missing profile filter. It's the fastest route to a complaint from a brand new customer.
• One send carrying four ideas. Founder story, proof, objection handling and a product grid in one email is a landing page with a subject line.
• Ending in a countdown timer. A fake deadline on a first purchase teaches the buyer that your prices are theatre.
• Judging the flow on open rate. Opens are inflated by design.
• No holdout, ever. Attributed revenue without a control group is a number the tool made up on your behalf.
• Letting the series trail off. It should end. Non buyers roll into your campaigns and your abandoned cart and browse flows.
• Building five splits before the first has data. Complexity you cannot maintain degrades faster than a simple flow you actually read.

The same handful of errors shows up account after account, which is the subject of this breakdown.

When this does not fit

• Signup and purchase are the same action. Subscription-first and membership brands have no gap to fill. The work moves to onboarding and the first renewal.
• High consideration, high ticket. If the decision involves a spouse, a contractor or a finance approval, a five-send sprint to an order is the wrong shape. The series becomes education and the metric becomes consultations booked.
• B2B and wholesale lists. The order is a purchase order and the sequence should route to a human.
• Very low signup volume. If you add a handful of subscribers a week you cannot read per-send differences, let alone run a holdout. Build the five sends, judge them qualitatively, and put the effort into list growth.
• A list you did not earn. Purchased or scraped addresses fail on deliverability regardless of what the emails say. Fix acquisition first.
• Regulated categories. Supplements, alcohol, anything with claim restrictions. The proof send is where brands get into trouble, and legal review beats a clever subject line.