Skip to content

Ecommerce email marketing: build the flows first

What is ecommerce email marketing made of?

Two jobs, not one.

Ecommerce email marketing is the email a store sends to people who gave it their address. It has two halves. Flows send themselves off a trigger, like a started checkout. Campaigns you write and schedule. Klaviyo ships its abandoned cart flow with a four hour delay built in, so flows work in hours.

Most owners treat email as one pile of work. It is two piles, and they fail in different ways.

A flow is built once and keeps running. A shopper acts, a timer starts, the email goes. The work sits in the setup and the wiring. When a flow breaks it breaks quietly, and nobody notices for a month.

A campaign is written this week and sent this week. The work sits in the writing. When a campaign is skipped you know the same day, because the calendar has a hole in it.

Both draw on the same list. Both spend the same sender reputation. So a store sending four campaigns a week with no flows is doing the hard half and skipping the easy half.

Part What starts it Where the work sits How it fails
Flow A shopper’s action Setup and wiring, once Silently. It stops sending and nobody looks
Campaign Your calendar Writing, every week Loudly. You see the gap on Monday
List and consent The sign-up form Once, then upkeep Slowly. Reputation drops before sales do
Segments Your rules Once, then upkeep Quietly. Right email, wrong people

Build the flows first. They are the half that keeps working after you stop paying attention to it.

Which email flows should a Shopify brand build first?

In this order, and not all at once.

Build in this order: welcome series, abandoned checkout, browse abandonment, post-purchase, winback. The first two cover the moments a shopper sits closest to buying, a fresh sign-up and a started checkout. Klaviyo fires the second on Started Checkout, after the shopper enters contact details and moves to shipping.

1. Welcome series. It answers the person who just gave you their address and is still on the site. It is also the only flow where you control the trigger completely, because your own form starts it. Start with a welcome email series built for a Shopify store, not the four generic emails every ESP ships in its template library.

2. Abandoned checkout. The shopper picked the product, entered an email, and stopped. Nothing else in your account speaks to a warmer moment. This is the flow with the most written-down craft, so work from abandoned cart emails with examples, a template and the Shopify setup.

3. Browse abandonment. Viewed a product, added nothing. Weaker intent, so a lighter touch. One email, sometimes two.

4. Post-purchase. Shipping, care, how to use it, and the ask for a second order. It carries the least risk, because the person already bought.

5. Winback. Bought once, went quiet. Build it last, because you need enough order history for the timing to mean anything.

Order Flow Trigger Why here
1 Welcome series Subscribed to a list You own the trigger, and intent is fresh
2 Abandoned checkout Started Checkout Highest intent in the account
3 Browse abandonment Viewed a product Real interest, no commitment yet
4 Post-purchase Placed an order Lowest risk, sets up the second order
5 Winback No order in X days Needs order history to time it

Five flows is a system. Fifteen is a maintenance problem you built for yourself.

What belongs in the weekly campaign calendar?

One send that earns its place. Not three that fill space.

A campaign calendar should hold one send a week you could defend out loud: a new product, a restock, a use case, a customer answer, an honest offer. Fill the rest with segments, not extra sends. Google tells bulk senders to keep spam complaints under 0.10 percent, and volume is how you break that.

Most stores do not have a content problem. They have a reason problem. If you cannot say in one line why this email exists this week, the email does not exist this week.

Five reasons that hold up every time:

  • Something is new or back.
  • Something is going away.
  • A customer asked a question other customers also have.
  • You know something about the category the buyer does not.
  • You have an honest offer with a real end date.

Then cut who receives it. A restock email belongs to the people who viewed that product or bought its predecessor, not to everyone. Segmenting is how you send less and keep more of the list.

Write the campaign to one person. Read it back as if it arrived on your own phone in a queue of forty other emails. If you would not open it, do not send it.

How do you stop two flows emailing the same person twice?

With filters, before you turn anything on.

Use flow filters. Klaviyo recommends two on an abandoned cart flow: Placed Order zero times since starting this flow, and has not been in this flow in the last X days. Add them at build time. Overlap is not a rare edge case, it is the normal result of running five flows on one list.

The overlap that bites hardest: a new subscriber signs up, gets welcome email one, adds to cart the same hour, starts a checkout, and now sits inside two flows that both want to send tonight. They get two emails from a store they met an hour ago.

Three fixes, in order of how much they matter:

  1. Exit conditions. A flow should stop when its job is done. Placed an order? Out of the abandoned checkout flow.
  2. Re-entry limits. Nobody should re-enter the same flow this week because they browsed twice.
  3. Send-time protection. Most platforms have a setting that stops one profile receiving repeat flow messages inside a short window. Turn it on before you need it.

The wiring is platform work, and it is the part most guides skip. Which Klaviyo flows to build, and how the filters stop them colliding covers the exact triggers and filters per flow.

Not sure which half of this is costing you the most? The assessment reads your week and names the one marketing job worth building first. Take the assessment. Free. No pitch. You get the answer on screen.

What has to be true before any of it sends?

Consent, authentication, and a way out of every email.

Set up SPF, DKIM and DMARC on your sending domain. Google treats more than 5,000 messages a day to Gmail as bulk sending and requires one-click unsubscribe on marketing mail. The FTC’s CAN-SPAM guide requires you to honour an opt-out within 10 business days. None of this is optional.

This is the unglamorous half, and it decides whether the other half arrives.

  • Authenticate the domain. Google’s email sender guidelines and Yahoo’s sender best practices both require SPF, DKIM and DMARC, with the From domain aligned.
  • Watch the complaint rate. Google asks bulk senders to stay under 0.30 percent reported spam in Postmaster Tools and recommends staying under 0.10 percent.
  • Make leaving easy. One-click unsubscribe, plus a visible link in the body. The FTC’s CAN-SPAM compliance guide sets the 10 business day limit for processing it.
  • Keep consent honest. A checkout is not a subscription. If someone did not opt in to marketing, they belong in transactional email only.

A brand that hides the unsubscribe link buys three months of extra sends and pays with the inbox placement of every email after that.

How do you judge whether the email programme is working?

Read revenue per recipient, not open rate.

Apple’s Mail Privacy Protection pre-fetches images, so open rate now reads high for reasons that have nothing to do with the reader. Judge a send on three numbers instead: clicks, orders inside a fixed attribution window, and revenue per recipient, which exposes the send that only worked because it went to everyone.

Four numbers, in this order:

  1. Revenue per recipient. Total attributed revenue divided by people who got it. It is the only number that punishes sending to the whole list.
  2. Click rate. Did the message do its job and move them back to the site.
  3. Unsubscribe and complaint rate per send. The cost of the send, paid later.
  4. Flow completion. For a flow, how many people reach the last email versus how many leave because they bought. High exit through the purchase is a working flow.

Set the attribution window once and keep it. Changing it mid-quarter makes every comparison meaningless, and it is the most common way a store convinces itself something worked.

Judge a flow after it has had enough triggers to mean anything, not after four days. A welcome flow on a store with thirty sign-ups a week needs a month before its numbers are worth reading.

How do you build this without it eating every week?

Build it once as a job, then rerun the job.

Treat each flow as one job with a written brief: trigger, filters, timing, the message of each email, and who approves it. Five briefs cover the whole programme. Clockwork runs its own marketing this way, and the rebuild cost is the reason most stores stop at two flows and stall there.

The five steps, in the order that works:

  1. Write the brand context down once. Products, buyer, promise, tone, the claims you are allowed to make, the words you never use. Every email you write for the next year starts from this file. Do it once and no run starts from zero.
  2. Pick one flow. The welcome series if you have sign-ups. The abandoned checkout flow if you have traffic and a live cart.
  3. Draft it against the brief. Every email gets one job and one link.
  4. Review before it goes live. Read the drafts as a sequence, not one by one, so you catch the same sentence appearing twice.
  5. Turn it on, then leave it alone. Come back after a month of real triggers.

This is the shape of work Clockwork keeps. The weekly marketing jobs a store already does, run from your own context and evidence, with anything questionable held and shown with the reason. Nothing goes live without your yes, and how a run works spells out where the held row and the record sit. Clockwork’s own marketing runs on that shape of work, which is why the record and the held row exist at all.

Where should you start this week?

With one flow, written down.

Pick one of the five flows, write the brief, draft the emails against your brand context, review them as a sequence, and turn it on. One finished flow beats five half-built ones, because a half-built flow still sends. The order Clockwork uses on its own marketing is welcome series first, then abandoned checkout.

The stores that end up with good email are not the ones with better ideas. They are the ones who wrote the job down once, so the next run started from something instead of from a blank page.

FAQ: what else do Shopify owners ask about ecommerce email marketing?

The four questions that come up before anything gets built.

The common questions are how many emails to send, whether to use Shopify Email or Klaviyo, whether discounts are required, and how long results take. The honest answers are: fewer than you think, whichever one you will actually maintain, no, and longer than a week of data can tell you.

How often should a Shopify store send ecommerce email marketing?

One campaign a week is a defensible floor, plus whatever your flows send to individuals. Flows do not count against the calendar, because each person only receives their own. Add a second weekly campaign when your complaint and unsubscribe rates stay flat after a month, not before.

Do I need Klaviyo, or is Shopify Email enough?

Shopify Email handles campaigns and basic abandoned checkout automations from inside the admin. Klaviyo adds branching flows, event triggers such as Started Checkout and Added to Cart, and segment rules. Start with the tool you will maintain. A working flow in the simpler tool beats a half-built one in the better tool.

Does every flow need a discount?

No, and defaulting to one teaches your best customers to wait. Use the message first: the reason to buy, the objection, the proof, the guarantee. Keep a code for the last email in the sequence, if at all. A store that discounts in every flow has trained its list to ignore full price.

How long before email is worth judging?

Long enough that the numbers come from behaviour and not from noise. A flow needs enough triggers to compare against something, which on a small store means weeks, not days. Set the attribution window once, write down when you turned the flow on, and read it at the end of the month.

Keep reading: which job do you build first?

Three pages, in build order.

Start with the welcome series, because your own sign-up form controls its trigger. Build the abandoned checkout sequence second, because it catches the highest intent moment in the account. Read the Klaviyo page before either one, because triggers and filters decide whether the two collide.

The jobs in this system

Everything we have written

Not sure which job to build first?

Eight questions, about two minutes. It names the one recurring shop job worth handing off first, and counts what it costs you across a year.

Take the assessment