Every article about Klaviyo flows gives you the same list. Very few tell you what goes inside each one, and none of them tell you the order.
The order is the part that matters. Build these out of sequence and the early work does not fund the patience for the rest, which is how flow projects stall at two flows and stay there for a year.
Here is the order, and what each one should actually contain.
1. Abandoned checkout
Three emails. 1 hour, 22 hours, 46 hours.
Someone entered checkout and left. It is the highest intent anyone will show you without buying, and it is the flow most likely to already exist as a single email.
- Email 1, at 1 hour. Not a discount. Just the cart, and one line asking if something
got in the way. A meaningful share of abandoned checkouts are a card that failed or a shipping cost that surprised someone, and both are answerable without margin.
- Email 2, at 22 hours. Handle the objection instead of repeating the reminder. What
the return policy is, how fast it ships, what the guarantee is. Reviews of the specific product in the cart do more work here than reviews of the brand.
- Email 3, at 46 hours. Now you may discount, if you discount at all. Make it expire.
The mistake to avoid: leading with a discount at hour one. You are paying margin to people who were going to buy anyway, and you are training your buyers to abandon.
2. Welcome
Four to five emails over ten days.
Someone gave you their email. They will never be more curious about you than they are in the next 48 hours, and most brands answer with one email carrying a 10% code and nothing else.
- Email 1, immediately. Deliver whatever you promised. If it was a code, the code is
the first thing they see, not the last.
- Email 2, day 2. Why the brand exists, told as something specific rather than a
mission statement. What you refuse to do is usually more interesting than what you promise.
- Email 3, day 4. The bestseller, or the product that most people should start with.
Not the catalogue.
- Email 4, day 7. Proof. Reviews, press, the number of people who have bought.
- Email 5, day 10. A reason to act now if they have not, and a segment exit if they
have.
The test that matters: if your welcome sequence would read identically with another brand's logo on it, it is not a welcome sequence. It is a coupon with steps.
3. Browse abandonment
Two emails. 4 hours, 30 hours.
Someone looked at a product and left without adding it. Lower intent than a cart, far more traffic, and cheap to build once the abandoned checkout emails exist because the structure is the same.
Keep it soft. This person has not committed to anything, and an aggressive browse flow reads as surveillance. Show the product, show one alternative, stop.
Suppress anyone currently in the abandoned checkout flow or you will send two near-identical emails within an hour of each other.
4. Post-purchase
Three to four emails, starting the day after delivery.
The window where a first-time buyer becomes a second-time buyer, and the flow almost every store leaves empty because the sale already happened.
- Email 1, day after delivery. How to use the thing well. No selling. This is the
email that reduces returns and it earns the right to the next three.
- Email 2, day 7. Ask for the review, once.
- Email 3, day 14 to 30. The natural next product. Timing depends on your replenishment
cycle rather than the calendar — a consumable is 30 days, a pair of shoes is not.
- Email 4, optional. Referral or the loyalty programme, if you have one worth mentioning.
Split it by what they bought. A post-purchase flow that treats every product the same is the clearest signal a store has that nobody is watching it.
5. Winback
Two to three emails, triggered at roughly twice your average repurchase interval.
Only build this once the four above are running, because it works on exactly the people those four failed to keep. Build it first and you are papering over the leak instead of fixing it.
Work out your own trigger rather than using 90 days because an article said so. Take the median gap between first and second order, double it, start there.
- Email 1. What is new since they last bought. People leave because they stopped
paying attention, not because they decided against you.
- Email 2. The offer, if you use one.
- Email 3. A genuine last email, and then let them go quiet. Continuing to mail people
who have ignored five sequences is how a sending reputation dies.
Why this order
It runs from highest intent to lowest. Abandoned checkout pays back within days, which funds the patience for post-purchase, which pays back over months. Start at winback and nothing pays back for a quarter, which is roughly when the project gets abandoned.
How long does it take to build all five?
Properly, this is around 20 to 25 emails, each one written in the brand's voice with the right products in it, plus the segmentation, the suppression logic, and the testing.
For one person doing it alongside everything else, that is a two to three month project that usually takes six, because it is the only work on the list with no deadline attached.
That is the whole reason it does not get done. Not difficulty. No deadline.
What to do after they are live
Leave them alone for six weeks, then look at flow revenue per recipient rather than open rate. The emails to change are the ones with traffic and no revenue, not the ones with low opens — opens have been unreliable since Apple started pre-loading images.
Then change one thing at a time. A flow you rewrite wholesale teaches you nothing about which change worked.
If you want the five built rather than described, that is what we do, and the machine that builds them is yours to keep at the end. Fifteen minutes is enough to know whether it fits.