A software company sets up a seven-email trial onboarding sequence: day one, day two, day four, and so on. One trial user invites her whole team and connects two integrations in the first hour. On day four she still gets “Have you tried inviting your team?”, and on day twelve a “Your trial ends soon” email arrives while she is already on a paid plan, because billing never told the email platform she’d upgraded. Each message went out exactly on schedule. To her, the company looked like it wasn’t paying attention.
We see this a lot. The account usually has enough flows. What it lacks is any link between those flows and what the customer has just done, so nothing knows who the person is right now.
Behavioral email versus the scheduled kind
A scheduled campaign goes out because it’s Thursday. A behavioral email goes out because something happened: a first order, a trial that started but stalled, a renewal date ninety days away, a subscriber who stopped clicking six months ago. The second kind is almost always more relevant, because it arrives when the reader is already thinking about you.
To build that properly, the email platform needs more than templates. It needs reliable data about each person, a clear definition of the state they’re in (new subscriber, first-time buyer, active customer, lapsed), and events it can actually detect. Rules then decide who enters a flow, who is excluded and what happens next. Only after that does the email itself matter.
When one of those layers is missing, the flow still sends. It just sends to the wrong people, or at the wrong time.
How do you decide which email flows you need?
You need the flows that address the points where customers stall in your lifecycle, and that your data can reliably trigger. Everything else is optional until those are working.
We usually work through four questions with a client before building anything:
- Where do people get stuck? Subscribers who never buy, trials that never activate, customers who buy once and disappear. The biggest drop-off is usually the first flow worth fixing.
- Can the system see the moment? If your ecommerce store or product doesn’t pass the relevant event into the email platform, the flow can’t fire. Sometimes the first job is the integration, not the email.
- Who should be left out? Every flow needs exclusions. Recent purchasers, open support tickets, customers already in onboarding. If nobody can define the exclusions, the flow isn’t ready.
- Is there enough volume to matter and to measure? A flow that touches eight people a month may be fine as a manual task.
The answers change a lot by business model. A B2B software company with a few dozen trial signups a month will get far more from a well-built onboarding sequence than from a win-back flow. A coffee subscription brand selling a consumable product lives or dies on replenishment and retention timing. A furniture retailer whose customers buy a sofa every eight years probably doesn’t need a replenishment flow at all, but its post-purchase and referral messages may be undervalued.
The flows we build, and when each one earns its place
These are the flows we’re most often asked about. Treat this as a menu to choose from, not a list to complete.
Notice that reactivation and win-back are different flows. Reactivation is about engagement with your email. Win-back is about buying behavior. Mixing them up is how a loyal customer who simply doesn’t open newsletters ends up receiving a “come back” discount.
Not sure which flows to start with?
We'll look at where your customers stall and what your data can detect, then recommend a short, prioritized flow plan.
What usually goes wrong in existing flows
Many of our engagements start with flows that already exist. The problems tend to repeat:
- Two flows trigger on the same event and nobody decided which one wins.
- The “has not purchased” condition is only checked when someone enters the flow, so a customer who buys mid-flow keeps receiving “complete your purchase” emails.
- Re-enrollment settings were never thought through, so customers either re-enter every time or never again.
- A field the flow depends on was renamed or stopped syncing from the store or CRM, and the flow quietly stopped sending.
- Success is judged on open rates, which Apple Mail Privacy Protection has made unreliable for a large share of recipients.
Platform details matter here. In Klaviyo, the difference between a trigger filter and a flow filter changes who stays in a flow. In HubSpot, enrollment triggers, re-enrollment rules and suppression lists in a workflow decide the same thing in a different way. We work inside the platform you use and explain those choices in its own terms. See our Klaviyo and HubSpot pages for specifics.
How a build runs
Map the lifecycle and the events
We document the customer states that matter to you and which events your systems already send. Gaps in event data get flagged before any flow is designed.
Write the rules down
For each flow: entry trigger, exclusions, exit conditions, timing, and what counts as success. This is the document your team can check the build against later.
Shape the messages
We define what each email needs to do and in what order. We can draft copy, or hand a clear brief to your writer or brand team if you prefer.
Build and test with real scenarios
Test profiles walk through the edge cases: the customer who buys mid-flow, the subscriber with no first name, the contact who unsubscribes from marketing but still needs order updates.
Launch and measure against a baseline
We compare flow recipients against a holdout group or a pre-launch baseline where volume allows, and look at movement to the next customer state rather than opens alone.
Measuring whether a flow works
A flow works if it changes what customers do next. For a welcome flow that might be first purchase rate within a set window. For onboarding, the share of new accounts that reach an activation milestone. For win-back, the share of lapsed customers who order again.
Attributed revenue numbers inside email platforms are useful but generous, because they credit any conversion within the attribution window. Where the list is large enough, a small holdout group gives a far more honest answer. Where it isn’t, we’ll say so and agree a simpler comparison.
Behavioral email also depends on reaching the inbox. A well-built flow sending from a domain with poor authentication or high complaint rates will underperform no matter how good the logic is. If that’s a concern, our email deliverability review looks at it directly.
How many flows should we start with?
Usually two or three. Pick the stages with the biggest drop-off that your data can reliably trigger, get them working and measured, then add more. Launching a dozen at once makes it very hard to tell which ones help.
What if our store or product doesn’t send the events a flow needs?
Then the integration comes before the email. A missing upgrade, refund or usage event can usually be added through the platform’s native connection, a workflow tool such as Zapier, Make or n8n, or a direct API call. We’d rather spend the first week of a project on that than build a flow that can’t see its own trigger. Our integrations page explains how we choose.
Do you write the emails?
We can. Some clients want the full flow including copy, others have a writer or brand team and want us to handle structure, logic and build. We’ll agree which before the project starts.
Can you fix our existing flows instead of rebuilding?
Often, yes. If the triggers and data are sound, tightening exclusions and timing is faster than starting over. If the underlying data is unreliable, we’ll tell you that patching the flows won’t hold.
Is email automation the same as lifecycle automation?
They overlap. Email marketing automation focuses on the messages. Lifecycle automation also covers the CRM side: lifecycle stage changes, sales tasks, routing and the rules that move customers between states.
How is a project priced?
It depends on the number of flows, the platform, and how much data or integration work is needed first. For reference, our lifecycle automation engagements start from $2,500. Our pricing page has current starting points.