Marketing Automation Strategy: 14 Flows Ranked by Return Per Hour

The strategies, not the six steps

Search this topic and you get the same article repeatedly: define your goals, map the customer journey, audit your stack, design your workflows, launch, measure. Six steps, no argument with any of them, and nothing you can build on Monday morning.

What is missing is the list. Which flows are worth building, in what order, what each one is triggered by, and roughly what it returns. That is what this page is.

Fourteen strategies below, ordered by return per hour of build time. The comparison table tells you what to build first. Build three, get them right, and then come back for the rest, because a programme with three well-maintained flows beats one with fifteen half-built ones every time.

Methodology: ordering is by build effort against typical impact, from running these programmes rather than from a survey. Where a widely recommended flow rarely earns its maintenance cost, it is in the "not worth it for most" section rather than padded into the list.

The comparison table

Rows 1, 2 and 6 are where almost every programme should start. They are cheap to build, cheap to maintain, and they cover the three moments where automation reliably beats a human: immediately after signup, immediately after abandonment, and long after a person would have given up following up.

1. The welcome sequence

Three to five emails over two weeks, starting within minutes of signup. This is the highest-engagement window you will ever have with a contact and most businesses spend it on a single "thanks for subscribing".

What to put in it: what you do, the one thing they most likely came for, proof it works, and a single clear next step. One idea per email, not a newsletter.

The detail that matters: send the first one immediately, not at the next scheduled batch. Engagement drops sharply within hours of signup and a delay of a day costs a meaningful share of opens.

2. Abandoned cart recovery

Someone adds to cart and does not check out. Three messages: one hour, 24 hours, 72 hours.

The first should assume a technical problem or a distraction, not reluctance. The second can address a common objection like shipping cost or returns. The third is where a discount belongs, if you use one at all.

Where most implementations lose money: discounting in the first message. You train customers to abandon carts deliberately, and it takes months to untrain.

3. Lead scoring and routing

Assign points to actions, and when a contact crosses a threshold, route them to sales automatically.

The scoring model is where this succeeds or fails. Score behaviour that indicates buying intent, such as pricing page visits, demo requests and repeat visits within a short window. Do not score newsletter opens, which mostly identifies people who read email.

Review the model quarterly. Scoring built once and never revisited is the most common reason sales teams stop trusting the leads marketing sends, and once that trust goes the whole programme is decorative.

4. The post-purchase sequence

The window everyone wastes. A buyer has just decided you were worth paying, and most businesses follow it with a receipt and silence.

Use it for delivery expectations, getting the product used successfully, and the review request. Successful first use is the strongest predictor of a second purchase, and it is almost entirely automatable.

5. Behavioural nurture

Instead of sending everyone the same drip, branch on what a contact actually looked at. Someone who read three pricing comparisons needs a different next message than someone who read an introductory guide.

This is where automation earns its licence fee, and it is also where complexity gets out of hand. Two or three branches, not twelve. Every branch you add is a branch somebody has to maintain.

6. Win-back campaigns

Contacts who bought or engaged and then stopped. Two or three messages triggered by an inactivity window appropriate to your cycle, typically 60 to 180 days.

The most effective version simply asks what changed rather than offering a discount. It gets replies, and replies tell you why people leave, which is worth more than the recovered orders.

7. Trial and onboarding flows

For any product with a free trial, the flow should be triggered by what the user has and has not done, not by the day of the trial.

A user who has not completed setup on day two needs a setup email. A user who completed setup and is using the product needs an advanced tip, not the same setup email. Time-based trial sequences send the wrong message to half the list on every send.

8. Browse abandonment

Viewed a product, never added to cart. Lower intent than cart abandonment, so one message, not three, and a lighter touch.

Cheap to build and it works. The failure mode is being creepy: reference the category rather than replaying the exact item someone looked at once.

9. Sales handoff alerts

Not an email flow at all. When a contact crosses the score threshold, notify the rep with the contact's actual behaviour attached.

This is a two-hour build with an outsized return, because the difference between a rep calling within an hour and calling the next day is large and well documented. It is also the flow most often skipped, because it does not look like marketing.

10. Replenishment reminders

For anything consumable, time the reminder to typical reorder cycle minus a week. Almost no maintenance, and it converts because it is genuinely useful.

11. Review and referral requests

Triggered by delivery or by a success milestone, not by the calendar. Ask once, ask well, and make it one click.

Timing decides everything here. A review request that arrives before the product does is worse than none at all, and that misfire is common when the trigger is order date rather than delivery date.

12. Churn-risk intervention

Usage drops below a threshold, and something happens: an email, an in-app message, or a task for a human.

The highest-value flow on this list and the most expensive to build, because it needs product usage data flowing into your marketing platform. Only worth it for subscription businesses, and only after the simpler flows are running.

13. Re-engagement before sunset

Before you remove inactive contacts from your list, one last honest message. Two benefits: some of them come back, and removing the rest improves deliverability for everyone remaining.

Sunsetting feels like losing list size. It usually improves revenue per send, because inactive contacts drag your sender reputation down and suppress delivery to people who would have bought.

14. Event and webinar sequences

Registration confirmation, reminders at 24 hours and one hour, then two different follow-ups for attendees and no-shows.

The split is the whole point. Sending attendees the "sorry you missed it" email is the single most common automation error and it is entirely avoidable.

Not worth it for most businesses

• Twelve-branch decision trees. Impressive in a demo, unmaintainable in practice, and no one on the team will understand them in six months.
• SMS for everything. Effective for delivery and appointments, intrusive for nurture, and expensive per message.
• Predictive send-time optimisation. Real but marginal, and usually the last thing worth tuning.
• Full personalisation of every field. The gain over first-name plus segment is small, and the failure mode, a broken merge tag, is visible to the customer.
• Automating the sales conversation itself. Automate the routing and the follow-up reminder. The conversation stays human.

How to sequence a build

Month 1. Welcome sequence, and one of abandoned cart or win-back depending on your model. Clean the list before either goes live.

Month 2. Post-purchase or trial onboarding, plus sales handoff alerts if you have a sales team.

Month 3. Lead scoring, with a scoring model you will review in month six.

Month 4 onward. Behavioural branching, then the specialist flows for your model.

Resist building everything at once. The failure pattern is a large programme launched in a single quarter, unmaintained by the second, and quietly turned off in the third.

Measuring each flow

Per flow, not per programme. Aggregate numbers hide the flow that is doing all the work and the four that are doing nothing.

That last row catches the real problem with automation, which is that flows keep running long after they stopped being true. Prices change, products get discontinued, and the welcome sequence keeps promising something you no longer sell.

For the money side of these decisions, our guide to marketing automation ROI covers the calculation properly, and lead generation cost covers what you are comparing against.

Choosing the platform

Briefly, because the platform matters far less than the flows.

Pick on integration with your existing data, not feature lists. The best platform is the one that can see what your customers actually do, because behavioural triggers are where the value is and they require data. A cheaper tool with a native integration to your commerce or product data beats an expensive one that needs a custom pipeline.

We compare the category in best AI marketing automation tools, and digital marketing strategy examples covers where automation fits in the wider plan.

Frequently asked questions

What is a marketing automation strategy? The set of triggered flows you run, the data that fires them, and the rules for maintaining them. In practice it is a list of specific flows like the fourteen above, not a philosophy. Where AI changes those flows rather than just writing the copy inside them is covered in our guide to AI marketing strategies.

Which automation should I build first? A welcome sequence, in almost every case. It is cheap, it touches every new contact, and it is the highest-engagement window you get.

How many flows should a small team run? Three to five, maintained. Teams under three people that run more than that generally have flows nobody has read in a year.

How often should flows be reviewed? Quarterly for scoring models, annually for everything else, and immediately whenever a price, product or offer changes.

Do I need a big platform to start? No. Most of the first six flows here run on entry-tier tools. Upgrade when a specific trigger you need requires data your current tool cannot see.

What is the difference between marketing automation and email marketing? Email marketing sends to a list on a schedule. Marketing automation sends to an individual because of something they did. The trigger is the whole difference. For the small business end of that distinction specifically, our roundup of email marketing automation for small business compares the platforms that do both.

Start with three

Pick the welcome sequence, one abandonment or win-back flow, and the sales handoff alert if you have a sales team. That is roughly two days of build for the flows that carry most of the return in the table above.

Get those three right, leave them alone for a quarter, and measure each against a holdout. Then build the next three. The programmes that work are the ones built at this pace, and the ones that get switched off are the ones built all at once.

If the constraint is the content those flows need rather than the flows themselves, Distribb writes and publishes it on a schedule, which is the input most automation programmes run short of by month three.