Best email marketing automation software: the branch is the product, and the tier line is almost always the branch

Every email platform sells automation, and the thing that separates them is whether a sequence can take a decision. A linear sequence sends three emails on a timer. A branched one asks a question at each step, which contact opened, who bought, who looked and left, and sends a different message to each group. Branching is where the money in email automation is, and at several vendors it sits one tier above the plan most buyers pick. Check that one line and the comparison is nearly finished.

The five flows worth building, in order

Flow	Trigger	The branch that matters
Welcome	A new subscriber	Opened the first email or did not
Abandoned basket or enquiry	Started and did not finish	Which product or which service page
Post purchase	A completed order	First order against repeat order
Lapsed	No open or purchase in a set window	Re engaged, or move to suppressed
Review request	Delivery or job completion plus a delay	Satisfied, or route to support instead

Those five cover most of the value in email automation for most businesses. Build them in that order, because each one earns from traffic you already have rather than from traffic you have to buy.

Where each vendor puts the tier line, read on 21 October 2026

• ActiveCampaign publishes send caps and seat counts as its tier lines, and at a thousand contacts its three visible plans publish 10,000, 12,000 and 15,000 email sends a month alongside one, three and five user seats. If several people will build and approve flows, the seat count will decide your plan before any feature does.
• Brevo publishes marketing automation and A/B testing on its second tier rather than its first, and states that the automation on it allows an unlimited number of multi step workflows. Its tiers are set by monthly email volume, from 5,000 a month on the first paid tier to from 150,000 on the third, with ten seats and phone support arriving on the third. It also publishes send time optimisation on the second tier and contact scoring on the third.
• Klaviyo publishes automations on its free plan, bounded by up to 250 active profiles and 500 email sends a month, which makes it the cheapest way to build and watch a branched flow before paying anything. The paid ladder then runs on active profiles.
• MailerLite publishes the full feature set on every paid plan, with price moving only on list size and sending, plus unlimited monthly emails and unlimited user seats. For automation specifically, that is the clearest published promise in this group, because nothing is behind a higher tier.
• Mailchimp bands on contacts and publishes that overages apply if the contact or email send limit is exceeded, so model the overage before relying on an entry plan for high frequency flows.

On the stated criterion, MailerLite wins outright for a buyer who does not want automation gated by tier. Klaviyo wins for building a branched flow at no cost first. Brevo wins for a large list mailed rarely, since volume rather than contacts sets its price. ActiveCampaign is the one whose seat count is most likely to catch a team out, which is a planning fact rather than a criticism.

The four things that break a flow after it is built

• No single source of truth for the subscribed flag. Two systems holding consent means somebody gets mailed after they left.
• No exit condition. A buyer who purchases in the middle of a nurture sequence must stop receiving it in the next step, not at the end.
• No suppression of bounces and complaints. Automated sending to dead addresses damages placement faster than any campaign.
• No reporting per contact. If you cannot export per contact data, you cannot tell which branch earned anything.

Mailbox provider requirements apply to automated mail exactly as they do to campaigns, which means authenticating your sending domain before the first flow goes live and keeping one click unsubscribe available on marketing messages. The placement, list health and flow detail is in email marketing deliverability, list health and the flows that earn, and if you are choosing the platform itself rather than the automation layer, start at email marketing platforms for a small business.

What buyers ask about automation software

How many flows should a first setup have?

Three. Welcome, one recovery flow for the thing people start and abandon, and one post purchase. Five is the ceiling for a first year, because each flow needs reading and editing to stay true.

Is a CRM with email built in enough?

For a sales led business where one record drives both email and pipeline, often yes, and the data proximity is a real advantage. Move to a dedicated sender when you need behavioural segmentation across channels or a content newsletter with its own calendar.

Does automation need more contacts to be worth it?

No. Automated flows send on the contact's timing rather than yours, so they earn at small list sizes where a monthly campaign does not. A welcome sequence to two hundred subscribers is usually the highest returning email a small business ever builds.

What should I test first?

Build one flow with one branch during the trial, authenticate the domain, send it to a real segment and export the per contact report. That single exercise reveals almost every limit a plan has.

The one trigger no automation platform can see

Automation fires on things your systems observe: a form, a click, an order, a date. There is one buying signal none of them can see, because it happens outside your property entirely. A buyer asks ChatGPT, Gemini, Perplexity, Claude or Google AI Overviews who to use for what you sell, and gets back two or three business names. No pixel records it, no flow triggers on it, and nothing in your reporting contains it.

We do not sell email automation and this page does not suggest buying us instead of one of the products above. The adjacent category, measuring whether assistants name a business at that moment, is the one we are in, and on the criterion used here, what the vendor publishes about its own limits, AI Knows Us publishes the whole unit openly: five surfaces named, one credit as one buyer question asked on one app once with Claude published at three, the questions watched and the answers read a month stated per plan, one login with full access, and United States prices rendered on the page. Where a rival publishes a stronger allowance, plainly: Semrush publishes unlimited prompts to track daily on its top tier, read on semrush.com/pricing on 21 October 2026, which is a wider allowance than any question cap we publish.

Build the five flows first, because they earn from the traffic you already have. Then ask your buyer's own question in one assistant, because a flow can only work on somebody who arrived, and that answer decides who arrives.

And whichever platform the flows end up on, the setup that decides whether they arrive is the same everywhere: authenticate your own sending domain and keep one audience as the source of truth. The worked version of that, on the platform most people start with, is in authenticating the domain and using one audience in Mailchimp.