Marketing automation

What sits in this category

Platforms that coordinate marketing across channels around one contact record: capture, segmentation, sequences, scoring, landing pages and the handover to sales.

The distinction from ordinary email tools is orchestration. A campaign tool sends to a list. An automation platform reacts to behaviour, updates a record and triggers something in another system.

Plain newsletters and sequences belong in email marketing. Outbound prospecting is sales engagement, and the deal record itself lives in CRM and sales.

Readiness matters more than features

This is the category most often bought too early.

The platform assumes several things exist: a steady flow of inbound interest, content worth sending, a definition of a qualified lead that sales agrees with, and somebody whose job is to run it. Where those are missing, the subscription becomes an expensive contact database.

A useful test is whether you can name the three sequences you would build in the first month, and who would write them. If that answer is vague, a simpler tool will produce more revenue for less money.

The pieces, and which ones earn their keep

• Forms and capture, on your own pages and in embedded places.
• Segmentation, on behaviour rather than job title alone.
• Sequences triggered by actions, not only by list membership.
• Scoring, to decide who deserves a human conversation.
• Landing pages, useful when the website team is a queue.
• Reporting, tying activity to pipeline rather than to opens.

The first three do most of the work. Scoring becomes valuable once volume exceeds what a person can triage, and before that it mostly formalises guesses.

Landing page builders inside these platforms are convenient and often slow, and they live on a different domain unless configured otherwise. Both facts matter for search performance and for consistency.

The handover to sales

Marketing automation succeeds or fails at the point where a person is passed to a salesperson.

Agree the definition before configuring anything: what a qualified lead is, what happens when sales rejects one, and how quickly contact must be attempted. Products support all of this, and none of them supply the agreement.

Then check the synchronisation with your CRM. Two-way, field-level, with clear rules about which system wins on conflict. One-way pushes create two versions of the truth, and the argument about which is correct outlives most marketing plans.

Measurement, honestly

Attribution in this category promises more than it can deliver.

First-touch and last-touch models are simple and wrong in opposite directions. Multi-touch models spread credit using assumptions you cannot verify. All of them are useful for comparing periods and campaigns, and none should be quoted as fact in a board meeting.

Measure the things that are actually observable: pipeline created, conversion rate between defined stages, time from first contact to opportunity. Those survive scrutiny.

What it costs, including the parts that are not software

Pricing combines contacts, sends and seats, with automation depth setting the tier. Inactive contacts usually still count, so cleaning the list reduces the bill as well as improving deliverability.

Onboarding fees are common at the upper end, and so is a minimum annual commitment. The larger hidden cost is staffing: a platform of this kind needs someone to build and maintain sequences, and the subscription is often the smaller half of the total. Related pricing habits are described in what pricing pages hide.

Building the first three sequences

A platform bought without a plan becomes a contact database with a monthly fee, so it helps to know what will exist in the first month.

The first sequence is a welcome, sent to anyone who gives you an address, explaining what they will receive and giving them something worth opening.

The second is a behavioural trigger tied to the action that matters most in your business: a trial started, a demonstration requested, a document downloaded.

The third is recovery, aimed at people who began something and stopped.

Those three cover most of the return this category produces. Scoring, complex branching and attribution modelling can wait until the basics are running and somebody owns them, and building them first is the reliable way to end up with a configured platform nobody uses.