Email marketing.

The list is the channel's asset, and it is built in only one way: someone gives their address freely and knows what it will be used for. Six things decide whether the foundation holds.

1. 01Consent has to be active. Swedish marketing law (section 19 of marknadsföringslagen↗) requires prior consent for electronic advertising to individuals, and the GDPR's requirements for consent↗ mean it must be freely given, specific, informed and as easy to withdraw as it was to give. A pre-ticked box meets none of that. The exception sits in the same section: if you obtained the address in connection with a sale, you may market your own similar products without asking first, as long as the person has not objected and an opt-out is offered free of charge both when the address is collected and in every send.
2. 02Say what they are opting into. Anyone signing up should know roughly what will arrive and how often. A promise of news and offers with no substance means the unsubscribe arrives with the first send.
3. 03Never buy a list. A bought address said yes to someone else, not to you, and carries neither legal basis nor interest. The effect shows up immediately in complaints, unsubscribes and in how future sends are treated.
4. 04Give something worth an address. The guide, the template or the analysis is what makes the trade reasonable for the recipient. Without that effort the form is a question with nothing offered in return, and the people who fill it in anyway are rarely the ones you wanted.
5. 05Keep the list clean. Addresses stop working, people change jobs and interest fades. A list that is never pruned looks bigger and performs worse, because inactive recipients both drag the numbers down and make them harder to read.
6. 06Make unsubscribing easy. A link that works in one click is both a requirement and a protection. Someone who cannot find the way out reports the send as spam instead, and that costs more than a lost address.

The through-line: the list's value lies in it being voluntary, not in it being large.

Once the foundation is there, the work is about sending the right thing to the right part of the list. Segmentation↗ is what separates a send from a conversation. Six steps:

1. 01Split the list by what you actually know. Role, industry, what the person downloaded, how long since the last visit. Attributes you guess at are worse than none, because they give a false sense of precision.
2. 02Let the signup source decide the first send. Someone who downloaded a comparison guide is closer to a decision than someone who signed up in the footer, and the two should not meet the same email.
3. 03Build a welcome flow. The days right after signup are when interest is highest. A couple of emails explaining what you do and what comes next do more good there than the same content does three months later.
4. 04Trigger on behaviour, not on the calendar. An email that arrives because someone did something is almost always more relevant than one that arrives because it is Thursday. That is also the line towards marketing automation↗, which is the system support for the same thing at greater scale.
5. 05Personalise the substance, not the greeting. A first name in the subject line changes nothing. Content that is about the recipient's actual situation does.
6. 06Let silence mean something. Someone who has not opened anything in six months should not get the same flow as an active reader. Either an attempt at re-engagement or a place off the list, but not the same send forever.

A small segment that gets the right email almost always beats the whole list getting the same one.

Most of what goes wrong in a send can be caught in five minutes. A check worth doing every time:

1. 01Read the subject line as a recipient. It decides whether the email gets opened and is the only thing visible before that decision. Do not promise more than the body delivers, because the disappointment costs more than the click you missed.
2. 02Check the sender name. A company name works for newsletters, a person for anything meant to be answered. What does not work is an address nobody reads, because then the conversation is closed before it starts.
3. 03Click every single link. A broken or misdirected link is the error that costs the most, because the email is already out when it surfaces and someone who clicked the wrong thing rarely clicks again. Check at the same time that tracking survived the link being shortened or edited.
4. 04Read the email on a phone. That is where most people open it, often in passing. The question is whether the message fits in what is visible before anyone has to scroll or zoom.
5. 05Send a test to yourselves. Not to the same address every time, but to a few different email clients, because they render mail differently and what looks right in one can fall apart in another.
6. 06Decide what you will look at afterwards. Which number would make you do something differently next time? If the answer is none, the send is not finished being planned, only finished being written.

The check takes five minutes and catches what is otherwise discovered only once the email is already out.

Two questions decide whether a send was worth anything: did it arrive, and did it make a difference. They are connected, because what makes recipients complain is the same thing that puts the next send in the spam folder.

• The sender needs to be verified. Receiving servers check against the domain's DNS that whoever is sending has the right to do so in the domain's name, in practice through the SPF, DKIM and DMARC records. The largest email providers now require them of anyone sending large volumes, and if that check is not in place the content matters less, because the email never reaches the inbox.
• Complaints and unsubscribes weigh more than opens. An unsubscribe is honest feedback. A spam complaint is more expensive feedback, because it affects how future sends from the same sender are treated.
• Open rate is a weak measure. It is counted through an invisible image that more and more email clients load automatically, which blunts the number. Why it does not work as a target is something we have written about in open rate is not a goal.
• Clicks say more, but not everything. A click means someone wanted to know more. What happened after the click is decided on the page they landed on, not in the email.
• Test one thing at a time. Subject line, sender name or offer, not all three at once. The method is the same as in A/B testing, and it assumes the segment is large enough for the answer to mean anything.
• What counts is what happened in the business. Enquiries, bookings, purchases and closed deals, traced back to the send. Same logic as in conversion optimization: measure the step you actually want to move.

If the measurement can only be read inside the sending tool, you are measuring the channel, not the business.

Most people searching the topic are really looking for a tool. We recommend no name, for the same reason we recommend no ad platform: the right choice depends on volume, integrations and who will run it day to day. What can be said is what a tool needs to handle.

• Consent management. Where the consent came from, when it was given and what it covered should be possible to show afterwards. Without that history the list is hard to defend the day someone asks.
• Segmentation on your own data. If the tool can only split the list by opens and clicks, the segmentation stays shallow. It needs to combine with what you know from forms and the business system.
• Flows triggered by events. Signup, download, purchase, silence. A tool that only sends manual campaigns ties the work to the calendar.
• Connections both ways. Contacts should be able to come in from the site and out to the business system. Otherwise you build yet another register nobody maintains.
• Export without obstacles. The list is yours. If it cannot be taken with you it is not an owned channel, it is a rented one.

The list of requirements is short on purpose. A tool that handles those five things lasts longer than one with a hundred features nobody uses.

A few recurring beliefs make the channel underperform long before the tool gets to matter:

• “Email is dead.” The claim comes back every few years and has so far always been about mass sending, never about the channel. As long as the inbox is where people run their working day, the channel is still there.
• “A bigger list is better.” A list that grows through bought addresses or forced signups grows in size and shrinks in value. The number of recipients is not a goal in itself, it is a by-product of someone wanting to hear from you.
• “The newsletter is the strategy.” The newsletter is a format. What you want to achieve, who you want to reach and what you have to say belong in the content strategy, and the newsletter falls out of it.
• “The more often the better.” Frequency without content to carry it produces unsubscribes, not attention. The pace should be set by what you can hold in a normal month, not by what would be optimal in a good week.
• “The tool solves it.” A sending tool does what you decided, faster and more accurately. It decides nothing for you, and it does not make thin content worth reading.

What all five share: they move the question from what you have to say to how you send it.