◈Campaigns in practice

In three steps that are unlocked one after the other — you cannot get to the next one until the previous one is in place.

1. Recipients. Who are you reaching? The number is shown before you move on.
2. Content. Subject and text, with a preview.
3. Check and send. A test send to yourself, then the real send.

What you see in every step: the status in plain language — how many recipients, whether a test send has run, whether anything is missing.

What you should do before pressing the last button:

• Read the number of recipients once more. Do not skim it — read it.
• Actually look at the test send, not just trigger it.
• Ask yourself whether the number matches what you wanted to reach.

The third point has already prevented sends that would have gone wrong.

Enlarge

Enlarge

Do not send. That is the most important sentence in this chapter.

An unexpectedly high number almost always means that a filter is not taking effect. The most common causes:

• A property is empty. A filter on a field that is not filled in for anyone narrows nothing down.
• Two filters that do not exclude each other. “Region A or Region B” instead of “and”.
• A category with no assignments. If no entry has been assigned the category, the selection is not empty but undefined.

The last case once proved expensive: a campaign with an unfilled category reached all contacts instead of the intended subset — 141 mails. Since then there has been a safeguard that catches exactly this case.

Your next step: go back to step 1, check the filters one at a time, look at the number after each step.

From coarse to fine, looking at the number after each step.

The procedure:

1. Start with the coarsest property — category, type of contact.
2. Look at the number.
3. Next property — region, status.
4. Look at the number.
5. And so on, until the set is right.

Why step by step: if the number stays the same after a filter, that filter is not taking effect. You only notice this if you go one at a time.

What you should also do: open a few recipients as a spot check. Do the three you look at fit? If not, the selection does not fit.

And the rule for unclear cases: too narrow is better than too wide. A campaign that reaches 40 instead of 60 people can be extended. One that reaches 400 instead of 60 cannot.

Seven things, in this order:

1. Does the mail arrive at all? If not, the problem lies with sending, not with the text.
2. Does it land in the inbox or in spam? Either tells you something.
3. Is the sender right? Name and address, as the recipient is meant to see them.
4. Is the subject complete? Phones cut it off early.
5. Are the placeholders filled? “Dear {Name}” in the test send means: in the real send too.
6. Do the links work? All of them, clicked one by one.
7. What does it look like on a phone? Most people read it there.

And the eighth point that people forget: read the text once in full, as if you were the recipient. Not as the author.

A test send that you trigger and do not look at is not a test.

Then the value is missing for this recipient — the placeholder itself is fine.

What to do:

• Check before sending how many recipients have this field empty. If there are many, do not use the placeholder.
• Set a fallback wording where possible — “Dear Sir or Madam” instead of an empty name.
• Complete the records, if there are only a few.

The pragmatic rule: only use placeholders for fields that are practically always filled. Name and company usually yes; contact person and position often not.

What comes across as particularly awkward: a placeholder in the salutation. It is right at the top and is the first thing that gets read.

And the test for this: a test send to yourself with a record in which the field is empty.

For business mail: Tuesday to Thursday, in the morning.

Why:

• Monday is the day the weekend gets worked through. Your mail sits behind fifty others.
• On Friday afternoon nothing new gets started.
• Mornings are for reading, afternoons for working through things.

What matters more than the day of the week:

• Do not send into a holiday period. Start of the school holidays, between Christmas and New Year, bridge days.
• Not into an event that takes up people’s attention.
• Not directly after the last mail. If you wrote two weeks ago, wait.

And the caveat: these are values from experience, not laws. For your audience it may be different — try it out and look at the open rates.

What we advise against: sending at night. The timestamp is visible.

Once, after about a week, with something new.

What makes a good follow-up:

• It does not repeat. “I just wanted to check in” is not a message, it is a reminder notice.
• It adds something — an example, a figure, a concrete suggestion.
• It is shorter than the first mail.
• It makes ending easy: “If now is not a good time, I will get back in touch in six months.”

How often: once. Twice at most. After that it is harassment — in the legal sense too, if an objection has been made.

And the most important point: anyone who has objected gets no follow-up. This applies without exception and is taken into account technically — but check it anyway if you put together a recipient list by hand.

Recalling is not possible. What you can do, in this order:

1. Stop immediately if sending is still in progress. Every mail not sent counts.
2. Establish who is affected. Exact number, exact list.
3. Assess what happened. Wrong text? Wrong recipients? Third-party data visible?
4. Decide whether to get in touch. With a wrong text, usually not. With visible third-party data, yes.
5. Check whether it must be reported. If personal data has reached unauthorised persons, a report to the supervisory authority within 72 hours may be required.
6. Record what happened — even if you do not report it.

What you should not do: send a second mail “Please ignore the previous one” to everyone. That doubles the attention paid to a mistake that many would not have noticed.

And afterwards: eliminate the cause, not just the case.

Clearly visible, in every advertising mail, and it has to work.

The requirements:

• In every promotional message, not just the first.
• Free of charge apart from transmission costs.
• Without hurdles — no login, no form with required fields, no justification.
• It must take effect, and promptly.

What goes wrong in practice:

• The notice is so small and grey that it is barely legible. That can be regarded as obstruction.
• The unsubscribe is recorded, but the next campaign had already been put together beforehand.
• An unsubscribe only applies to one list, not to the sender.

Check it yourself: unsubscribe once with an address of your own and see whether the next campaign still reaches you.

It is usually due to the sender configuration, not the text.

Check the technical side first:

• Have the sender authorisations for your domain been set up? Without them, many receiving systems downgrade every mail. This is the most common reason.
• Does the sender address match the domain from which mail is sent?

Then the content:

• Many links to different destinations look suspicious.
• A pure image mail without text does too.
• Trigger words in the subject — “free”, “guaranteed”, strings of exclamation marks.
• Attachments in the first mail.

And a point that cannot be bought: if many recipients delete your mails unread or mark them as spam, delivery drops for all your mails. That is the strongest reason to choose recipients carefully.

If the technical setup is missing, contact whoever looks after your domain.

Less than you would hope — but enough to learn something.

What is reliable:

• Delivery. How many mails arrived? Anything below 95% points to bad addresses.
• Bounces. Invalid addresses should be cleaned up, not reused.
• Replies. The only figure that means business.

What is uncertain:

• Opens. Measured via a tracking pixel that many programs block. The number is a lower bound, not a value.
• Clicks. Somewhat more reliable, but not complete either.

What you do with them: compare campaigns with each other, not with industry figures. If one brings three times as many replies as the other, you know something — even without absolute figures.

And the only figure that counts: how many conversations came out of it?

Yes — with two restrictions, both of them important.

First: not to the same recipients. Anyone who has already received the mail does not get it again. It looks like an oversight and annoys people.

Second: not to those who have objected. That is mandatory.

What makes sense:

• The same campaign to new contacts who have been added in the meantime.
• A revised version to a different audience.
• A repeat after a longer time — six months or more — to recipients who did not respond, but did not object either.

How you do it in practice: take the campaign as a template, put the recipients together afresh, go through the text and update it. What was current six months ago may not be today.

And always: a new test send. Even with a familiar campaign.

Yes, and then you need agreements, otherwise contacts get three mails in one week.

What you should settle:

• Who sends to whom. By region, by industry, by responsibility — some kind of split.
• A shared calendar in which sends are entered before they take place.
• A minimum gap between two campaigns to the same person.

What helps technically: look at the contact to see when something last went to them. If it was only two weeks ago, wait.

What causes the most trouble: two colleagues working on the same company without knowing it, and writing to two different contact people with the same offer. That gets noticed within the company and looks disorganised.

The simplest safeguard: before every send, briefly say whom it is going to.

Yes, as long as it is running — and then speed matters.

What happens: sending is paused. Mails already sent are gone and do not come back. Those not yet sent no longer go out.

How much you save: depends on how quick you are and how large the campaign is. With a large campaign, that can be the greater part.

What you do afterwards:

1. Establish how many are already out.
2. Decide whether there is a reason to get in touch.
3. Fix the cause.
4. Only then send again — to those who have not yet received anything.

Point 4 is often forgotten: if you send again to everyone after the correction, some people get two mails.

And prevention: the test send is the point at which stopping would not have been necessary.