The short answer
For a service business, results come from a small list of past customers, timing tied to their last service date, and a plain email from a person. Subject lines that state a specific fact beat clever ones. Authentication decides whether any of it arrives.
Who is on the list matters more than the writing
A roofing company with four hundred past customers and a plumber with four hundred past customers have the most valuable email list in their market, and most of them never send anything to it. Meanwhile the same owners will pay for a campaign to reach strangers who have never heard of them.
Start by exporting every customer you have invoiced in the last five years. That list already knows you, already has your work on their house, and already decided you were worth paying once. Everything below applies to them first and to anyone else second.
What actually gets opened?
Subject lines that state a specific fact. "Your furnace filter is due" gets opened. "We will be in Middletown next Tuesday" gets opened. "Spring Newsletter" does not, and neither does anything with a question mark and an emoji, which is the format most templates push you toward.
Send from a person's name at the business, not from a department. "Dave at Coastal Heating" outperforms "Coastal Heating Marketing Team" for the same reason a truck with a name on it does. And send plain-looking email. The heavily designed template with a hero image and three columns signals mass mail before anybody reads a word of it.
One email, one job
Decide the single thing you want the reader to do and remove everything that competes with it. One link, repeated twice if the email is long. A newsletter with five sections and five links gets a fraction of the clicks of a short email asking for one thing, and it takes four times as long to write.
Keep it to the length you would use in a real message. Three or four short paragraphs. If you are including a price or a date, put it on its own line where a scanning eye will land on it.
Timing beats targeting for a service business
Retailers segment by browsing behavior. You do not need any of that. Segment by the date of the last job, because that is what predicts the next one.
A dental practice sends hygiene recall at six months because that is the interval. Heating and cooling sends the tune-up reminder in the two weeks before the first genuinely cold night, not on September 1 when nobody is thinking about it. A landscaper sends before the first hard frost. A roofer sends after a storm, to the streets that took the storm. That is the whole segmentation strategy, and it beats anything more sophisticated.
Write the send date on a calendar a year out as soon as you know the trigger. The reason most service businesses send nothing is not that they cannot write. It is that the right week arrives during their busiest fortnight and the email never gets drafted. Write it in the slow month and schedule it.
Why do my emails end up in spam?
Usually because the sending domain is not authenticated. Your sending platform needs SPF and DKIM records published in your domain's DNS, plus a DMARC record. The major inbox providers now require authentication and a one-click unsubscribe from anyone sending in bulk, and they enforce it by quietly filtering rather than bouncing, so you will not get an error telling you what went wrong.
Two more things sink deliverability. Sending from a free address in the from line, which fails authentication by design. And a list you have not mailed in three years, where a chunk of the addresses are now dead and spam traps. If you are restarting a cold list, send to the most recent customers first and work backward over a few weeks. Whoever manages your email setup should be able to show you all four DNS records in about a minute.
Measure replies and booked work, not open rates
Open rates stopped being reliable when inbox providers started pre-loading images, and they were never the point anyway. Count the replies, the calls that come in the two days after a send, and the jobs booked. For a list of a few hundred past customers, a handful of booked jobs from one email is a good send.
What email will not do
It will not create demand that does not exist. Nobody replaces a working water heater because they got an email. What it does is make sure you are the one they call when it fails, which is a different and slower kind of win.
It will also not rescue a purchased list. Those addresses did not ask for anything, complaint rates run high, and a few complaints can damage the domain reputation you need for the customers who did want to hear from you. Unsubscribes are not a failure either. A smaller list of people who want the email is the asset.
Follow-up questions
Monthly at most, quarterly for most trades, and always with something specific to say. A service business that emails only when it has a real reason gets better response than one on a fixed schedule with nothing to report.
Yes, once you are sending to more than a few dozen people at a time. Bulk sending from a normal mailbox will get filtered and can damage your domain. Any of the mainstream platforms will handle authentication and unsubscribes correctly.
Leave them for a year, then send one email asking whether they want to stay on the list, and remove the ones who ignore it. Carrying dead addresses lowers the odds that your active customers see the next send.
Let’s build something.
Twenty minutes on the phone and you’ll know whether we’re the right fit. Either way you’ll leave with something useful.