Why does one email take two weeks?

I get asked this a lot, usually by someone trying to decide whether their team is slow. The honest answer is that it depends on what you count. There are two numbers hiding inside the question and they have almost nothing to do with each other.

The first is touch time: the hours a person actually spends building the thing. The second is elapsed time: the days between someone asking for the email and the email going out. In most enterprise teams the first is measured in hours and the second in weeks and the interesting question is what happens in between.

Where the hours go

Take an ordinary campaign. One email, an existing template, an audience that already exists or a list somebody sends over, a standard approval. Built by marketing ops from a ticket. Here is roughly where the hands-on time goes in the teams we work with.

• Reading the brief and asking what’s missing. Which segment exactly. Which sender. Is the copy final. Is this the right template. Half an hour if the brief is good and it usually isn’t.
• Building the email. Copy into the template, images sized and placed, links checked, preview text written. An hour or two if the template behaves. Longer if it needs a new block or the request came as a Word document with formatting nobody can reproduce.
• Setting up the campaign or program. Cloning, naming by the convention, campaign fields, email group, schedule. Half an hour to an hour, with a nonzero chance of a typo in the name that shows up in reporting for the next year.
• The audience. Picking a segment is minutes. A list that needs cleaning before it can be uploaded is an hour on a good day and an afternoon on a bad one.
• Testing. Test sends, a read on mobile, a click on every link. Half an hour, then another half hour after the first round of feedback.
• Scheduling and launch. Ten minutes, once everything else is done.
• Pulling the results afterwards. Twenty minutes per request and there are usually two or three requests.

Add it up and a routine send is somewhere between two and eight hours of a skilled person’s time. When 3M measured theirs before they changed anything, it was eight. Eight hours for a campaign that nobody would describe as complicated.

Where the weeks go

Now the other number. None of those hours happen back to back. Between them the campaign waits.

It waits in the queue behind the fourteen requests that arrived earlier. It waits while ops asks which segment and the marketer is in meetings. It waits for the copy to be final. It waits for the list. It waits for the test send to be looked at. It waits for approval which means it waits for a specific person to open a specific email and if that person is on vacation it waits for them to come back. Then it waits for the send window.

Most of the two weeks is nobody doing anything. That is worth saying plainly because teams trying to get faster almost always attack the hours, with better templates and tighter briefs, when the days are mostly made of waiting.

Most of the two weeks is nobody doing anything to the campaign at all.

What should be faster

Look back at the list of hours and notice how much of it is the same every time. The email goes into the same template. The program is cloned from the same template program. The name follows the same convention. The audience comes from the same handful of segments. The test is the same test. The results are the same four numbers.

Everything that is the same every time should take almost no time, because software can do it. A template with locked zones means the email is built the moment the copy is pasted in. A program template with tokens means the clone is ready when the dates are filled in. A naming convention applied by a rule means no typos. An upload that is validated before it is imported means no afternoon of cleaning. Results that show up in front of the marketer mean no ticket for the numbers.

More to the point, when those steps are automated the marketer can do them. The hours stop being ops hours. The queue is where the weeks came from and it gets shorter because most of what used to be in it never enters it.

What shouldn’t be faster

One step on that list deserves to stay slow, or at least human. Somebody should read the content and look at the audience before the send goes out. The template already handles the fonts and the footer. The reviewer is there for whether the words are right and whether these are the right people.

The trick is to make that review quick by giving the reviewer less to check rather than by skipping it. If the template can’t be broken, the sender is fixed, the audience came from an approved list and the upload was already validated, then approval is a read of the copy and a glance at the count. Five minutes rather than a day, with a deadline on it so it doesn’t turn into a week.

What fast looks like

At 3M the same campaign that took eight hours of ops time now takes about 45 minutes, most of it the marketer’s. Regional teams send 120 to 200 campaigns a month that way without the ops team growing. Across Campaign Launchpad customers the average campaign launches in about 30 minutes from the marketer opening the wizard. Nobody works faster. The eight hours were mostly work that didn’t need a person and the two weeks were mostly waiting that didn’t need to happen.

Do the arithmetic

Multiply your hours per campaign by your campaigns per month and you have the ops capacity being spent on routine work. Multiply the days by the same number and you have how long your marketers are waiting, in aggregate, for things that could have gone out the same afternoon. The calculator on the Launchpad page does this with your own numbers and the guide to campaign request intake goes through what to automate first.

The hours matter more than the speed. Get them back and ops spends them on the campaigns that need an expert which is the work the team wanted to be doing all along. The marketers get Thursday’s email out on Thursday. Nobody has to wonder anymore whether the team is slow, because the question stops coming up.