Self-service marketing without the cleanup afterwards

I spent a long time on the requesting side of marketing operations before I spent any on the building side so I have been both people in this argument. The marketer who needs the email out by Thursday and can’t understand why that takes a ticket and a week. And the person looking at the ticket queue thinking that if one more team gets a login to the platform she will be cleaning up after it until spring.

Both are right which is why self-service projects go around in circles. Here is how I’ve come to think about it.

What marketers mean by self-service

When a marketer asks for self-service they are picturing something specific and it is smaller than ops fears. They want to send the monthly newsletter this afternoon instead of next week. They want to change the date in the reminder without filing a ticket for a one-word edit. They want to load the list from the event they ran yesterday. And the morning after a send they want to see how it did without asking anyone.

They are not asking to build segments, configure programs or learn what a decision step is. Most would be horrified to be handed the campaign canvas. Self-service to them means the things that feel like theirs: the words, the pictures, the timing, the audience from a short list of sensible options, the results.

What ops hears

Ops hears “login.” And a login means the all-contacts segment is one click away, every template is editable including the footer with the legal text in it, the upload wizard will overwrite ten thousand job titles with blanks if the mapping says so and the advanced sending options include a check box for emailing unsubscribed contacts. Ops has seen each of these happen and isn’t being precious.

The trouble is that a login is how the platforms offer self-service, because they are built for people who work in them. You can narrow it. Security groups in Eloqua and roles in Marketo decide what a person can open and the simple email campaign hides the canvas. Even narrowed, a login gives a marketer far more surface than they wanted and far less guidance than they need. It tells them what they may touch and nothing about whether what they did with it was right.

A login gives a marketer more than they asked for and less than they need.

Decide what to hand over before you decide how

Skip the tooling question for a moment and write two lists. The first is what marketers should decide for themselves. The second is what ops should never give up. In the teams I work with the lists come out looking roughly like this.

Hand over:

• The copy and the images, inside whatever space the template leaves open.
• The subject line, within a length that fits.
• The send date and time, inside the windows ops allows.
• The audience, chosen from a short list ops approved for that kind of campaign.
• A list upload, as long as the file is checked against ops’s rules before a record is written.
• Their own results, the morning after, without a ticket.

Keep:

• The templates and which parts of them can change.
• How audiences are defined, including every suppression and exclusion.
• What a valid upload looks like and which fields it may touch.
• Sender addresses, email groups, tracking and everything else that lives in the platform’s settings.
• Frequency rules and whatever else keeps one contact from hearing from six teams in a week.
• Naming conventions and campaign fields so reporting still works.
• Who approves what and how many times.

Once the lists exist the tooling question answers itself. Whatever you use has to let marketers do everything on the first list and make it impossible to do anything on the second. If it only does the first half you have a login with extra steps.

Roll it out one campaign type at a time

Nobody should turn on self-service for everything at once. Look at a quarter of requests and find the type that is most of the queue. It is almost always a single email to an existing audience: the newsletter, the webinar invitation, the product update. Build that one path properly. Lock the template, define the audiences, write the upload rules, decide who approves and switch it on for one team.

Then measure three things. How long a campaign takes from request to send. What share of campaigns went out without anyone from ops touching them. How often something had to be redone. If the first two move and the third doesn’t, add the next campaign type. If the third moves, the guardrails need work before anything else does.

Keep a path back to ops for the exceptions. Some campaigns will always need a step a template can’t anticipate and the worst outcome is a marketer improvising around it. The path should be part of the same process rather than a side email so ops sees the request with everything the marketer already filled in and the campaign stays in one queue with the rest.

The governance job gets bigger after self-service

One more thing from the requesting side. The easier it is to send, the more gets sent. That is the point and it also means somebody has to watch the total. I wrote about that in Who owns the inbox? and it matters more after self-service than before. The frequency rules belong on the keep list for a reason.

How Campaign Launchpad draws the line

This split is what Campaign Launchpad is built around. Ops builds a wizard for each kind of campaign and decides what it shows: the templates, with the zones that are open, the audiences, the fields a marketer fills in, the rules an uploaded list has to pass, the approvers. Marketers run the wizards without a login to Eloqua or Marketo so everything on the keep list is out of reach because it was never on the screen. Campaigns that need a step ops does by hand are routed to the right person with a deadline and every campaign reports its own results back to the marketer who sent it.

3M put 250 people onto it that way, with one small ops team behind them and no new hires. The marketers got what they meant by self-service. Ops kept what it meant by control. That is the whole trick and the tooling is the easy part once the two lists are written.