There is something inherently satisfying about setting up an automated marketing system. Whether it’s retargeting, welcome series, scheduled social updates, automated ad buys, there is a temptation to "Ron Popeil" these efforts (“Just set it, and forget it!”).
However, when done this way, automated marketing can not only multiply traditional marketing challenges, they can create a new set of challenges. To wit:
- The building of bubbles. As individuals, we unfriend those who don’t share our viewpoints and consume news that doesn’t offend our preconceptions. As a result, we train algorithms to give us what we want; they train us to want what we are given.The same thing happens with organizations. We acclimatize to the supporters we have, without thinking, about who else might support us.
- Customized cruelty. When marketing knows a great deal about you, then plays it back, the results can arouse suspicion at best. At worst, consider Facebook’s algorithmic “Year in Review” album for its users. One user saw a post “Eric, here’s what your year looked like!” with a picture of his daughter who had died that year. He shared this in a post titled “Inadvertent Algorithmic Cruelty.” The fact that it was inadvertent and algorithmic did not subtract from the fact that it was cruel.Or three years ago, OfficeMax sent a man a letter with the second address line, “Daughter Killed In Car Crash,” which had happened.
As someone who has lost a daughter, my first reaction of these would not be to share more information with Facebook or to buy office products.
- A stratification that goes against many of our missions. Cathy O’Neil, author of (the excellent) “Weapons of Math Destruction” wrote that the moment she quit her job to write that book was:
“… when I heard a venture capitalist describe his ideal for the future of tailored advertising: a world where he’d receive only offers for jet skis and trips to Aruba, and where he’d never again have to see ‘another University of Phoenix ad,’ because that’s not for people like him.”
It's an unfortunate fact that a person who can give you $10,000 is more valuable that one who can give you $100. But as many work to be voices for the voiceless, we must listen to and engage with all who want to be a part of our missions.
- One-way marketing. Algorithms don’t listen. We drop a mail piece, blast an email, broadcast our ads and publish blog posts. Even our verbs betray the lack of conversation with our benefactors. And we learn nothing when we don’t listen.
- Volume and frequency. Often when algorithms are built, there is little consideration for the tolerance of people to hear from you. This leads to blindness to your messages at best and active disgust at worst.
So, how do we get the advantages of automation without the disadvantages? A few tips:
- Listen and respond. A recent DonorVoice highlights how getting free responses from your supporters can help your retention rate, both for the individuals who comment and by improving your systems based on their comments.Feedback also functions as an early warning system for campaigns you may have that are creeping people out or not landing as you might wish. A “Year in Review” campaign that sounds good in concept can be pulled back after the comments start rolling in.
- Build in spontaneous interactions. If your data says that Tuesdays and Thursdays at 9 a.m. ET is the best time to do your Facebook posts, the algorithms will eventually make you irrelevant to the people who check Facebook most on weekends. Make sure you are building in some messages that are suboptimal on average, but help with a certain audience.
- Speaking of, test to see if you have difference audiences. There are messages that may not work on average, but do for a subgroup. We worked with one health charity that found that their in-home nursing care program didn’t sell overall. However, for the people who had received or were receiving their services, they cared about little else. This allowed them to create an audience and a donor identity that had a better response to a new message (and they weren’t ham-handed enough to say “because you’ve had nursing care in your home”).
- Test different frequency and cadences of communications. Every donor will have their breaking point for how much is too much communication. You can test this in three ways: 1) by reducing frequency overall, like some charities are doing (click here for an example); 2) by reducing frequency algorithmically; and 3) by asking donors what they want.
This last point is the number one way to have effective automated marketing: Build from explicit data, rather than implicit data whenever possible. A cat person won’t be freaked out by getting all cat ads, if they told you they were a cat person. The defense of an algorithm marketer is always that targeted ads are better, because they give people what they want. And that is true, if and often only the person has told you that information themselves. If not, they may not want that to be a customization point.
The good news is not only that people are often willing to tell you how and how often they would like to be communicated with, but also that asking them this makes them more likely to sign up for your organization, as extensive testing by Kiki Koutmeridou, PhD, has shown.
These simple tips can prevent your automated marketing strategies from going off the rails and make it so your donors and supporters welcome your communications.
And, since I try to practice what I preach, what do you think of this post? Please let me know at firstname.lastname@example.org or in the comments below. Thanks in advance!