Does this bring in money or votes? Newsletter 23
Hello, everyone, and welcome to the latest edition of “Does this bring in money or votes?” -- I’m Anna Kelsey, the DNC’s Grassroots Prospects Manager, back to tell you more about the reactivation program that my colleague Jose overviewed last week: how we stood up the DNC’s first-ever supporter reactivation program in the last three months of the cycle -- one that reactivated millions of supporters and accounted for over 16% of the DNC’s online fundraising revenue in the last quarter of the cycle -- and the learnings you can use to guide your own reactivation work!
The DNC has been running an email program for nearly two decades, which means we have millions of email subscribers who aren’t active on our list anymore. But with the most important election of our lifetimes on the horizon, the DNC team realized this was the moment to try to get those supporters back on board -- and that’s why they brought me on board the email team!
My mission: reactivate as many of our lapsed subscribers as possible and get them back to engaging with and donating to our main email list, without sacrificing our excellent and hard-won inbox placement.
As you can see above, we were successful on that second piece! Read on to hear more about what we did for part one of that mission and our thought process behind it -- but before we get into that, here’s...
What We Learned
The work we did was possible because of the proprietary trove of historical data that the DNC has collected over early two decades of emailing supporters -- but there were still some takeaways that other organizations can use when thinking about reactivation:
Keep your data, no matter what! Our historical behavioral data was incredibly important to the success of this program, and throwing it away -- even data as granular as individual opens and clicks -- would have cost us millions of dollars.
Reactivation requires planning and care. If we’d simply mailed random groups of inactives blindly, we would’ve had serious deliverability problems: when we tested sending to a tiny group of names we deemed “risky,” the bounce rate was 94% higher than on our regular inactive sends. Taking the time to pick the right audience is worth it!
However you approach reactivation, find a way that allows you to be both systematic and nimble. More on the specifics of what we did below -- but what it really came down to was both having detailed plans and being ready to change course quickly in response to results. Your reactivation program might look different from ours, but you should find a way to juggle both of these needs -- otherwise, you’re almost certainly leaving money on the table.
Targeting our inactive subscribers wasn’t as straightforward as it might seem -- before we could actually email any inactives, we needed to identify who was available to us and what data we had about them. We also aggregated all of the data points we had for each lapsed subscriber, like how many times they had opened or clicked in the past, which cycles they’d engaged in, their donation history, and their status on the OFA and HFA lists.
We eventually used these points to divide our inactives into three main buckets -- very recent inactives, deep inactives with little to no historical data points, and deep inactives with many historical data points -- and ran two distinct programs.
Program 1: The Reactivation Series
The first program was a five-message automated reactivation series which went to (relatively) small cohorts of inactives with similar historical behavior. The series featured the types of messages that have been most successful on our main list -- a “greatest hits,” if you will!
Breaking this group up into cohorts had two main advantages:
First, it allowed us to prioritize targeting inactives that I believed had the highest propensity to reactivate through the series first, getting them into our primary email program with the most time to donate before Election Day.
Second, this program helped us identify what criteria are the best indicators that an individual will reactivate -- which will help us build on and optimize this work in future cycles.
Program 2: One-Off Sends
Our reactivation series was great for bringing in supporters we thought had a very good chance of reactivating back into the fold and for comparing cohorts -- but it didn’t take advantage of important campaign moments or the urgency that kept ramping up the closer we got to Election Day. That’s why, in conjunction with the series, we implemented a strategy of sending one-off emails at key moments to much larger groups.
For this approach, we mostly sent to larger swaths of more recent inactives and deep inactives without much historical data, and chose sends around big moments of excitement, like the VP announcement and the debates, along with emails from high-performing surrogates, like Barack Obama, Hillary Clinton, and (of course) Kamala Harris and Joe Biden.
Altogether, the reactivation program was hugely successful: Between our two approaches, we sent every single emailable inactive at least two emails and reactivated millions of supporters -- who accounted for 16% of our online fundraising revenue in the last quarter of the election.
Having these two distinct -- but complementary -- approaches was critical to the success of the program overall. Our one-off sends helped us ensure that we weren’t leaving anything on the table in critical moments, and were a source of a higher volume of reactivations overall. The series gave us actionable data we can use in future cycles -- and takeaways that we were able to use towards the end of the cycle to cut the widest possible reactivation prospect audiences, as we shifted more resources into one-off sends. We’re also excited to look at the long-term results of this program and measure which of these two approaches leads to more sustained engagement with our primary email program after Election Day and into 2021.
🧵Thread we think you should read (and retweet) 🧵
Thank you for joining me on this reactivation journey! It was an incredibly exciting project to work on over the last few months, and I hope you enjoyed reading about it. If you’re dying to know how we did this, stay tuned for a later newsletter with more of the practical details about the tech tools & data crunching that made this all possible!
Happy last month of 2020(!!!),