Letting the calendar go – why evergreen marketing wins

Marketers are bound to their calendars, and rightly so. Events must be advertised, creative must be designed, mailers and emails must go out – a marketer without a calendar is surely in a muddle. And of course, some customer journeys cannot be divorced from a calendar – a Black Friday sale can only happen on Black Friday, and New Year’s Resolutions to join a gym are only made in January. But many of a lifecycle marketer’s core campaigns – e.g. winback, cart abandonment, referral, renewal, repurchase, cross-sell – make more sense, and will be more effective, structured as evergreen. 

It may feel natural to plan and execute cart abandonment or winback as a calendared campaign – to follow clear steps with a beginning, middle, and end. But insisting that every campaign be tied to a calendar has two significant disadvantages – one obvious, and one more subtle. In this blog post, we unpack the advantages of embracing evergreen campaigns.

Content creation is an unending burden

Consider a retail brand running a repurchase and upsell campaign. Every two weeks, they refresh the campaign with new creative to highlight different products. In effect, the brand has turned what is naturally an evergreen campaign into a series of calendared campaigns, each lasting two weeks. One problem is that this approach is simply a lot of work. The marketer is constantly ordering new creative, writing new copy, and approving new designs. Generative AI may one day ease this burden, but most enterprise marketers have the same workflows for producing creative that they did before the arrival of chatGPT.

All this effort can create an illusion of productivity – the emails themselves become the deliverable. But let’s be real. Marketing emails aren’t Super Bowl ads, and your customers probably don’t find them especially memorable. Would a customer really notice if they saw the same creative in two emails two months apart? You might not want to use the same subject line with the same customer twice in one week, but with reasonable guardrails, a marketer could get significantly more value from their existing assets.

How do you tell what’s working?

However calendared campaigns present a more fundamental challenge – how do you know if they’re working? Of course marketers can measure outcomes – like conversions, revenue or profit – but these are trailing indicators which are not necessarily actionable for the marketer. Knowing that revenue was up in the last two weeks does not, in and of itself, tell you which subject lines, offers, or sending times were effective. While it may be clear enough which email was the global “winner” – or at least which email had the most conversions overall – what marketers really want to know is how to personalize and pick the "winner" for every customer. 

Measuring success with manual A/B testing or multivariate testing is always slow, but for a brief calendared campaign, it’s hopeless. By the time a marketer has data, the campaign is already over. It will be hard to apply lessons to the next campaign when so many variables are changing, and constantly running new tests is prohibitively laborious. All too often, the outcome is to simply send this week’s emails to every customer this week, and then send everyone the next week’s email next week. The rapid pace of the campaign means that we don’t have time to personalize, or at least not to personalize in a way we know is optimizing the campaign -- we don't have the capability to learn which subject lines, offers, or sending times work best for each individual customer.

Rather than trying to optimize with manual testing, a marketer could personalize a calendared campaign with AI Decisioning – AI that finds the best message, creative, offer, sending time, etc. for each customer. However, learning quickly in a short campaign can be challenging even for AI. (Essentially there are two requirements – the AI Decisioning agents must apply transfer learning from similar campaigns, and the messages and creative must be parameterized in such a way that the AI can make comparisons across campaigns. OfferFit’s AI Decisioning Platform has these capabilities, which provide 1:1 personalization for campaigns that truly must be calendared.)

Nevertheless it's often better to determine if the campaign needs to be calendared in the first place.

Why evergreen marketing can win

Evergreen campaigns give marketers two key advantages over a calendared approach. First, letting marketing assets – creative, messaging, subject lines, headers – stay in the rotation longer simply saves time and effort. But there is a more fundamental advantage. The longer a campaign runs, the more time a marketer has to gather data and measure what’s working. That makes it possible for the marketer to run A/B or multivariate to discover what's working best for their global audience, or perhaps for segments. 

More importantly, evergreen campaigns are a natural fit for continuous automated experimentation. Marketers hoping to discover the right message, creative, offer, sending time, and frequency for each individual customer – that is, marketers hoping for the dream of 1:1 personalization – will want to deploy AI Decisioning. This type of AI, called reinforcement learning, continuously learns and empirically discovers the best choices for every customer. Of course, marketers will change offers and creative over time – the AI will adapt and continue to optimize the overall campaign. While it’s possible to use AI Decisioning in calendared campaigns, this complexity is often unnecessary. Most lifecycle marketing campaigns are naturally, and most successfully, structured as always-running and evergreen. 

Nathaniel Rounds writes about AI and machine learning for nontechnical audiences. Before joining OfferFit, he spent 10 years designing and building SaaS products, with an emphasis on educational content and user research. He holds a PhD in mathematics from Stony Brook University.