Phillip Wild

From inflexible customers journeys to engagement-based marketing

Discussion created by Phillip Wild on Jul 9, 2018

-- Posting as a discussion since I don't have blogging rights --


Original post here:


Creating ‘customer journeys’ is all the rage right now. This sexy-sounding phrase describes building out a series of automated communications for customers who interact with your brand in different ways. Great. Right? Well, not always.


While it’s true that triggering emails or other automation based on an interaction will be more effective than the old-school “batch and blast”, if you’re not careful, you can create a house of cards that will be impossible to manage and just doesn’t scale.

Let’s start with a conversation many marketers have had in the past few years.


Big Boss: Phil, what happens when someone orders one of our catalogues? Let’s build out a journey for that.

Phil: Sure thing! Ok, so autoresponder email when they order, wait 3 days, send sales email.

Big Boss: Phil, what happens when someone has been visiting our website a lot? We need a journey for that.

Phil: Ok, send email pushing them to sales, wait 1 day, then if they’ve purchased send them a thank you email, if not, send the same email with a different subject line. Ah, but during that time they could have ordered the catalogue — so suppress anyone from THAT journey during this one.


Expand this over a few months and now you have not two journeys, but perhaps a dozen. Or fifty. Hey, I don’t know how involved your CMO gets or how much of a sucker for punishment you are. So now there are two possibilities:


  • You have many customer journeys in action, each with extremely complicated logic and wait steps to ensure they don’t interact with each other in an undesirable way (for example, sending three emails in one day, just by coincidence). Whenever you are asked to create a new one, you die a little inside.
  • Each journey is not aware of any other journey, and you have customers going through different pathways at the same time, leading to a very confused customer.


Invariably, your marketing automation instance looks the back of a TV set in 2012 - a mess of cables and cords! This is not good. You’re getting confused. Your customers are getting confused. It’s a mess. So what can you do instead? Well, here’s our thinking.


  • We have a complicated product with no clear markers between first knowing a customer and them booking a tour. They could order a brochure, but they don’t have to. They could call up our call centre, but they don’t have to. They could visit our website, but they don’t have to. Unlike many B2B purchases, we don’t have any “lines in the sand” where we can funnel customers — they interact with us on their own terms, and in the channel they prefer.
  • Regardless of which channel they use and how they got there, a customer who is ready to book is ready to book. A customer who is simply researching is, simply researching. Yes, some actions are more important than others, but the output should be that a customer is ready for a particular message (discover our brand, view our range, look at dates, book now), and that’s the most important thing.


So we designed, from scratch, a lead scoring model that would place our prospective customers into different buckets. You get a large number of points for high-value actions (such as calling in or requesting a brochure) and a lower amount of points for lower-value actions (such as subscribing to our email newsletter or entering a contest). These points will place you somewhere on a “Propensity to Buy” scale. Long story short, how likely are you to buy our product, right now?


This scale determines what messages you are ready for. Someone who has been all over our website, ordering brochures, and calling in will be ready for a sales message. Someone who has only once or twice clicked a blog story in our newsletter is likely not. And to ensure this isn’t simply a point-in-time look at our database, scores decay with inactivity. So now we have different buckets of people, each at different stages of the buying cycle. Now, we need to create appropriate, “evergreen” content for them. Our most engaged people will enjoy a link to our Travel Deals page. Our least engaged might enjoy a Buzzfeed-style list of where to travel this year, as voted by our subscribers. Once this is built out, you can use your marketing automation platform to “drip” this content out, at a set cadence, as people enter and exit the particular buckets. No one will receive the same content twice, but as they move up and down in engagement, they will receive appropriate communications. And if they book, they will be removed from these streams entirely and pushed into a “pre-travel” stream.


Here’s what this could look like with a series of weekly drip programs:


  • John goes to our website and signs up for our email newsletter on a Monday. He’s now in our database at a low level of engagement.
  • Tuesday our drip stream fires out a “Where to Travel” email. John clicks it but doesn’t move beyond the blog post.
  • Several days pass and John seems to have lost interest for now. The next Tuesday, our new low engagement email goes out to him: it’s speaking about the best places to hike in the world.
  • John loves it and clicks through past the blog article to investigate our tours. He finds one he likes and starts looking at dates. Fantastic! It seems we run a tour in the week he had vacation time planned this year. He adds the tour to his wish list. His points start to stack up. He’s now in the “Highly Engaged” grouping.
  • Next Tuesday, John gets our first “highly engaged” email: a push to check out our travel deals. The tour on his wish list is on sale, and he books online. Fin.


So what happened here? Well in the traditional “customer journey” approach, John might have triggered three different journeys here, all competing for his attention: the subscriber start journey, the website visit journey, and the wish listed tour journey. All potentially sending overlapping communications, or at the very least, making a very complicated back-end of your marketing automation software. But in our system based on the level of engagement, the difficulty is removed: we know that John is going to receive something relevant to his current situation — whether those emails deploy daily, weekly, or monthly.


As an added bonus, we now know that if as a business we need to send something out to the database (or a subset of it) we only need to avoid the days that we have earmarked for nurture deployments. Everything else will be free space. If you have journeys built upon wait steps of a particular duration, then that will never be the case: there will always be someone who happens to be receiving the journey email that same day, since “3 days since original interaction” would be different for everyone who flows through that step.

This solution might not be the right one for all types of businesses, but if you have a complicated product with a long lead time and few set milestones, you may find it reduces your headaches significantly and gives a better customer experience. Win-win.