AC equivalent of "The Orbit" for non-linear purchase paths?


has anyone seen or developed anything similar to this in AC? ->

“The Orbit” is Jermaine Griggs’ answer to non-linear purchase paths. It’s basically a circular automation that runs and is constantly determining the optimal product/offer to present to the prospect based on past behavior, purchases, etc. It deconflicts against past offers and current sequences so it’s like having an intern who looks at each prospect case-by-case constantly and serves the most ideal content at all times.

I’m looking to build an elegant nurture “engine” for Pagely that will run for all prospects and drip them the most optimal content based on their stage in the sales cycle, past behavior, stuff extracted from phone/email conversations (tags for pain points, current hosting provider, industry vertical, etc). I have a loose idea of how this all needs to work but was wondering if anyone has developed something similar and would be willing to compare notes on how you approached it


Wow, this is a great use case, and we have discussed this internally for feature release cycles, and would love to bounce some ideas back and forth.

I would approach like this:

Create a “persona-feature” matrix. Personas in the rows, and features in the columns. I would go through each persona and assign a unique number to the features based on what matters the most, like a sudoku.

Each persona would essentially get its own automation, and each feature adoption would have a goal (probably using Event Tracking). So you end up with a prioritized “todo” list for each persona. If a user has already completed a goal (adopted a feature) move them on to the next stage in the master “index” sequence. If they complete the feature adoption process mid sequence, jump them along to the next “todo.”

As you continue to roll features out internally, slot them in to each existing sequences. Even if the new feature appears “early” in the sequence for a persona, the next time they hit the “loop” they will be checked for adoption and fail (since it is a new feature) and enter into the new sequence.

Is this kinda like what you are thinking?


Jordan, thx. The approach you’re suggesting is very straightforward but various reasons not what we need. Problems with that approach:

  1. It’s inherently 2-dimensional (persona & feature). We need to be able to do multi-dimensional touches and have a gatekeeper function decide which content is most optimal given all the variables

  2. It violates the DRY (do not repeat yourself) principle. We need something that’s modular so you can reuse messages across personas and have them reside in one place and be invoked based on different criteria

  3. If I’m understanding it right it still assumes a forward flow through a set of goals. It’s non-linear in the sense that you can skip forward over a goal but it’s not the “Orbit” I’m referring to in the sense that you can hop around within the

What I have in mind is more like a slot machine that gets pulled at pre-defined intervals. Content is chosen at runtime based on a number of variables (persona, aspirations, objections, industry vertical, current provider, challenges, etc). It performs essentially the function of a good sales person checking in on a prospect and delivering what they need to hear based on past interactions, their site & email interaction in the interim since last conversation, stage in the sales cycle… I’m seeking to basically extract what I do into a sort of AI.

At this point I know what I want it to do and I’m more wrestling with how to implement it. I’m not entirely sure if I can do this by breaking it into sub automations, using goals and conditional logic and doing it all within AC or if it’s easier to write this logic as an external webservice that I invoke from within AC. Like a good sales person it should be able to accelerate or decelerate the frequency of touches based on how hot the prospect is (ie if they want to ascend faster through the cycle we compress the timeline intelligently).

Anyways, I’ve been poking around to see if anyone had something in the marketplace that accomplished this or had accomplished it another way. It sounds like you guys have a similar need but were thinking down a different path than I am. The engine I’m envisioning is very flexible. I’ll update this thread as I progress in building it and share any lessons from the process. Please others do the same if you make advancements towards this vision in a different way.


I’ve created my own system for implementing ‘The Machine’ on ActiveCampaign. I use different controller automations, such as Funnel Controller, Segmentation Controller and Value Loop…along with custom fields for Last Lead Magnet Requested and Last Product Purchased. This “Orbit” approach kinda builds on ‘The Machine’ by adding additional conditions to which Engagement/Ascension series is presented next. It’s a ton of logic to implement, but would be fun for sure.


@activeautomations presumably you’re referring to Ryan Deiss’ framework? ->
If so, yea would be interested in comparing notes on how you approached things. I’m basically envisioning this as if all the potential emails I could send someone are like a huge deck of playing cards. I’ll have pre-ordained “slots” at which Cards (emails) will be pulled at runtime based on a number of variables (vertical industry, role of the recipient, phase in the sales cycle, hot buttons, etc). The cadence can be accelerated/decelerated based on how hot to trot the recipient is per his/her interactions w/ the emails & site.

I haven’t investigated “The Machine” yet. It’s mildly annoying their LP is just a “sorry we’re at capacity” message on what is undoubtedly something that could be taught via self-guided eCourse. I’ll poke around and see if any of the public videos about it have any worthwhile lessons. I’m still trying to determine whether the logic for the email selector is something best done within AC or if this needs to be an external webservice invoked by webhook. I’m assuming the system you developed resides all within AC?


So I finally built this. It’s taken about 80hrs all told from concept to completion and blossomed in complexity but we now have essentially an AI for Pagely that takes knowledge of the prospect from past interaction, factors in site activity and interaction with email behavior and runs perpetually in the background dripping them the most optimal content based on what we know about him/her. Hat tip to Ed Haskin’s course on how to implement Ryan Deis’ “The Machine.” That gave me critical epiphanies on how to do the logic. It’s basically a set of 5 automations that invoke one another running in a loop and create a “plinko board” that filters the contact down the optimal path. It then invokes a webservice that queries the right messages from a Google Sheet and returns the appropriate piece of content, stuffs it in an email and emails the prospect. For now it’s proprietary IP for Pagely but if they let me share the logic of how it all works I’ll post here at some point.


Sounds awesome. Please do post about it when you can. I’d be interested to see how it all works.