Tl;dr

The lifecycle marketing problem is finding the best journey for each user. Its an iterative process. Teams are constrained by bandwidth - data, content, ops, strategy. AI changes this. Sortment changes this.

As they say, “Show, don’t tell”

Example: See Sortment’s Marketing Ops AI Rho creating a custom attribute

In the current world, this would take a few weeks at the least:

  • figuring out if its possible
  • raising ticket to engg
  • getting them to prioritise, make a business case etc
  • waiting for a couple of sprints

Now it takes 20 seconds.

Defining the “lifecycle marketing” problem

Every business would want each of their customer to come to them everyday, become “daily active” in  industry terminology. At least, show up when they need to buy the business’s services, and not go to any competitor. It's obviously better for the business if this happens more frequently.

Unfortunately, very few products have the ability to become a daily habit for their users. Think a coffee shop which people feel is “needed” to start their day.

For everyone else, they need to

  • create intent
  • nurture intent
  • capture intent to turn into $$

However, every business is competing with Mn others for very limited attention of their users. More importantly, competing with everything else that the user could be thinking about. There’s so much worth thinking about.

Unlikely, they’re thinking of your business & its services, unprompted on a daily basis.

This is the core problem of lifecycle marketing

The opportunity comes from the fact that humans are fundamentally social. Hence, We have anointed platforms that enable social interactions as places where we become “daily active”. These are

  • Social media platforms: Think Instagram, TikTok, LinkedIn, X
  • Messaging platforms: Email, SMS, Push notifications, etc

Naturally, every business realised presence on these is crucial to create, nurture, and capture intent. But everyone knows this. Its competitive. Competition for user’s scare time and attention. And that makes this a problem worth solving. This is a matching problem of user’s current / future needs to a business's services.

Clearly, it is one of the most valuable problems to solve.

Social media platforms collectively have a revenue of $500 Bn and market cap of $3-5 Tn

  • Highly concentrated between a few platforms, order of 10-20
  • Walled gardens
  • They do the matching between users and businesses to make money via their ad engines
  • Use large amounts of data to find patterns at scale, sometimes even know us better than we think.

Marketing technology platforms collectively have a revenue of $100-300 Bn and market cap of $1 Tn

  • Fragmented, order of 10,000+ (one of largest categories in Software)
  • Business is responsible for solving the matching problem
  • Data is first party and limited (compared to social media platforms scale of data)

Sortment falls in the marketing technology space.

Being 1 in 10,000 might seem scary to some, but to us it seems very exciting. Clearly it is a large problem that so many companies have been created here. Its likely it has a lot of nuances which allow for so many companies to survive offering some differentiation from each other.

Problem definition: Finding “the best journey for each user”

The lifecycle marketing problem can be defined as an optimization exercise with

Goal: Find a sequence of actions (messages, push, ad placements, offers) for each user that maximises revenue from them, reacting to their dynamic actions & preferences

Constraints

Large companies solve this by throwing “AI at the problem”

The approach to solving this problem differs by scale of the business. A larger company like Home Depot or Amazon will solve this by throwing money & the brightest minds after it. It would leverage the scale of data it has. It would then hire the best minds in Machine Learning to train custom models to answer the question - “At this point in time, for each user, what is the next best action to max revenue from them”

Likely, most businesses in the world don’t have the resources for this.

And even if they did, they don’t have the data for it.

Everyone else uses human data-backed intuition & A/B experimentation

For companies that are not Amazon, the typical approach is

  • Analyze user-behavior data, think conversion funnels, drill-downs, path analysis
  • Come up with hypotheses on “combination of right users + right message + right timing” that would create, nurture, or capture intent
  • See if the data is available to actually execute this in their marketing tool
    • If not, check if it is even available in the organization
      • If yes, “convince” engineering to start sending that data to their marketing tool
      • If not, “convince” them to start capturing it
      • This convincing is not super easy for businesses where engineering bandwidth is very limited
  • Setup this experiment / workflow in the marketing tool
    • Create content
    • Setup the right triggers, audiences
    • Check everything is running fine
  • Once the experiment is running, measure the results until it reaches statistical significance
  • If it works, make it live future users as an automation
  • Rinse & Repeat

^This is not scalable.

One can think of this as a truth seeking mission. Peeling layers of consumer behaviour onion wrt of the business’s products & needs one by one. Its a super iterative process.

To effectively do this, an organisation needs to constantly need to answer questions like

  • What data should we capture about our users?
  • How should this be captured? How do we centralise it in one place?
  • How does one turn this raw tables of data into actionable for users?
  • How to continuously generate high quality hypotheses?
  • How do I allocate limited resources to discover the best “journey for each user”  

The resource constraint and hence allocation is the central problem here

You can only

  • come up with so many hypotheses
  • get engineering team to spend so much bandwidth in making required data available to you
  • create so much content
  • setup so many campaigns or journeys
  • run so many experiments

This directly affects how personalised your campaigns are (reflected by how granular your segments / logics are). Naturally meaning, how close are you to the “best journey for each user” dream.

AI has changed this resource allocation problem in 2025

Its likely someone reading this is using ChatGPT (or equivalent) daily (or at least weekly). Its super human intelligence at everyone’s fingertips. Its not perfect but If someone told me in 2020 the world in 2025 will look like this, I wouldn’t believe them.

In its first avatar, it was generating things - text, images, music, video

In lifecycle marketing, that meant it could help in reducing the time to create content. Over time, it is only getting better at this

Solving (read attempting to solve) one of the constraints of lifecycle marketing among

  • Hypothesis generation
  • Content creation
  • Data accessibility
  • Ops bandwidth

So far, the other constraints were unsolved because

  • They require a lot more context about the business
  • Need access to internal first party data
  • Text output is not enough, they need action like setting up campaigns, journeys
  • Accuracy is crucial

Enter Sortment:

AI Agents for Lifecycle marketing

You might have recently seen AI has gone beyond a chat bot. It can now plan a full trip: book flight tickets, plan an itinerary, book stay etc. Basically instead of telling what to do, it can actually do those things.

This development has huge ramifications on the Lifecycle marketing problem

Instead of humans relying on other humans, they can use these agents to

  • suggest hypotheses
  • crunch data for it
  • create content
  • setup experiments

This would mean shortening the time between each iteration from months to days.

The reduced cycle time to test a new hypothesis simply means → more experiments → getting closer to the “best journey for each customer”. A good way to think about it is you have 2x / 3x the bandwidth you have today on everything operational. Pushing the bottleneck to the imagination of the Lifecycle marketing strategist.

How do Sortment AI agents work?

Unlike other tools which operate on a silo, Sortment is connected to companies' sources of truth, generally sitting in a data lake or a warehouse. It taps into real time events for topical information. It has its own marketing execution layer, think ability to cut users into audience, create metrics, pass this data to other tools (like Braze, iterable), and functionality to run campaigns & journeys across all channels on its own.

The AI gets access to the data & the execution layer to become a 24x7 co-worker with unlimited bandwidth.

Sortment's architecture (simplified)

Shaping the future of Lifecycle

Jury is still out on what is hype vs reality. One thing is clear that AI is disrupting everything. We see this vision for 2030 for Lifecycle marketing

  • Smaller teams consisting of thinkers who are super close to the consumer & the product
  • They complement AI with empathy and context
  • They make strategic decisions and shape the direction of work for AI to execute
  • They iterate towards the goal of “best journey for each user” 10x faster than ever possible before
  • AI takes on 90% of operational work, with humans acting as reviewers of their work
  • Consumers get fewer but more relevant and valuable information
  • There’s lesser spam

We’re excited about this future. We’re going to help make it happen.

If you’re looking to understand / brainstorm implications of AI on your Lifecycle strategy, book time with us.