At the foundational level, growth marketing is about building a system of inputs and outputs that predictably lead to revenue and user growth. And like any system, there are a number of steps that need to be followed in the right order to ensure success.

So what does this have to do with AI?

As we’ll explore in this post, AI is going to have an epic impact on a number of the most important steps in the growth marketing formula. From identifying patterns in massive datasets, to accelerating the creative process, AI is going to reshape how many companies in the near future will approach growth marketing.

Many of the concepts we’ll be discussing in this post come from a live Founder Fireside between Mutiny’s CEO Jaleh Rezaei and Y Combinator’s Managing Director Anu Hariharan.

Listen to the full conversation: Spotify, Apple Podcasts, Twitter Spaces

The current growth process

Here’s what a typical growth process would look like at a modern software company:

It starts with connecting data. Data is the backbone of the entire growth process because it allows you to aim your technical resources at the right problems to solve. Without a large amount of real-time data showing the health of your entire customer journey, it would be nearly impossible to know where the opportunities for improvement are.

Next, you need to identify what improvements can be made and what the revenue impact will be. This requires either a lot of experience to predict what will work, or a mature data model based on historical conversion metrics.

Then comes brainstorming and ideating about what the content and creative should actually be. The best companies are able to inform their creative decisions using proprietary data from each of their customer segments. Netflix famously does this by creating dozens of thumbnails for a piece of content based on engagement metrics.

Now that a project is scoped, it will then be shortlisted. Engineers are expensive, so in order for a growth project to actually get built, projects are prioritized on three things:

1. Time to build

2. Confidence is success

3. Revenue impact

Only once a project is deemed “high-impact” is it actually given to a growth engineer to bring it to life. Depending on the size of the project, this may take anywhere from a day to a few weeks to implement live and begin collecting new baseline data.

Finally, once a change has been made it must run for long enough and be seen by enough users for it to be deemed statistically significant or not.

From insight to launched experiment, the typical growth process can take between 4-10 weeks.

Now, Mutiny makes it possible to collapse that timeline down to a few minutes using a no-code tool that allows marketing teams to unlock their growth with the help of AI.

Solving the art and science of conversion

To truly understand how AI is going to impact the next era of growth marketing, let’s look at each of the 5 key steps outlined above individually and extrapolate how AI is going to change each of them.

1. Identifying areas of improvement from a mountain of company data. It’s not hard to imagine how AI is going to change data analysis because it’s already happening. AI is relentless at identifying patterns in massive datasets that humans would never be able to see. The resulting outputs are recommended conversion opportunities based on positive buying signals. Examples include new audience segments to target, under-performing industries, or target accounts with real-time purchasing intent.

2. Choosing high-impact changes to be made to the existing marketing asset. Ok, so you know a landing page isn’t converting as well as you’d hope. But what specifically about the landing page isn’t working?

   Identifying the root cause of a conversion issue is more art than science, but it often still comes down to a few key elements on a page. But as AI continues to become better at identify patterns, it’s going to become increasingly better at making recommendations about what to change and why.

3. Brainstorming new content and creative to get the desired result. This is where things start to get really interesting. AI is already being used by creatives to generate entire written pieces of content, and the results are starting to be really impressive. Pictures and video are on the horizon, hinting at the not so far future where creative is no longer the bottleneck for getting collateral launched in-market.

4. Prioritization based on revenue impact and confidence level. As AI gets better at identifying patterns, it’s going to become really good at predicting which changes are going to have the biggest impact on conversion rates driving revenue. It will also make use of benchmark data from previous experiences to predict whether an experiment will be successful.

5. Implementation of new customer experiences. The act of building these customer experiences has always been a huge hurdle to overcome because engineering resources are so valuable to a software company. This means that marketers have had to beg to borrow engineering time to implement their experiments. But with no-code tooling like Mutiny, marketers are going to be able to build their own growth experiments without dependencies on engineers to build it.

The evolution of Mutiny’s AI

So how are we using AI to improve Mutiny? CEO and cofounder Jaleh Rezaei explains how she’s been thinking about the power of using AI within Mutiny:

We’ve always approached the use of AI in terms of how it makes the life easier for Mutiny end-users easier, not just for the sake of using a flashy new technology. This means we use AI to guide and assist growth teams to make high impact decisions.

Currently, Mutiny has one of the largest engineering teams with production experience in marketing AI, and still growing.

But this was not always the case, especially in the early days. Here’s a quick timeline of how Mutiny’s AI has evolved.

V1: Humans as the AI

In the beginning, we didn’t yet have the proper dataset to train an AI and get meaningful guidance. So we started by guiding customers based on previous experiences and how others were seeing success.

But even at this stage, we knew that we wanted our users to have access to standardized benchmarks, suggested audiences, and a creative assistant for writing website copy. To accomplish this, we created standardization across the platform so all anonymized data is comparable.

V2: Community-sourced playbooks

This led us to launching community playbooks. When a user has a winning experience, we’d ask them if they wanted to share it with the rest of the user community. With their permission (we had near 100% acceptance rate) we’d extract what they did to drive more conversions and the metadata about the audience segment and conversion rate.

The product then began to recommend that playbook to other users who serve the same segment, always solving the problem: how do we make it easier, faster, and more high-impact for our customers. 

This increase in product usage and velocity allowed us to continue to build the dataset. As the database grew, we were able to identify anomalies in their audience and suggest how to convert them. 

This began to give our model the necessary inputs to predict how a change on their site could increase conversion lift. We also began to build capacity in making suggestions to specific parts of the webpage a user could improve.

The rapid acceleration of the natural language AI model GTP-3 allowed us to introduce a writing assistant that provides high-converting website copy based on your desired value proposition and page type.

Mutiny even identifies under-performing segments of your audience and predicts the impact you’ll have on your revenue by launching an experiment.

All these AI capabilities add up to a platform that allows growth teams to:

• Identify the highest impact experiments to be running.

• Setup the experiment without engineering dependencies in a fraction of the time.

• Scale their experimentations across new audiences, customer segments, and pages.

But what does that look like in practice?

When Jimmy Flores was at ClickUp, he was able to increase the testing velocity of the experimentation program by 850%, allowing him to launch over 1,600 experiences in 18 months. This led to a step-function increase in the number of conversions the website was creating.

The future of growth marketing with AI

I’m not able to predict the future (if I was, I’d be a billionaire!), but what I’m seeing is a truly remarkable piece of technology now becoming accessible for anyone to use.

I’ve heard stories of parents playing with their kids by giving DALL-E fun prompts and seeing what the result it. I used a writing assistant to help with parts of this post. Everyday I’m seeing a new use cases on Twitter of how people are using AI to generate videos, lookbooks, and audio.

This is a totally new canvas people are going to use to express themselves creatively.

As new AI technologies and capabilities develop we’re going to continue to build Mutiny with our end-user in mind by solving both the art (strategy) and the science (tactics) of conversion.

If you want to be a part of building the growth team for every company, apply today.