Audience Segmentation: How to Segment SaaS Customers Into LTV Cohorts

To keep your SaaS business profitable, you need to continuously monitor the relationship between your customer lifetime value (LTV) to your customer acquisition cost (CAC). Ultimately, the lifetime value of a customer should be three times more than the cost of acquiring them.

So, how do you improve LTV:CAC ratios after arriving at an average LTV for all your customers?

For the majority of SaaS businesses, it all comes down to audience segmentation.

Let’s take a closer look at how to get started.

Segmenting Customers Into Cohorts: Where to Start

To start getting more granular LTV data, you will need to break the total pool of customers into cohorts.Each user in a cohort must share a distinguishable trait that separates them from the other cohorts, such as a particular demographic, statistical, or product usage characteristic.You could consider factors such as the user’s proficiency – from the beginner to intermediate to advanced. Or the country where the user is based, the product seats he’s using, or how actively he’s using the product.When tackling cohort audience segmentation, remember to first consider the creative implications and how that could help you do a better job with value-based bidding.

Think About the Creative Implications of Audience Segmentation

Once you separate your trial users into cohorts with different lifetime values, you can really lean into remarketing campaigns to move them to paying customers.

Done well, remarketing can accelerate the buyer’s journey and help leads move through its different stages as seamlessly as possible. Effectively, to get the most out of your remarketing campaigns, you need to think about what different customer segments you can have different creative for.

For example, you could consider the user’s degree of proficiency and tailor your ad message to the specific cohort to help them use your product more efficiently. A hobbyist using your SaaS solution would have different pain points and goals than someone using it professionally.You could also consider the country where the user is based. This would allow you to develop content in the local language and local references.

Look For a Wide Distribution of Value Within Each Cohort

Value-based bidding, or target ROAS, is quickly gaining traction within the SaaS industry. Because target ROAS lets you assign unique values to different types of conversion, Google’s algorithm can take a portfolio bidding approach across an entire group of campaigns.As you get started with value-based bidding, remember to look for the dimensions among your users where data is most polarized, or widely dispersed. These could be geography, level of experience, industry, or role in the organization, for example.The wider the range in expected value (from the low point to the high point in your chosen cohort dimension), the better you'll be able to manage ROI for that cohort.Why? Because value-based bidding for the acquisition of those users will have a clearer separation of values and will be able to bid more effectively and deliver more precise ROAS.Consider this: the lifetime value of a beginner versus the lifetime value of an advanced user. As they are likely highly polarized, putting different LTV or total value against those different cohorts would allow value-based bidding to ultimately have a greater impact.

Get Started With Remarketing to Cohorts

Segmenting customers into LTV cohorts offers valuable context, helps you determine which behaviors to target, and enables you to tailor your messaging more closely to the buying cycle.

Executed well, audience segmentation can yield a few strong cohorts with high-value customers, giving you plenty of opportunities for ongoing engagement.

On a roll? Check out our blog and discover the most common remarketing mistakes SaaS businesses make (and how to avoid them).

Delcho Stanimirov

Head of Paid Media

I lead the PPC and Analytics teams. My professional goals are happy clients and colleagues. Also I love numbers, charts, and data-driven decisions.