Customer Segmentation Strategy Beyond ARR

The first question in a customer segmentation exercise usually determines everything that follows.

Over the years, I have seen many segmentation exercises begin with some version of the same questions:

• Where should the ARR thresholds be?
• How many accounts should each CSM manage?
• Which customers should receive dedicated, pooled, or digital engagement?
• How many segments can the organisation realistically support?

These are important questions. They influence hiring plans, coverage models, cost-to-serve, capacity and the overall economics of Customer Success.

The problem is not that companies ask them. The problem is that they ask them first.

When segmentation begins with staffing ratios and revenue thresholds, the organisation starts optimising its operating model before agreeing on what the segmentation is actually supposed to achieve. The result may be operationally convenient, but it is not necessarily aligned with how customers succeed.

Segmentation Is Not the Same as Tiering

Most companies use the words segmentation and tiering interchangeably, but they are not quite the same thing.

Tiering divides customers based on the level of investment the company intends to make in them. Segmentation should explain why different groups of customers need to be served differently in the first place.

The purpose of customer segmentation is not simply to divide a customer base into enterprise, mid-market and SMB buckets. It is to identify groups of customers that have meaningfully different needs, behaviours and paths to value, so that each group can be supported through the right engagement model.

A useful segmentation strategy should help answer questions such as:

• What makes one group of customers meaningfully different from another?
• What outcomes is each group trying to achieve?
• Which challenges or constraints are unique to that group?
• What must happen for customers in that segment to realise value?
• Where would additional investment produce disproportionately better outcomes?

Only after these questions are answered should the company decide how many CSMs it needs, how accounts should be distributed and which activities should be delivered through people, programmes or technology.

The Operating Model Should Follow the Customer Model

Coverage models, staffing ratios, playbooks and technology are all downstream decisions.

Tiering answers how much attention a customer gets. Segmentation answers what kind of attention works.

They should be designed around an understanding of customer needs, rather than becoming the basis for defining those needs.

Consider two customers paying the same amount. One may have a simple implementation, a strong internal champion and a repeatable use case. The other may operate across several business units, depend on multiple integrations and require significant organisational change before value can be realised.

From an ARR perspective, they belong in the same segment. From a customer success perspective, they may require entirely different engagement models.

The reverse can also be true. Two customers with very different contract values may share the same use case, operating complexity and path to value. Treating them as fundamentally different simply because one pays more can create unnecessary variation in how the company serves them.

Good segmentation identifies these differences before assigning a service model to them.

Why ARR Is an Incomplete Foundation

ARR is an important commercial metric. It tells the company what a customer is worth today and helps ensure that the cost of serving the customer remains economically sensible.

But ARR says very little about why the customer will succeed.

It does not necessarily tell you:

• How complex the customer’s environment is
• How difficult the product is to implement or adopt
• How many stakeholders must be aligned
• How mature the customer is in the problem being solved
• How much behavioural or operational change is required
• Whether the customer has significant expansion potential
• Whether the current contract represents a small or large share of the addressable opportunity

A high-ARR customer may already be close to its natural ceiling. A smaller customer may have the potential to become one of the company’s largest accounts over time.

Similarly, a high-value customer may require less support because it has mature internal teams and well-defined processes. A smaller customer may need significantly more guidance to reach the same level of adoption.

None of this makes ARR irrelevant. It means ARR should be one input into segmentation, rather than the entire foundation.

Current Value and Future Potential Are Different

One of the most important distinctions in segmentation is the difference between what a customer is worth today and what the relationship could become.

Many segmentation models are heavily weighted towards current revenue because it is visible, measurable and readily available in the CRM. Future potential is harder to estimate. It requires understanding the customer’s organisation, business model, use cases, product footprint and strategic priorities.

But if the objective is long-term growth, segmentation cannot focus only on protecting existing revenue. It must also identify where additional investment is likely to create disproportionate future returns.

That potential might be driven by:

• The number of business units or teams that could adopt the product
• Transaction or consumption growth
• Geographic expansion
• Additional products or use cases
• The customer’s own growth trajectory
• Strategic influence within an industry or ecosystem
• The gap between current adoption and the total addressable opportunity

A customer with modest ARR but high potential may warrant more attention than a larger account with limited room to grow. A segmentation model based only on current value will often miss that distinction.

What Should Segmentation Be Built Around?

There is no single segmentation variable that works for every business.

For one company, employee count may be a reasonable proxy for organisational complexity. For another, transaction volume may be far more meaningful. Industry, product maturity, operational complexity, business model or number of use cases may explain customer needs better than company size ever could.

The right characteristic is the one that most reliably predicts how customers realise value and what support they require along the way.

In practice, strong segmentation strategies often combine three dimensions:

1. Customer Characteristics

These describe what the customer is and how it operates. They may include company size, industry, geography, business model, organisational maturity or regulatory complexity.

2. Path to Value

These explain what must happen for the customer to succeed. They may include implementation effort, integration requirements, number of stakeholders, time to value, change-management needs or breadth of adoption.

3. Commercial Potential

These describe both the current economic value and the future opportunity. They may include ARR, expansion headroom, product whitespace, consumption growth or strategic importance.

The objective is not to create the most sophisticated model possible. Complexity for its own sake does not make segmentation better.

The objective is to identify the smallest number of customer groups that are genuinely different enough to require different engagement models.

A Better Sequence for Segmentation

A more effective segmentation exercise follows a different sequence.

First, identify the customer groups that have distinct needs, constraints and paths to value. Then determine what outcomes the company wants to create for each group and which interventions are most likely to improve those outcomes.

Only after that should the company design:

• Coverage models
• CSM capacity and account ratios
• High-touch, pooled or digital engagement
• Onboarding and adoption programmes
• Success playbooks
• Escalation paths
• Technology and automation
• Measures of success

This sequence matters because it prevents the operating model from defining the customer model.

The engagement model should be a response to customer differences, not an internal organisational preference imposed on them.

Start Without ARR

One of the most useful exercises is to temporarily remove ARR from the discussion.

Imagine that you had to build your segmentation strategy without knowing what any customer paid. What characteristic would you use instead?

Would it be employee count, company revenue, transaction volume, business maturity, operational complexity, industry, product footprint or something else?

More importantly, why would that characteristic matter?

The answer forces the organisation to articulate what actually makes customers different, how those differences affect their ability to realise value and where additional investment is most likely to change the outcome.

ARR can then be brought back into the model to ensure commercial discipline. But by that point, it is informing the engagement strategy rather than defining it.

That is the difference between dividing customers into tiers and building a segmentation strategy that can genuinely improve retention, expansion and long-term growth.