Understanding cost drivers in the abstract is useful. Having a repeatable process to surface them is what actually changes decisions. Learn how to identify the right cost drivers in your business to calculate margin profitability at the customer and product level.
Your gross margin looks fine. It almost always does — right up until the moment it doesn’t.
Most SaaS companies discover a profitability problem at the worst possible time: during board prep, a fundraise, or a renewal that goes sideways. The business looked healthy. The numbers looked acceptable. And then someone started asking the right questions.
That’s how profit leakage works. It doesn’t announce itself. It hides inside blended averages, accumulates across customers, and compounds quietly until it’s expensive to fix.
The companies that catch it early have one thing in common: they know their cost drivers. Not just their cost categories — the specific activities, behaviors, and structural factors that determine what it actually costs to serve each customer. That clarity changes how they price, how they manage renewals, and how they make every decision that touches margin.
What Are Cost Drivers in SaaS?
A cost driver is any factor that causes costs to increase. In software businesses, cost drivers are the specific activities and behaviors that determine what it actually costs to deliver your product to a given customer.
The reason they matter is simple: not all customers cost the same to serve. Not all product features carry the same margin. Revenue and profitability are different things — and the gap between them is where profit leakage lives.
When finance teams understand their cost drivers, they can connect those drivers to specific customers and segments. Pricing decisions get grounded in real data. Customer acquisition strategy gets smarter. And the conversation about whether to renew a complex enterprise account gets a lot more honest.
The Hidden Cost Problem Behind SaaS Profit Leakage
A SaaS business running at 74% gross margin feels healthy. Benchmarks put the target range at 70–80% for well-run software companies. No alarm bells.
But inside that 74% might be an enterprise segment at 51% and a mid-market segment at 82%. The average looks fine. The enterprise business does not.
This pattern is more common than most finance leaders realize — and it surfaces most often in companies that have added product complexity, mixed customer sizes, or new AI capabilities without updating how they track costs.
What gets hidden inside a blended margin is telling: the enterprise account that opens 40 support tickets a month. The implementation-heavy deal that consumed six weeks of professional services before generating a dollar of recurring revenue. The customer running AI-powered workflows at 10x the volume of comparable accounts at the same contract value.
None of it shows up as a problem on a company-wide P&L. Revenue is revenue. Costs get pooled. The margin looks acceptable. And the leakage continues.
The Most Common SaaS Cost Drivers
Most SaaS profit leakage tends to come from a small number of recurring cost drivers. Here’s where it most commonly shows up:
| Cost Driver | Common Signal | Why It’s Underestimated |
| Cloud & Infrastructure | High compute, storage, or data transfer | Tracked as a single aggregate invoice |
| Customer Support & CS | High ticket volume, complex escalations | Reported as headcount budget, not per-account |
| AI Feature Usage | Heavy inference, frequent AI-powered workflows | Buried in infrastructure; unfamiliar pricing model |
| Product Complexity | Custom integrations, unique workflows | Maintenance burden not tracked against contract value |
| Third-Party APIs | High transaction or API call volume | Multiple vendor invoices, rarely linked to customer |
| Implementation & Onboarding | Long ramp, data migration, training | Treated as GTM cost, not customer profitability factor |
Infrastructure and Cloud Usage
Cloud costs — compute, databases, storage, data transfer — typically represent 20–30% of SaaS revenue. They’re also the most usage-variable costs on the P&L.
Most companies track cloud spend in aggregate. They know what they’re paying AWS or GCP at a company level. They don’t know what they’re paying per customer. A handful of accounts running batch jobs or storing years of data are driving meaningfully higher costs than light users at the same contract value — and without per-customer visibility, that difference is invisible.
Customer Support and Customer Success
Support in North America runs $15 or more per ticket once fully loaded. A customer generating 30 tickets a month costs $450 in support labor alone — before CSM time, escalations, or downstream engineering hours.
Because support is reported as a headcount and tooling budget rather than allocated by account, a high-touch customer at $40,000 ARR can quietly run at a loss while appearing profitable at a blended level.
AI Compute and Feature Usage
AI features carry per-call inference costs that scale directly with usage — and usage varies enormously. A heavy AI user can generate three to five times the compute cost of a comparable account that doesn’t. As covered in our piece on how AI features are compressing SaaS customer margins, the gap between what a customer pays for AI functionality and what it actually costs to deliver is often significant, and it compounds as adoption grows.
Product Complexity and Customizations
Every custom integration, unique workflow, or feature built to win a specific deal creates ongoing maintenance obligations. Multiplied across a product surface that expands year over year, complexity becomes a meaningful cost driver — especially for enterprise accounts that tend to accumulate the most of it.
Third-Party Vendor and API Costs
Payment processing, messaging APIs, data enrichment, identity verification — these carry per-transaction pricing that scales with usage. They’re easy to undercount because they arrive as separate invoices. With the right allocation methodology, as detailed in our piece on how to allocate costs across customers and products in software companies, they’re actually highly traceable.
Implementation and Onboarding
Implementation-heavy deals can require weeks of professional services time before generating any recurring revenue. When those costs aren’t modeled against account economics — and they rarely are, because they’re treated as a sales expense — companies can’t distinguish which customer types justify the investment and which don’t.
A Framework for Identifying Your Cost Drivers
Understanding cost drivers in the abstract is useful. Having a repeatable process to surface them is what actually changes decisions.
Step 1 — Start With Your Largest Cost Buckets
Rank your cost categories by total spend: infrastructure, support, CS headcount, vendor and API costs, implementation. Your top three are where the most material insight lives. Start there.
Step 2 — Match Costs to Customer Behaviors
For each major cost category, identify the behavioral driver behind it. Cloud tracks against usage and storage. Support tracks against ticket volume. AI tracks against inference calls. The question for each: which customers are generating the most activity?
Step 3 — Look for Cost Concentration
A small number of customers usually drives a disproportionate share of costs. A customer representing 8% of ARR but 20% of support tickets is a conversation worth having before the next renewal.
Step 4 — Compare Revenue vs. Cost-to-Serve
Set allocated costs against ARR to see actual gross margin by account. The distribution is almost always more skewed than the blended number suggests. Accounts below your margin threshold deserve a conversation. Accounts above it deserve to be understood and replicated.
Step 5 — Review Quarterly
Costs shift. Products evolve. Usage changes. Build a quarterly review into your operating cadence — annual is too slow to catch the drift that causes the most damage.
Three Mistakes Companies Make
Looking Only at Blended Margins
Blended gross margin is a useful summary. It is not a diagnostic tool. A company running at 73% gross margin can have individual segments running below 50%. The average obscures both the problem and the opportunity — and decisions about pricing, CS coverage, and customer acquisition get made on the wrong data as a result.
Assuming High-Revenue Customers Are Most Profitable
Enterprise accounts often carry the highest ARR and some of the worst margins. High-touch support, custom development, complex implementations, and intensive usage can push cost-to-serve well above what the contract economics justify. Companies that don’t know this tend to prioritize retention investment on the accounts that are quietly destroying margin at scale.
Treating Cost Drivers as Static
AI features get adopted at rates finance didn’t model. A new product line creates support volume existing resources weren’t sized for. A vendor changes pricing. An account triples its usage. Companies that review cost drivers once end up with models that are accurate at a point in time — and misleading going forward.
What You Can Do With This Visibility
Cost driver visibility doesn’t just tell you where money is going — it sharpens every decision that touches margin.
- On pricing: when you know what it costs to serve a customer at different usage levels, usage-based tiers, AI add-ons, and support packages can be priced on actual economics rather than competitive instinct.
- On renewals: a renewal conversation looks different when you know what the account actually costs to serve. Finance teams with this visibility can support CSMs with real data — and flag accounts where the renewal needs to be structured differently.
- On customer acquisition: when you know which customer profiles produce the strongest margins, you can make deliberate decisions about where to focus sales and marketing investment.
- On margin erosion: when you can see which accounts, features, or behaviors are compressing margin, you can address them before they compound. A proactive repricing conversation is significantly less painful than discovering a segment has been running below break-even for two years.
The Bottom Line
Profit leakage in SaaS is gradual, structural, and almost always hidden inside averages. A blended margin that looks acceptable is often masking real erosion at the segment or customer level — and the longer it stays hidden, the more expensive it becomes to fix.
The companies that manage this well aren’t doing anything complicated. They know their largest cost categories. They’ve matched those costs to the customer behaviors that drive them. And they review the picture regularly.
That discipline is what converts a company-level P&L into an actual management tool.
For teams trying to operationalize this level of visibility, platforms like Cogs’z can help automate customer-level cost allocation and profitability analysis that becomes difficult to manage manually.
Brad Perry is the CEO of Cogs’z, a profitability management platform built for B2B SaaS companies. Brad co-founded DealerSocket, an end-to-end platform in the automotive industry, where he experienced firsthand the margin challenges that Cogs’z is designed to solve. Cogs’z automates customer-level cost allocation so finance, CS, and sales teams share a single, accurate, view of who’s profitable and why. Learn more or request a demo at cogsz.com.
References
- SaaS Capital — Annual SaaS Gross Margin Benchmarks
- OpenAI / Anthropic — LLM API Pricing Documentation
- Gartner — SaaS Cost Management and Cloud Spend Trends