The Data-Driven Path to Durable EBITDA Improvement

When the board asks for earnings before interest, taxes, depreciation and amortization (EBITDA) improvement, the first conversation is almost always about headcount. Sales efficiency ratio review. Marketing spend as a percentage of revenue. G&A rationalization. These are real levers. They're also the harder levers to pull without damaging the growth trajectory, because the same costs that look inefficient from a finance perspective are often the investments that generate the expansion revenue that shows up 12 months later.

The commercial EBITDA lever is faster, larger, and fully under the company's control. It doesn't require layoffs. It requires looking at what the transaction data says about where realized revenue is leaking.

Where the Guessing Happens

EBITDA improvement discussions typically start from the expense side of the P&L and work backward to identify waste. That's a legitimate process. It's also an incomplete one when the revenue side of the P&L has its own efficiency problem.

The guessing is that revenue quality is fixed by the commercial model and can only be improved through growth. That belief is wrong. Price realization, net revenue retention (NRR), and cost-to-serve are all variables that respond to commercial architecture decisions, and improving them generates EBITDA improvement faster than almost any cost reduction initiative.

The data to diagnose the commercial EBITDA opportunity is already in your CRM, billing system, and product analytics. Most companies have never systematically mined these three sources for their EBITDA impact. They look at cost lines. They miss the revenue quality story.

What the Transaction Data Actually Reveals

Price realization data shows the gap between your list price and what you actually collect after all discounts, concessions, and adjustments. In most B2B SaaS companies, this gap is 18% to 30%. Closing it by even five points produces direct EBITDA improvement at full contribution margin. There's no cost of goods change. The revenue simply realizes at a higher rate.

In a $70M annual recurring revenue (ARR) business with a 22% pocket price gap, a five-point improvement in price realization generates $3.5M in additional revenue at approximately 85% contribution margin. That's nearly $3M in additional EBITDA from a pricing governance change that takes 60 to 90 days to implement.

Cost-to-serve data tells the second half of the story. Not all customers cost the same to serve. Customers who were heavily discounted at close, who have complex custom configurations, or who are outside your true ideal customer profile (ICP) typically generate 2 to 4x the support volume and customer success (CS) time of in-ICP customers. When you segment your customers by cost-to-serve versus revenue contribution, you find a subset of accounts that are EBITDA-negative on a fully loaded basis. Renewing these accounts at current terms, or even retaining them at all, isn't always the right decision.

NRR is the third commercial EBITDA lever. Each point of NRR improvement reduces the go-to-market (GTM) investment required to maintain a given revenue level. The math is straightforward but underweighted in EBITDA discussions because NRR improvement is a 12-to-18-month cycle, while cost reduction is a 30-day cycle.

The Framework

A commercial-first EBITDA improvement program requires three analyses run in sequence.

Analysis 1: Price realization waterfall for the trailing 12 months. Build this from transaction records, not CRM fields. Identify the five largest categories of revenue leakage. Estimate the EBITDA impact of closing each by half. Prioritize the two with the best effort-to-impact ratio.

Analysis 2: Cost-to-serve segmentation by customer profile. Pull support ticket volume, CS time allocation, and professional services usage by customer. Segment by ICP match score and by original deal discount. Find the cohort of customers who are EBITDA-negative on a fully loaded basis. Calculate the P&L impact of running off or restructuring these accounts over 24 months.

Analysis 3: NRR bridge from current to target. Model the EBITDA impact of a 5-point NRR improvement: reduced gross churn requires less new logo acquisition to maintain revenue, which reduces GTM spend as a percentage of revenue. Show this as a five-year compounding impact to give the board the correct time horizon for evaluating commercial investments.

The Failure Case

A workflow automation company at $83M ARR had 21% EBITDAs and a board target of 28%. Finance proposed a plan built primarily on reducing headcount in CS and marketing.

A commercial analysis found an alternative path: price realization was 23% below list on a pocket price basis. Cost-to-serve segmentation identified 18% of the customer base as EBITDA-negative accounts. NRR improvement from ICP tightening was modeled to add $4M in annual EBITDA within 24 months.

Before: 21% EBITDA, cost-reduction plan requiring 40 FTE reduction to reach 28%, estimated $6M in severance and productivity loss.

After: Commercial improvement path implemented with no headcount reduction: price governance tightened, EBITDA-negative accounts restructured or not renewed, ICP definition tightened. EBITDA moved from 21% to 26% over 18 months. The board approved one more year before any headcount decisions.

What to Do This Week

Open your last 12 months of CRM data. Find the 10 customers with the highest support ticket volume and the highest number of CS touchpoints. Calculate what each of those customers actually paid net of all concessions.

If the bottom five on your cost-to-serve list are also in the bottom quintile on pocket price, those are your EBITDA-negative accounts. That analysis is worth a board conversation before any headcount decision.

Assess Your Commercial Health to get a structured view of the commercial EBITDA levers available in your current model.

For a deeper look at the monetization layer of EBITDA improvement, see Why Your Instincts Are Wrong About Monetization Strategy and Stop Guessing: Monetization Strategy Driven by Data.