Which KMX Cost Lines Are Most Vulnerable to Cuts — and Where Does Cutting Risk the Customer Experience?

CarMax has committed to delivering at least $150 million in incremental SG&A savings by the end of fiscal year 2027 (February 2027). With total SG&A running at $2.44 billion in FY2025, the target implies a roughly 6% reduction from the current run rate. But SG&A is not a monolith. A line-by-line breakdown reveals that some cost buckets offer clean savings opportunities while others sit dangerously close to the customer experience that defines CarMax's brand.

Anatomy of CarMax's SG&A

CarMax's 10-K filings decompose SG&A into four categories. Here is how the $2.44 billion broke down in FY2025 (ended February 2025):

Compensation is by far the largest line at 58% of total SG&A. Within that, cash compensation was $1,289.7 million and stock-based compensation was $126.9 million. The "other overhead" bucket — which includes IT expenses, non-CAF bad debt, insurance, travel, and preopening costs — grew the fastest at 15.4% year-over-year.

Where the $150 Million Will Come From

1. Compensation & Benefits — The Primary Target (~$80-100M potential)

Management has already shown its hand. In Q3 FY2026 (November 2025), CarMax executed a 30% reduction in its Customer Experience Center (CEC) workforce. This is the centralized call-center operation that supports online and phone-based customer interactions.

The math is straightforward: with ~30,000 total employees and compensation running at $1.29 billion (excluding SBC), even a modest net headcount reduction yields outsized savings. A 5-7% reduction in the total compensation line would deliver $65-90 million annually.

The enabler is AI. CarMax's virtual assistant "Sky" now answers over half of inbound customer questions, driving a 30% improvement in containment rate and 24% improvement in consultant productivity. This is not speculative efficiency — it is already showing up in the numbers. Q3 FY2026 SG&A of $581 million was essentially flat versus Q3 FY2025's $576 million, despite inflationary wage pressure.

CX Risk: Moderate. The CEC cuts are the highest-risk move in the savings plan. CarMax's Net Promoter Score reached its highest level since the digital rollout in Q1 FY2026, suggesting the AI-assisted model is working. But the company acknowledged in Q3 that it needs to "improve the digital selling voice" — a tacit admission that fully automated interactions are not yet matching human quality for complex transactions. If containment rates plateau or customer satisfaction dips, this is where it shows up first.

2. Advertising — Already Lean, Limited Upside (~$15-25M potential)

Advertising has been on a steady decline: from $325.9 million in FY2022 to $260.7 million in FY2025, a 20% reduction over three years. Q1 FY2026 showed a further 5.3% year-over-year cut to $67.9 million.

However, management signaled in Q3 FY2026 that marketing spend per unit would increase year-over-year in Q4, reflecting the launch of the new "Wanna Drive" brand campaign. This is not a cost line management is willing to gut — CarMax needs to drive traffic at a time when retail unit comps turned negative (-6.3% in Q2 FY2026).

CX Risk: High if cut further. Advertising is a revenue driver, not pure overhead. With used vehicle comps under pressure, cutting deeper risks a volume spiral where lower awareness leads to fewer store visits, which leads to SG&A deleverage on the fixed cost base.

3. Other Overhead — The Quiet Opportunity (~$30-50M potential)

The "other overhead" bucket ballooned to $472.8 million in FY2025 (+15.4%), after a sharp decline from $565.6 million in FY2023. This category includes IT expenses, insurance, travel, charitable contributions, preopening costs, and non-CAF bad debt.

Several sub-items here are ripe for optimization:

• IT spending can be rationalized as AI tools replace legacy systems
• Preopening costs will moderate as the store expansion pace stabilizes (6 new stores planned for FY2026 vs. 5 in FY2025)
• Travel and administrative expenses are standard corporate efficiency targets
• Non-CAF bad debt is partially cyclical and may decline as credit conditions normalize

CX Risk: Low. These are back-office and infrastructure costs that are largely invisible to the customer. This is where management can extract savings with the least risk to the brand.

4. Store Occupancy — Mostly Fixed, Mostly Off-Limits (~$5-10M potential)

Occupancy costs of $285.3 million are predominantly lease obligations and facility maintenance — largely fixed and contractual. With CarMax actively expanding its store and reconditioning footprint, this line will grow, not shrink. Any savings here come from operational efficiency (energy, maintenance optimization) rather than structural reduction.

CX Risk: Low but savings are minimal. You cannot cut rent on a signed lease. This is not where the $150 million comes from.

The Scorecard

The Real Question

The $150 million target is achievable — the numbers add up across the four buckets. The harder question is execution sequencing. Through Q3 FY2026, SG&A as a percentage of gross profit has worsened to 87.5% from 74.5% a year earlier, driven primarily by gross profit compression rather than cost growth. If volumes do not recover, even $150 million in savings will not restore SG&A leverage to FY2022 levels (70.7% of gross profit).

The CEC workforce reduction is the boldest bet. CarMax is wagering that AI can maintain customer satisfaction at 70% of the prior staffing level. Q1 FY2026's record NPS suggests early success, but the Q3 commentary about needing to "improve the digital selling voice" reveals the tension. The savings are real; whether they come at the cost of the differentiated CarMax experience is the variable investors should monitor through FY2027.

Sources: CarMax 10-K (FY2025, FY2024), 10-Q (Q1 FY2026), earnings call transcripts (Q4 FY2025 through Q3 FY2026)