CoStar Harvests the Margins It Promised — and Discovers It Cut the Sales Force Too Deep

For two years CoStar Group told investors to be patient: the money going into Homes.com and into a 600-person inside sales floor in Richmond was an investment, and the profit would come later. In the third quarter of 2026, the profit came. It arrived attached to a problem the company had been deferring, and this time the deferral ran out.

The harvest is real — and partly borrowed

Revenue rose 14% to $952 million, the 62nd consecutive quarter of double-digit growth, and adjusted EBITDA doubled year over year to a record $229 million. That is a 24% adjusted EBITDA margin, ten points better than a year ago and the seventh straight quarter of expansion. Full-year guidance of $820–840 million implies a 22% margin against 14% in 2025 and 9% in 2024. “Revenue grew 14% year over year to $952 million, marking our 62nd consecutive quarter of double-digit growth.” — Andy Florance, Founder and Chief Executive Officer · 2026-10-27

Then read the footnote. CoStar now expects 2026 operating costs to land $150 million below its February plan, but new CFO Robin Rossmann was explicit that only about two-thirds is durable: roughly $100 million is run-rate savings, and “the remaining $50 million relates primarily to lower than anticipated sales headcount.” — Robin Rossman, Chief Financial Officer · 2026-10-27 Some of the margin beat is not efficiency at all. It is under-hiring.

The genuine efficiency work is the $114 million of effective SG&A, down 7% year over year even as revenue grew, a ratio that now sits near 12% of sales. Rossmann, on his inaugural call, put another $200 million of opportunity on the table across organizational design, acquisition integration, and procurement-plus-marketing-ROI — and pointedly noted that in Europe he had once found savings equal to 25% of the cost structure, which at group scale would be roughly $750 million. Analysts asked the obvious question: if that much is coming, why merely reaffirm the 2028 target of $1.25 billion of adjusted EBITDA rather than raise it? “We haven't gone through a whole new planning cycle and redone forecasts for the long term.” — Robin Rossman, Chief Financial Officer · 2026-10-27 Believably cautious — and also a signal that the new CFO is not yet willing to put a bigger number in print.

"Did you cut too deep?"

I'm just trying to understand sort of what's going on with the sales force... did you cut too deep? I think there might be somewhat of an issue of not having the right tenured people in the right place.

That is the quarter in one question. Annualized net new bookings were $37 million against $84 million a year earlier, and revenue guidance ex-Zonda came down $35 million, or 1%. CoStar now runs about 2,000 sales professionals, down roughly 10% year over year; Homes.com's inside team has been cut from a peak of 655 to 267. Apartments.com ended the quarter with 91,155 paid properties, down 2% sequentially, with the monthly renewal rate slipping 20 basis points to 98.4% and average revenue per property flat. The company that spent four quarters arguing it would not trade price integrity for penetration quietly discovered it lacked the bodies to do either.

Andy Florance, who has never been shy, conceded the execution point: CoStar's and Apartments.com's recruiting and management “could have done a better job and continue to grow” — Andy Florance, Founder and Chief Executive Officer · 2026-10-27, while the Homes.com experiment of hiring new graduates into a centralized floor produced exactly the churn anyone outside the building would have predicted. The remedy is a decentralised field model — recruiters in roughly twenty city hubs, compensation plans rewritten to reward headcount growth, a new CHRO announced within two weeks, and 30% field sales growth targeted.

What makes this a reversal rather than a stumble is the tape of the company's own language. In July, Florance was still describing a deliberate programme: “we continue to invest on optimized sales force growth... shifting to more field salespeople in both homes and apartments, which we think will have a benefit.” — Andrew Florance, Chief Executive Officer · 2026-07-28 In April, after a third consecutive sequential decline in bookings, he called the pattern “like a groundhog day” — Andrew Florance, Chief Executive Officer (CEO) and Founder · 2026-04-28. And in this quarter's keyword set, the two dominant themes have shifted wholesale from demand to org chart: sales teams and inside sales now top the list, while lead to lease conversion and price integrity — the competitive arguments that carried the last several calls — have dropped out of the top ranks entirely. When a management team stops defending its pricing and starts describing its recruiting funnel, the debate has moved.

Buying altitude instead of building it

The genuinely new thing this quarter is not a cost programme; it is $730 million of capital. CoStar has agreed to acquire Vexcel, the private aerial survey business behind the imagery in Bing Maps: roughly $140 million of trailing revenue growing better than 30% on multi-year data contracts, about a 20% EBITDA margin, and — remarkably — six salespeople. Vexcel flies its own fleet, designs its own cameras in Graz, and collects more than 60 million images a year across 45 countries at 7.5 centimetres, which Florance framed as sixteen times the pixels per square metre of commercial satellite imagery, feeding computer vision that can read a thousand attributes off a roof.

This is the aerial imagery bet, and it is the logical extension of an argument management made in February: “AI cannot create data from nothing. If it doesn't know something, it can't create something it does not know.” — Andrew Florance · 2026-02-26 Rather than spend four more decades driving researchers past buildings, buy the camera in the sky. It is funded from the $500 million of quarter-end cash and an undrawn $1.1 billion revolver, against $266 million of effective net cash, down 90% year over year. Rossmann's accretion maths — accretive in year one before synergies, four times more accretive than a buyback by year three — is doing a lot of work in a quarter where the company also bought back $82 million of its own stock.

What is left in the drawer

The second engine is depth advertising. Homes.com grew 50% to $31 million with more than 36,000 members, and the September launch of Platinum pushed average subscriber pricing from $271 to $429 in October, a 58% step-up. That is the playbook from Apartments.com, where roughly 80% of revenue is depth advertising, and from Domain at 85–87%. It is the first honest layer on a twenty-year pricing ladder.

Domain remains the drag: down 5% on a pro forma basis as Sydney home sales fell 27% and Australian house prices declined in every city for the first time in fifteen years. Zonda, closed 21 August for $815 million, contributed about $42 million pro forma with subscription up 4% — the keyword surfaced a quarter before the deal hit the numbers, which is how CoStar's language tends to lead its disclosure. And the AI narrative has rotated: Homes AI was the single biggest keyword riser two quarters ago and the biggest faller the very next one; today's AI conversation is about efficiency opportunities and ChatGPT referral traffic rather than product awe. Florance's most durable line was defensive: “75% of our content is not visible to the broader internet.” — Andy Florance, Founder and Chief Executive Officer · 2026-10-27

None of this is a market wave. CoStar's keywords appear nowhere in the broader market's top themes this quarter, which are dominated by rate anxiety — rising rate environment, Fed hikes, the cost of deposits. CoStar touches that only at the margin, where Florance flagged likely softness in investment sales as rates rise while banks grow more interested in portfolio risk. This is a company-specific story about execution capacity, and the equity reflects it: the shares had rallied 9.5% over ninety days to $35.45 on 22 October, then gave back 9.7% into the print, still 68% below the October 2021 peak of $99.74 and now at roughly 3.3x revenue, down 72% year over year.

The bull case now rests on one falsifiable claim: that Q3 was the trough in bookings, with sequential improvement in Q4 as field reps ramp. Everything else — the $200 million, the Vexcel synergies, the Platinum ladder — is downstream of whether CoStar can hire salespeople faster than it did the last four quarters. The margin machine works. The distribution machine is being rebuilt in public.