CoStar's Chaos

Hedge fund founder Dan Loeb doesn’t do gentle nudges. His company Third Point – an activist investor in OnTheMarket’s parent CoStar – launched a broadside at CoStar this week, one which also lands squarely in UK agency inboxes because the thread runs straight from Washington to Wolverhampton: if the hedge fund gets its way, the oxygen feeding OnTheMarket thins fast.

The numbers are ugly. Around $5bn sunk into CoStar’s residential portal landgrab in the US, UK and Australia; c.$60m US residential revenue in 2024; a five‑year share price drift that flatters no one; and a $37m CEO payday that invites headlines. CoStar’s riposte – buy back $1.5bn of stock, keep calm and AI on – tees up the real tension: can their much-trumpeted Feb-launch of “Smart Search” be the saviour of its residential portal play?

Homes.com’s conversational search seems pretty strong. It fuses language models with computer vision to understand “vaulted ceilings”, “south‑facing garden” and “walkable to Northern line”‑style intent – and then ranks results around lifestyle rather than blunt filters. They may not have the property volume to make it truly fly elsewhere – something like 60% in the UK and about the same in Australia – but it ought to get traction in the States because the Multi-Listing System over there means they’ve got 100% of the properties available.

However, that does produce a soon-to-be pounding headache for them too. In America, a Buyers’ Agent system helps their broker agents (or ‘realtors’) to double their 3% fee by helping people to +find+ their new home through all of the available stock, as well as sell the one they’re in. If an AI can do this discovery piece – and for broadly $0 – then that second 3% looks under serious threat. That will not go down well with agents staring at 50% of their revenue evaporating.

It’s not like these agents aren’t represented, either – the National Association of Realtors is a fiercely protective organisation. So CoStar’s big launch of its new AI will either face hostility at home from its customers – or apathy away, as people can’t find enough of the stock they want on the paid-for portals in the UK and Australia.

Hence the angry investors. This is a space – residential – that doesn’t operate with the same juicy margins that CoStar is used to from its commercial property background, unless you are already a dominant market leader.

Two clear paths emerge:

- Hold the line: CoStar backs the residential bet, pipes Smart Search into OTM and Domain, and uses the UK to showcase AI‑led discovery without detonating US agent relations. Apartments.com, one of their first residential portals in the US, took years; board and investors grit their teeth again.

- Trim the sails: new directors, tighter capital allocation and a pivot back to core commercial. International residential becomes a “for sale” sign and OTM is back on the block before the Homes.co.uk stickers have finished printing.

And what of perennial bridesmaid Zoopla? Time was that it seemed certain to be swallowed by CoStar’s insatiable acquisitiveness – now there can be virtually no chance that a takeover would be allowed by CoStar’s board. Silver Lake – Zoopla’s current owners – has a focus elsewhere (they were part of the consortium that is reported to have acquired TikTok in the US so will be loath to spend too much time on a 7-year stagnating asset when there’s much more fun to be had with 7-second videos for billions of Stateside teens.

 A break‑up across Zoopla, Hometrack, Confused and Uswitch is plausible. Rightmove doesn’t need it. CoStar won’t be allowed it. Trade buyers are thin. Private equity enthusiasm for portals has cooled. A bargain, yes; an obvious buyer, no.

So what should UK agents do now?

- Cut back. If you’re spending on more than one paid-for portal, are they all delivering? If Scrutinise marginal value. Spend where the next instruction will actually come from and withdraw from the others.

- Prepare for conversational search. Rewrite listings for semantic engines: plain‑English features, lifestyle landmarks, neighbourhood nuance. Feed the model or be buried by it.

- Own your demand. Invest in first‑party data: email capture, WhatsApp broadcasts, SEO for “streets and schools”, and community content that ranks when paid-for portals are bickering. Paid-for portals shouldn’t be your only tap.

- Measure lead quality, not applause. If OTM ships Smart Search, watch intent depth, duplicate rates and time‑to‑viewing. If quality spikes, lean in. If it floods you with tyre‑kickers, push back hard.

- Trial AI on your side. Deploy your own discovery layer: natural‑language site search, image‑matching from saved homes, and AI triage on enquiries. If the portal is getting smarter, your funnel must too.

Whether Loeb forces a retreat or Florance forces a rethink, the centre of gravity is moving from portal placements to intent orchestration. The real question isn’t whether OTM survives. It’s whether you’ll still need any paid-for portal as much once consumers can describe their life and be shown the home – by your brand and on your free-to-list channels. That’s the edge to build while the giants trade blows.