Zillow announces competition with $1M prize for person or team who most improves its Zestimate algorithm, with the final round ending Jan 2019
Kurt Schlosser / GeekWire :
Context & Ripple Effects
Coming off a quarter where Zillow beat expectations on Q4 revenue of $227.6M, up 34% YoY and guided past $1B for 2017, the company is now treating its core asset — the Zestimate — as something outsiders can improve, putting $1M behind the best submission by January 2019. That openness sits oddly beside its history of guarding proprietary advantage: two years earlier it paid $130M to settle the Move dispute over trade secrets and executive poaching.
The bet matters because the Zestimate is what converts Zillow's audience — approaching 140M monthly uniques after the Trulia deal — into advertising value. How far algorithmic accuracy can carry the company gets tested four years later, when Zillow shuts down its home-buying arm with a $500M+ write-down and ~2,000 layoffs.
First-order effects
- External data scientists gain a sanctioned path into one of real estate's most closely held models, with a $1M prize concentrating talent on Zillow's hardest problem through January 2019.
- Zillow's own engineers shift from sole custodians of the Zestimate to evaluators of outside submissions, changing how the company staffs and sequences model work.
Second-order effects
- A more accurate Zestimate strengthens the trust that drives Zillow's ad business — agents and premium advertisers pay for placement on valuations buyers believe, so accuracy gains flow straight into the revenue line that just grew 34%.
- Rival portals face pressure to match the transparency play or defend their own valuation models as less scrutinized, turning estimate credibility into a competitive metric rather than a back-office detail.
Third-order effects
- If the competition format works, it points toward platforms treating proprietary algorithms as contestable assets — opened periodically to outside improvement rather than hoarded, a reversal of the defensiveness behind the Move settlement.
- The longer arc cuts both ways: the same confidence in algorithmic pricing that motivated the prize later fails to protect Zillow's iBuying business, suggesting valuation accuracy is necessary but not sufficient for algorithm-driven transactions.
The trend: Consumer platforms are shifting from guarding proprietary algorithms as trade secrets to crowd-sourcing their improvement through prize competitions — while discovering that model accuracy alone does not de-risk algorithmic business models.