External Control Layer

A home-buying algorithm buys more homes than the business can carry

Each cash offer fit Zillow’s pricing model, but together the offers bought more homes in the third quarter of 2021 than the company could renovate or resell, at prices above what it later expected to sell them for.

Date
July 2021
Kind of system
Predictive or matching model
Outcome
Harm

Gaps found

G1G2G3G4G5G6G7G8G9G10G11G12G13

G5 Blind to the total

What happened

In February 2021 Zillow began using its Zestimate as a live initial cash offer for eligible homes. In the third quarter, higher-than-anticipated conversion rates, in its words, led it to buy 9,680 homes, “further pressuring our renovation and resale capacity constraints,” while it sold 3,032. It ended the quarter with 9,790 homes in inventory and 8,172 under contract.

Zillow paused new purchase contracts in October 2021 and on 2 November announced it would wind the business down. It wrote down $304 million of inventory “as a result of unintentionally purchasing homes at higher prices than our current estimates of future selling prices.” Its chief executive said “the unpredictability in forecasting home prices far exceeds what we anticipated.”

What allowed the action to take effect

  • G5 Blind to the total

    Each action was allowed one at a time. Nothing tracked what the actions added up to in volume, rate, exposure, or repetition until the total did the harm.

    Purchases were not held to the company’s capacity to renovate and resell. Conversion rates “resulted in the purchase of 9,680 homes, further pressuring our renovation and resale capacity constraints,” and new contracts stopped only when Zillow suspended them in October 2021.

    Kind of evidence: admission. Source: Zillow Group, 2 November 2021, “Q3 2021 Shareholder Letter (Exhibit 99.3 to Form 8-K)”, Homes segment results.

    Related guidance: P8 Sequence awareness. Contract clause C-8.

Considered and left out

G7 Review in name only
A widely repeated claim that Zillow told its pricing experts not to adjust the algorithm’s valuations comes from reporting whose sources could not be checked, so it is not counted.

Entries with gaps in common

Sources

Cite this entry

External Control Layer. (2026). Entry 9: A home-buying algorithm buys more homes than the business can carry. In Control Gap Record (Version 0.3). G-14, Inc. https://externalcontrollayer.ai/record-9

@misc{ecl_cgr_9,
  author = {{External Control Layer}},
  title = {Entry 9: A home-buying algorithm buys more homes than the business can carry},
  howpublished = {Control Gap Record, version 0.3},
  publisher = {G-14, Inc.},
  year = {2026},
  url = {https://externalcontrollayer.ai/record-9},
  note = {Released 2 November 2026}
}

Uses Control Gap Classification 0.2.

Correct this entry

Email ecl@g14.ai with “ECL record 9” in the subject line.

Write the email