Optimizely

Malicious code inserted through an Optimizely Web Experimentation flaw reached customer websites in November 2026. Optimizely said it removed the code and completed remediation. Customers continued investigations and precautionary script disablement; Alaska and Hawaiian said flight operations were unaffected.
Unauthorized access to systems, accounts, networks, or data.
Malicious software other than ransomware used to compromise or disrupt systems.
Compromise of a trusted third party, supplier, software provider, or service provider used to affect downstream victims.
Exploitation of a software or hardware vulnerability to gain unauthorized access or execute malicious actions.
Data was intentionally changed, falsified, manipulated, or otherwise modified without authorization.
Cloud-hosted infrastructure, software, storage, identity, or platform services were unavailable or materially impaired.
The incident caused operational effects at customers, affiliates, subsidiaries, partners, tenants, or other dependent organizations.
Available evidence indicates that no extortion demand, threat, communication, or related pressure tactic was identified.
Optimizely confirmed that an unauthorized party exploited a vulnerability in Web Experimentation and added malicious JavaScript to customer projects. The altered scripts reached public websites through the software customers used to run web experiments. The earliest affected script in Report URI’s investigation was observed late Nov. 15 Eastern time, or early Nov. 16 UTC. The full intrusion start time is unknown.
Report URI found 188 affected snippets associated with 178 organizations in its sample, including U.S. sites such as CNN and Salesforce. Site visitors received altered content through a live production service. The researchers also reported verified cryptocurrency theft associated with the activity. The number of affected visitors and the extent of payment-card or other data theft remain unknown. Successful compromise of visitors’ devices has not been established.
Salesforce investigated altered scripts on certain branded sites and said it was confirming they were disabled in preview and production. Zoom said it disabled Optimizely across production environments on Nov. 17 as a precaution and had identified no impact to customer data. Equifax told Reuters it stopped using Optimizely and had found no evidence that its systems were compromised.
Optimizely’s Nov. 19 advisory said it blocked delivery of malicious code and closed the account-signup flaw within three hours of detection Nov. 16. It said it removed injected code from all affected projects and completed remediation. It said fewer than 5% of its customers were affected and that no malicious code remained. The vendor characterized the incident as exploitation of a signup-flow flaw, not a credentials breach.
Salesforce said Nov. 18 that it disabled Optimizely scripts pending verification, with no evidence of impact to Salesforce platform or subscription services. Smartsheet said its public marketing and Help Center pages were clean and its signed-in application and customer data were unaffected.
Hawaii News Now reported Nov. 18 that Alaska and Hawaiian airlines were investigating effects on websites using Optimizely. Regional communications director Alex Da Silva said flight operations and safety were unaffected.
Completed vendor remediation and the available customer service findings support an assessment that material operations have likely resumed. Continuing investigations and disabled experimentation scripts leave complete customer-level recovery unconfirmed; they do not independently establish interruption of core services.
The vendor’s disclosure and independent observations establish a software supply-chain compromise with downstream effects. The overall scope of customer and visitor impact, actor attribution, ransomware involvement and the full extent of data access remain unresolved.
The full exposure window, affected visitor population and amount of any loss remain unknown. Researcher assessments about the attackers’ location do not establish a named actor or state sponsorship. Customer investigations and removal of malicious scripts do not independently establish the security of every downstream environment.



San Francisco anchors Salesforce’s headquarters. The documented defensive response involved certain Salesforce websites.
Corporate headquarters anchor for a documented affected customer. This records the entity’s physical place and does not assert a citywide or premises-specific outage.
Corporate headquarters anchor for a documented affected customer. This records the entity’s physical place and does not assert a citywide or premises-specific outage.
Hawaiian Airlines corporate anchor in Honolulu; no flight, airport or citywide disruption asserted.
New York City anchors Optimizely’s global headquarters. The documented compromise affected its hosted Web Experimentation service.
Corporate headquarters anchor for a documented affected customer. This records the entity’s physical place and does not assert a citywide or premises-specific outage.
Alaska Airlines headquarters anchor in SeaTac; affected public websites do not establish an airport or citywide outage.
The publication reported Optimizely’s acknowledgment that an unauthorized party exploited Web Experimentation and documented downstream effects on major sites.
Salesforce investigated altered Optimizely scripts on certain branded websites and said it was confirming that the scripts were disabled across preview and production.
Optimizely confirmed an unauthorized party added malicious code to customer projects after exploiting a vulnerability. The vendor removed the code, fixed the vulnerability and secured affected accounts.
Researchers identified 188 affected snippets associated with 178 organizations and documented delivery to public websites. They reported verified cryptocurrency theft and traced the first observed affected snippet to late Nov. 15 Eastern time.
Optimizely said it detected malicious project changesNov16, blocked code delivery and closed the signup flaw within three hours of detection. All injected code was removed, remediation completed and affected customers contacted. The incident was limited to Web Experimentation and was not a credentials breach.
Signed-in members can report an error, update, or missing source.