The Sunnyvale Charter Review Committee on Oct. 23 received a summary of a community survey that drew 302 responses, which staff described as roughly 15 hours of public input.

Connie Bersellas, deputy city manager, told the committee the survey produced “over 300 responses, 302, to be exact,” but she warned that staff provided only a summary because their platform is "very, very clunky" and retrieving full verbatim responses would have required heavy staff time. Bersellas also said a question about whether respondents were registered to vote was accidentally left off the Sunnyvale results and apologized for the error.

The presentation gave committee members and members of the public a high-level snapshot of attitudes on topics including how to fill council vacancies, whether mayor and council pay is appropriate, and the city manager’s settlement authority. Committee members repeatedly cautioned that the online survey was self‑selected and not a scientific poll: several said differences between registered and unregistered respondents were pronounced and that small sample sizes limit precise interpretation.

Bersellas said staff followed the outreach plan and used the city’s social platforms to promote the survey, and that staff are beginning to explore better survey tools after receiving user feedback. She said staff did not include the full written responses in the meeting packet because of turnaround time and the practical burden of printing hundreds of free‑form answers.

Public commenters and committee members flagged specific design problems that could bias results. Zachary K. and another commenter questioned question wording and omitted response options; Zachary asked whether respondents realized a missing choice existed in Question 1 (filling council vacancies) and said several questions appeared “loaded.” Steve S., a frequent commenter, urged the committee to remove the word “modernize” from ballot or explanatory language, arguing it can mislead voters, and he and others asked the committee to examine registered vs. unregistered answers.

Committee members recommended use of the full dataset for follow‑up analysis. Several members noted that registered respondents (72,759 registered voters was cited from county records) appeared to answer differently than unregistered respondents, and that staff and any subcommittee could extract and compare registered and unregistered responses from the platform if a deeper dive is desired.

The committee treated the survey as a “pulse” rather than a statistically valid sample. Members suggested next steps including targeted polling if council chooses to advance specific charter changes and offering subcommittees access to the raw responses so they can filter and analyze comment fields.

The committee closed the item and returned to its draft reports. Staff and committee members said the survey results will inform subcommittee work but do not by themselves determine recommendations to City Council.