Before retiring from politics, I ran internal polling for 2 guber campaigns.
The 2nd one used the data. They tweaked messaging in LGAs based on it. It also helped them know where to deploy (very limited) surveillance tech on election days, to get proof for tribunal if needed.

On the 1st campaign, polling was my idea. I got great buy-in from the candidate, but not enough from the old school politicians that ran the campaign departments, so usefulness was limited.
On the 2nd one, polls were already running before I joined. And there was more buy-in.

Off-topic, but my frustration about 1st campaign's dismissal of polls showed me my limitations with convincing certain types of people about certain things.
It made me re-evaluate how I sell, and I got better.
It could have been easy to just blame "old heads".

Those lessons have also helped me as a Product Manager.
Looking back at those veteran politicians, they were customers with jobs to be done. I was telling them polls were great, but I wasn't clearly tying polls to a pain point in their jobs, or listening to their pushback.