Admittedly it’s been a while since I did my research project and it didn’t involve a poll but I feel the answers presented here could have been given some more thought.
There are gaps in the options given, such as I mentioned above. No option exists between visiting a branch every 1-2 months and never.
I understand why you wouldn’t want to include a freetext option as that would make the dataset a nightmare to analyse but you could have broader answer options that accurately cover most reasonable scenarios. Or depending on the detail you want to go into, you could include more granular options and then present them in two sets: one broader and one more focused.
For lack of a better example yesterday’s Coronavirus briefing had a great example of what I mean. Chris Whitty presented the % of positive tests in different age groups from 0-89 years old in groups of 10 years each. But the next slide included a more detailed view of the incidence rates for under 21s in 3-year groups.
Again, depending on what you want to present and the relevance, it might not be the right way of presenting it but could be an option to consider.
Having more options is never bad - it’s easier to group them together later than try to split them out when you discover you don’t have enough (quality or quantity) data to analyse
Edit: also, you mention potentially editing the poll. Again, as I didn’t use surveys as part of my research I can’t comment on whether this is standard practice. However, I can’t help but feel that it would invalidate your previously gathered data since the methodology changed halfway through