We deployed a form with 5 sections, two of them have repeating groups and we also have several cascade select questions, my choices sheet has around 1500 rows, and my survey sheet has almost 400 rows.
The final section is a summary of some of the answers in the form, and it sums some of those responses.

My coordinators complain about the form being too slow, I’ve read that I can take the sheet choices to another document and upload it to the form to make the response faster, does this work without internet? I have some people in the field and they have problems with the internet connection.
I was giving a training session with the form and I have a good internet connection and it was really slow too, so I don’t think the internet is the problem.

I need to improve the performance of the form, as soon as possible, PLEAASE, What can I do? :sob: :sob: :sob:!!

Thank you in advance

Hi @analucia_espinoza,

Would you try out the following to see whether the rows (1500 rows in your choices sheet and almost 400 rows in your survey sheet) or the long complex syntax within your survey form is causing the lag.


  • Download your xlsform from your KoBoToolbox server (and create a backup if required).
  • Delete the expressions from the relevant column.
  • Also delete the calculate question types and the expressions used under the calculation column.
  • Deploy as a different survey project and see if its still lagging.

If and in case, the newly deployed survey project with 1500 rows in your choices sheet and almost 400 rows in your survey sheet is not lagging, then it should be an issue of the expressions used within your survey form. Try making it short and simple. Sometimes keeping them complex also hampers the performance of the form being deployed.

Have a great day!

Hello @analucia_espinoza ,
you may also start the testing, Kal_Lam proposed by reducing the long lists of choices to very few choices and see how this changes speed. If this is the problem, you may change to pulldata approach (doesn’t require Internet with KoBoCollect).

For testing only parts of the App you may enclose other parts with a (test) begin_group and end_group and a relevant = FALSE on this group level…

Are you using KoBoCollect OR Enketo? OCHA Server?
Kind regards

1 Like

Hola @analucia_espinoza

Did you manage to get a breakthrough with the provided suggestions? Also note some devices may be slower.


A post was split to a new topic: Sharing the findings of the survey