Using multiple cross-docs to reduce size of large docs and increase performance

I am looking for some feedback on best practices regarding large docs and cross-docs. I currently have a “Clients and Projects” doc for our construction company that syncs a lot of project information through Zapier, from quotes to invoices, permits and inspections and transactions. Before I deleted as much information as I could, the doc had 14K rows and failed to open up on a mobile device. We have about 6 different departments that interact with the doc, so I also have a lot of views and it again started to fail opening up on a mobile device even though I got down to about 8K rows. Many of our guys are out in the field and referencing their docs on a mobile device is important.

I currently use a lot of cross docs to sync information but I was thinking about creating separate docs for all the different pages (invoices, quotes, transactions, hours, etc) and different departments and syncing those. I’m wondering if I have every doc heavily dependent on cross-docs if it will just cause more headache that benefit. It will certainly be more work to mimic each view, conditional formatting, etc but it might help me get the doc in the range of optimal performance.

The other option is to kind of wait out performance for Coda handling large docs.

Polite bump… :slightly_smiling_face:

I also wonder if using multiple cross-docs in a doc is better (lighter) or worse (heavier) than one massive “native” doc? (so bigger but without cross-doc tables)

Rather know this before I invest many hours building multiple cross-docs only to find out it hurts performance…

Another polite bump… Same question: Is “using multiple cross-docs in a doc is better (lighter) or worse (heavier) than one massive “native” doc? (so bigger but without cross-doc tables)”

EDITS:
Let me put a finer point on it - my current structure is to pull in data from third-party apps (Strava, Todoist, Toggl, etc) that may each generate 1500-4000 rows per year. Each of these apps currently gets a dedicated Coda doc. I also have other tables of similar sizes but which are no longer growing (at least as rapidly). In total, let’s say 10 tables. To do some quantified self work, I am bringing slices of data from each table together for daily, weekly and monthly reviews.

From a performance perspective (especially with regards to mobile), I have three alternatives to consider:

  1. Mega doc (contains all 10 data tables)
  2. Cross doc dashboard (the dashboard doc contains nothing but 10 “full” cross-doc tables – e.g. all rows)
  3. Sliced cross docs dashboard (the dashboard doc contains about 30 “small” cross-doc tables – each based on a view of between 1 and 31 rows)

How would I analyze these three approaches? Total number of rows? Total number of tables? Cross-doc versus “local” tables?

Any insights are greatly appreciated!

:wave: Hey there!

Its really super specific depending on your use case. There are some tricky complications that come in using Cross-doc when wanting to scale to more and more docs that I’ve run across with some larger organizations.

Cross-doc can definitely provide a valuable resource for your larger scale Coda infrastructure, but I would caution in various different ways. Again, hard to actually provide real and specific advice without sitting down on a call to hear your very specific use case.

All in all though, less rows does = more performant documents. So its just asking the question: “Are the hurdles, set up, and maintenance of using cross-doc in a rather robust way worth the smaller doc size in terms of rows that I get?”

Sometimes the answer really is yes - and sometimes no. Again very specific.

Happy to jump on a free call to help consult if you sign up for my next GetUnstuck meeting and I can help give you more specific answers in person (Sign up right now is on Twitter) : https://twitter.com/thecodaguy/status/1501257311813201923

There are different creative methods for making your doc more performant as well like archiving in unique ways, optimizing formulas, or setting overall structure and schema differently. All these options (along with cross-doc) should be explored when you really start pushing your Coda use into the stratosphere.

2 Likes

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.