Hi all, founder @ Retool here. We've made a lot of progress over the past few years, but we first started on HN, and certainly wouldn't be here without all of you! For example, here's when we launched: https://news.ycombinator.com/item?id=14515494 . Hilariously, we described Retool as "Excel with higher order primitives", and people understood it! (Only on HN!) After that, we spent around a year polishing the product…
I've been really interested in trying retool since my head of ops pointed me to it a few months back. I think it should meet my use-case but I'm not entirely sure.
I run an experimentation program and I think retool might give me a way to build views on experiment performance for me and my stakeholders, but what I think it may be missing is some way of running statistics on the data, perhaps using python or R.
Data "about" my experiments (descriptions, tags, dates & times, stakeholder details etc) lives in airtable. I think I can get that into retool using sequin, right?
Performance data lives in bigquery. Sessions, clicks, conversions - all that stuff.
Right now I build ad-hoc reports in R/excel/data studio but as our velocity increases I need to streamline this.
I think what I'd like is for a stakeholder to be able to select the experiments and timeframes that they're interested in from a list, then behind the scenes the ID details to then be used to run a query in bigquery and that to pull back performance data. I'd like to show that in charts and tables on screen with how much impact the experiments have had (I think so far this should be reasonably straightforward with retool), but crucially I want to calculate confidence bounds for each metric.
Is there any way of bringing Python or R into the data flow somehow to do the heavy statistical lifting?
Am I looking at this the wrong way? Should I be trying to pre-calculate this nightly as some sort of scheduled job and just use retool as a presentation layer?