I love spreadsheets. They let you touch your data like no other tool can. They're the universal interface for data across industries and disciplines because they're the easiest way to create a graph, clean-up data, forecast demand, or model how $BUSINESS_OUTCOME varies with $SOME_INPUT.
When I say spreadsheets, there are only two players - Excel and Google Sheets. They are terrific tools - provided your work fits in their constraints:
Your data is small. Desktop Excel has a 1 million row limit. Workbooks start freezing well before that. Google Sheets is worse.
Your data is on your laptop. Excel made sense as a desktop application when it launched in 1985. But now data's in the cloud. So you have to download the data first, and once you do, you're limited to the memory and compute resources of your laptop.
Google Sheets [1] and Excel 365 have the same problem. Even though you access them via a web browser, the formulas run on your laptop - not the server. They're slower than Desktop Excel because Javascript is doing the compute. As incredible as v8 is, it can't compete with optimized Rust/C++. As a result, most serious spreadsheet work is still done in desktop Excel.
Your model is simple. If your formulas are simple MINs or COUNTIFs, everything's fine. But this is what we see all the time:
No one - not even the author - is sure this is correct. Multiline expressions like these are begging to be decomposed with a function. Writing a function in Excel/Sheets means VBA / AppScripts. Both are terrible.
In my previous roles as a manager and engineer, I ran into these constraints all the time. To do some analysis, first I'd write a SQL query and extract a csv. Then I'd discover the data had more than 1M rows, so I'd have to go back to the database and filter it to some subset. Then I'd import the data into Excel (which would lock the UI for a minute) and discover one of the CSV columns actually had a JSON blob in it. sigh. Rather than try to parse that in VBA, it's easier to go back to the database and use a built-in database function like JSON_EXTRACT_PATH_TEXT to pull out what I needed. Re-query, download, and import into Excel. Create a computed column with a double click-drag. Excel locks up for 30 seconds. Then pivot to aggregate. At last, I have a graph. Now I need to repeat the entire process for 3 other different subsets because the data was too big for Excel. FML.
My peers in BI, product management, operations, and finance all had the same problem: We hit Excel's limits.
“But why don't you just use Jupyter?” My business partners in finance and operations need to consume my analysis and layer their own on top of it. They don't know Python. The deliverable has to be a spreadsheet.
“Just use Tableau/PowerBI/Quicksight/Looker!” These are all a different interface/skill-set than Excel. They make it hard to see individual rows and write custom formulas, and my business partners don't know how to use them. Even if they did, my organization wouldn't want to pay nearly $1000/year for everyone in the org to have an author license.
What I need is a better spreadsheet.
And that's what we are building: Row Zero is a blazingly fast, connected spreadsheet. It's for Excel/Sheets users who need more power and Tableau/Power BI/Looker users who miss the flexibility of a spreadsheet. We have all the features you'd expect, like formatting, keyboard shortcuts, real time collaboration, pivot tables, and Excel's built-in formulas.
We are deeply integrated with the cloud - workbooks run on AWS and connect directly to Snowflake, Redshift, Postgres, and S3. And it's fast - a 1 GB, 10 million row CSV imports in 5 seconds. Once it is imported you can filter, sort, pivot, and plot instantly.
And finally: No. More. VBA. You can define custom functions in Python and consume those as formulas in spreadsheet cells. Your functions can use popular Python modules like pandas, numpy, scipy, and yfinance to import data or express complex business logic.
I've been comparing Row Zero to Excel and Sheets, because that's what the product feels like. In terms of where we fit within an organization's data strategy, we're a faster, affordable alternative to Tableau, PowerBI, and Looker. Many data scientists prefer us to Jupyter.
We're SOC 2 Type II certified. We do regular penetration tests. The engineering team is all ex-Principals from AWS S3, Tableau, and Airtable. We know what it takes to build secure, durable, available, and performant cloud services. We obsess over security so you can trust us with your most sensitive data.
Please give Row Zero a try and let us know what you think!
Footnotes [1] Google Sheet's choice to do compute locally makes sense - Sheets launched 10 days before AWS. But since then, data has moved behind APIs and into cloud databases.