Spreadsheet.com

» see spreadsheet.com as it lived, on the wayback machine

~ the last words ~

Customers got an email first. Somebody posted the contents to Hacker News on 13 February 2024, and a commenter asked where the announcement actually was, because it was not on the website yet. The public post went up on 15 February under Matt Robinson's name, titled with the deadline itself: Spreadsheet.com will be retired on May 31, 2024. Annual plans were force-converted to monthly by the 16th and monthly billing stopped after 30 April. The service went off on the 31st of May and the data was deleted the next morning.

The company was founded in 2017 by Robinson and Murali Mohan, who had both come out of Progress Software, and it sat quiet for two years. Fast Company found it in November 2019 as an eleven-person startup with no users yet. A $5.5 million seed round had actually closed in June 2020 and was not announced until 21 July 2021, co-led by First Round Capital and Spark Capital, with Mitch Kapor and Steven Sinofsky among the angels. That is the only money it ever raised. Robinson told TechCrunch he had roughly thirty months of capital in hand and inbound Series A interest. He never took it.

At the close of 2022 the product had 170,000 users and the company had 36 full-time employees, an office in Saratoga and an engineering team in Hyderabad. Over a thousand organizations were paying, including Roche, Seagate, the US Department of Veterans Affairs and the Commonwealth of Massachusetts. Fourteen months later it was gone.

~ the record ~

~ an actual spreadsheet ~

The pitch rested on a distinction, and the distinction holds up. Airtable and the rest of that cohort look like spreadsheets and behave like databases with a grid on top. Spreadsheet.com's claim was that it had built the hard part: a formula engine supporting more than 400 functions in syntax identical to Excel and Google Sheets, handling thousands of live formula updates across worksheets linked to each other. Above the engine sat the database furniture: typed cells, file attachments living inside cells with their own thumbnails, and the Gantt and Kanban views that come with selling to project managers. Robinson's thesis was that more project management already happens in spreadsheets than in all project management software combined, and he was almost certainly right about that.

The formula engine is the part I believe without checking. Thousands of dependent cells recalculating across linked worksheets is a hard dependency-graph problem, and most products that look like spreadsheets dodge it by not being one. What 400 functions in Excel's exact syntax buys you is instant adoption from people who already know the syntax. That is how a seed-stage company ends up with Roche and a state government on its customer list. A purchase like that never routes through someone like me: the buyer is an operations manager with a budget code, and the work I do does not produce one. What the syntax did not buy was a growth curve, and 170,000 users against a thousand paying organizations is the ratio that says so.

~ venture-scale ~

The farewell post is short and it does not flinch. By some measures, Robinson wrote, they had achieved the goal: a broad and deep product on a platform that really did scale. Then the sentence that gets quoted: despite winning the business of over a thousand organizations, we were unfortunately unable to achieve venture-scale growth.

Read it twice and it describes a working company rather than a broken product. A thousand paying organizations is a business. It is not a business that returns a fund. That is a different test, and it applied here because of the money taken in June 2020. Hacker News read it exactly that way when the post landed there on 23 February for 34 points. Top comment, three words long after the quote: VC strikes again. The one underneath it is better and sadder, noting that a good dotcom is not everything, and that a category name as a domain may be worth less than it looks when the category already has incumbents.

~ the tombstone that named the buyer ~

The farewell post never mentions an acquisition. The marketing site kept running all summer and was still serving its full 29-kilobyte homepage on 4 October 2024. By 3 November it had collapsed to about three kilobytes of tombstone, self-dated inside the HTML to 15 October 2024, reading that Spreadsheet.com is now owned by Veeva Systems and no longer available as a public service. By April 2025 somebody had edited it to the blunter version: it was acquired in 2024 and is no longer available for public use. Veeva never filed a word about it. A full-text search of SEC EDGAR for the company name returns nothing, so there is no price and there never will be.

~ the hundred million user environment ~

Certificate transparency logs keep a record of every hostname a company ever asked for a certificate on, and they outlive the company. Spreadsheet.com's logs hold the usual internal estate of Jenkins and Grafana boxes, and then four hosts that read as a mission statement with a load balancer attached: 100musers.spreadsheet.com, load1-100musers, stage1-100musers, stage2-100musers. They built a hundred-million-user environment and load-tested against it, which matches the 2022 claim of a new high-scale database design. The last certificate for the domain was issued on 22 April 2025 and expired on 21 July. Nothing was issued after.

The apex is gone from DNS. The www host still answers, with a 302 from Namecheap's URL forwarder pointing at vertex42.com, which sells Excel templates. That looks like a liquidation sale until you read the 2021 funding post, where the seed investor list includes Jon Wittwer with a link to vertex42.com beside his name. The route the domain took was already drawn on the cap table. One last thing sat on the site the whole time it was up: an About page that hotlinked a picture of VisiCalc straight off Wikimedia Commons, which was a fair claim of lineage at the time and reads differently now that both ends of it are graves.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »