Today, I’m excited to announce that Coda is entering a strategic partnership with Snowflake, the world’s leading Data Cloud company. In addition, Snowflake Ventures is leading a large investment in Coda.
The partnership originated after realizing that Coda and Snowflake share a vision for the future of work: one where the divide between data and action is eliminated across the enterprise. As Christian Kleinerman, Snowflake’s Head of Product, framed it:
The partnership started with a very simple insight that most companies are split-brained: one side is the company’s data, which is more accessible to data professionals. The other side is made up of the team at large and the tools they use that take action. Because the two halves can be divided, there’s a big opportunity to reduce friction and make enterprise data more actionable for every team, regardless of role or technical skillset.
There’s a lot of work in flight right now, but in the meantime, I encourage you to read the full story on the blog to get a first glimpse at what the partnership will do you for you and your teams.
This community continues to inspire us—and we can’t wait to show you what’s next.
wonderful news.
it may open the door to an enterprise level data storage as well inside a doc, kind of SF tables in Coda that sync in milliseconds.
The investment I read as a token of trust.
Some will ask, ‘why enterprise only?’ and I can see where that comes from. I read this announcement as an indication that coda will focus on specific clients: larger companies. I like this kind of clarity and I am sure it will generate a massive need for competent makers.
Less financial dependence on the lower tiers. That might translate into more features at stable, possibly favorable pricing.
More consultants driven by enterprise-grade requirements will have a backwash effect for SMBs and solo workers who benefit from a wider tree of solutions and templates.
Increased sustainability of security and data privacy as new threats emerge. Everyone benefits when enterprises demand timely improvements.
A key focus on data at scale and in real-time; two aspects of modern apps that Coda cannot presently provide and Snowflake is best-in-class.
AI. Coda AI has driven into a cul de sac; its decidedly OpenAI-centric integration leaves no room for the LLM-agnostic world that is advancing quickly. If Snowflake itself doesn’t nudge Coda AI into a better place, the new investment will surely make this probable.
Enterprise training will accelerate and reward Coda providers and template-makers. They love to buy training. These resources tend to find their way into the hands of lower tier users at low cost and often free.
This is precisely what we’ve been doing with ElasticSearch and Coda since 2019. In mid-2023 we began to lean away from ElasticSearch and toward Snowflake. Internally we use Coda + Packs + ElasticSearch to make it easier for every worker to perform AI tagging and outcome validations. The AI → Data → Coda train has been a key differentiator making it possible for us to perform AI work far quicker than ever before.
Feeling happy for Shishir and the Codans. The idea behind the collaboration makes total sense and I am sure it will give even more power to how data is used.
But at the same time I am feeling sad for the solopreneurs and small businesses who dedicate their time and passion to Coda. It has been the enthusiasm that Coda is giving the power of solopreneurs and small businesses to be creative, agile and maintain their competitive advantage towards the big enterprises.
We all collect data in Coda. Our understanding of big enough data. Why only enterprises would benefit from Coda’s efforts on AI development?
You asked “what it meant” for users like your self who do not work in large enterprises.
Data at scale is one (of dozens) of dimensions that may positively affect your work in smaller businesses.
As an example…
Today, Coda AI is unable to digest your entire document and make much sense out of it. The future of Coda AI depends on the ability to digest lots of data. If they partner with Snowflake, there’s a good probability they will have to resolve this issue at scale, and that means they will resolve this challenge for all small business users. Ergo, you will get the benefit of enterprise-grade AI and most likely without Coda increasing your price. Enterprises, after all, often subsidize many costs of lower tier customers.
I have a hunch you are not a NASCAR fan, and neither am I. But I am a fan of not dying in an automobile crash. Despite NASCAR not being ‘our world’, we both benefit greatly from NASCAR. The firewall in our cars came from racing. Our seatbelts, durable brakes, and crush zones – all NASCAR.
Stefan, I hear you about the solopreneurs and small businesses who may not feel like this launch is for them. And I completely understand the idea that Coda gives superpowers to these small companies to be creative and agile—I’ve seen it (and felt it myself).
What I would say is that these partnerships and product investments inevitably improve the product for everyone over time. True, the Snowflake Pack may not be for every solopreneur today, but maybe it will be in the future. And no doubt, the improvements we’re making across the product will be felt by everyone.
And finally, please continue to share this valuable feedback with us! It’s super helpful in understanding your (and others’) hopes and dreams for Coda.
i was thinking the other day when coda will be useful for REAL data analysis, and this kinda answer my doubts
I mean, when i will be able to have more than 10.000 rows in a sync table? 'cause after a while reaching 10.000 rows is likely will this allow to move the data in the external platform?
In my experience of working with both enterprise and small businesses, I’ve found that both have different approach to how they use and what they expect from tools. I’m guilty to have made a decision to allocate 10k hours in development for a group of users, inadvertently neglecting others. I now understand the need for a more equitable distribution of resources. I’m hopeful that Coda will avoid such imbalance.
The magnetism of a new partner in an entirely different strata of customers will undoubtedly attract Coda’s resources. Some will view this attraction as a distraction from features and functionality important to them.
This partner happens to be an investor, perhaps a significant investor. Historically, this is a well-known pattern. At the very least, Coda is now obligated to pursue and achieve the objectives of the partnership. Ultimately, Coda may be acquired, and such a step could result in a fracture from its roots.
Uncertainty affects everything despite the outward message it’s business as usual. We hope it’s a bright future that embraces the legacy of Coda. This is a very difficult tightrope to walk.
The paradox of software is that we expect nothing to ever change, but we always want improvements.
This suggests we’re competing mostly with enterprises. I don’t think this is a predominant case. Solo consultants and small businesses often have a vision to grow large enough to need platforms such as Snowflake.
Hopefully, Coda can convince their new partners that smaller firms, SMBs, SMEs serve as incubators for future Snowflake customers. Coda has the capacity to be a growth conduit.
Another key Coda advantage is the increasing move for enterprise departments to operate as autonomous business units that are not unlike small companies. Coda could become a powerful data conduit for the most challenging of data types - unstructured information. Snowflake probably understands that there is a much bigger story than a Pack that syncs data to the edge.
Arctic is ideally suited for SQL code generation, sophisticated instruction following, and complex data analysis. I have a hunch that it won’t feel right as a wholesale replacement of whatever Coda is presently using for Coda AI.
Instruction-following typically describes the ability of the LLM to generate and execute small code fragments. Perhaps – and I only say this with pure speculation – this aspect of Arctic could be used to generate Pack code, or CFL.
Certainly, the data analytics in Arctic seems spot on if we have hopes of using it to understand data inside of and synchronized with Snowflake.
Blending Arctic with other models would be necessary and maybe that’s what’s coming soon in an AI-related announcement.
will this allow to move the data in the external platform?