Chapters: 0:00 Intro 0:22 Buffet's Value Investing Philosophy 1:41 Struggles of Berkshire Hathaway 2:41 What is Snowflake? 3:20 Analytical v. Transactional Databases 4:18 Rise of Business Analytics 5:21 Snowflake Value Prop 6:43 EA Sports Adopting Snowflake 8:21 Snowflake's Top Down Sales Strategy 11:15 Incredible Business Performance 12:38 World-Class Customer Retention 14:15 Monopoly in the Making 15:07 What Modern Tech Startups Get Wrong
Great video. Small error when you talk about snowflake nps, the nps range is -100 to +100 as the scoring is %promoters - %detractors meaning a negative score is quite possible
As a data engineer, the problem with Snowflake is that while everyone wants turnkey big data solutions, there are a ton of competitors in this space that are also building similar things. And this ecosystem is evolving so insanely quickly that if you had a state-the-art data pipeline 5 years ago, it would be nothing but technical debt now. Cloudera for instance looked like it was on top of the world, and then it really struggled. Hard to tell whether going forward Snowflake is going to be able to have the decades long dominance that Oracle had (or if that kind of thing is even possible for any current companies).
I disagree with you. This means you don't understand, ground up redesign of snowflake or the stronger features, zero copy cloning, staging, micro partitions, that use cloud data ware house techniques. Brilliant thinking and benefits customer a lot.
I don't know if it's worth your time doing this full time but if you are on the edge please go ahead. Top notch quality content, no fluff or click bait and condensed, just the way I like it.
Came to say just that. The algorithm led me to the channel today and I'm blown by the quality of the content. I have no doubt this channel will be immensely successful soon.
I would almost bet that it was based on some company he is heavily invested in that found this product so useful that they were willing to bend the rules to look into it. Because before looking at the business prospects they have to be convinced that it fits with their core investing strategy
I got a BA in literature and am now pivoting to communications and marketing. These videos are very understandable, compelling, and have a great story structure to them too. I’m thankful for them
I doubt it was a position Buffett took. It is much more likely that one of the investment managers at Berkshire took this decision. Buffett rarely deploys this little capital.
Snowflake has one big disadvantage, they don't host your data. Furthermore Cloud platforms (AWS, Microsoft Azure, Google Cloud Services) provide or will provide similar services. These platforms charge you to move data outside their platforms. Snowflake takes a multiple-cloud approach where it allows you to deploy it to any cloud platform to avoid these fees. This makes it more limited and adds a lot of complexity to the product Since they can't host their user's data almost all of the revenue Snowflake makes goes to the cloud platform themselves which is why they have such low profit margins. In essence the more the business grows, the more they have to pay the cloud platforms. As far as I know they are actually subsidizing some of these costs for their customers already and everyone complains snowflake is already really expensive. So the more they grow the more they lose money, if they increase the costs to become profitable they will stop growing and might actually lose customers Meanwhile those providers have their own alternatives or are planning to provide them, they are/will be A LOT cheaper. In essence Snowflake is competing with cloud platforms behemoths which are also their main partners (conflict of interests for the cloud platforms as they also offer alternatives to snowflake) I wouldn't be surprised if Snowflake got acquired during this downturn and became single-provider native solution. Cloudflare is probably the one that would benefit the most from it Another way out is if Snowflake becomes a cloud platform themselves, good luck getting the capital for that Without anti-monopoly regulation in this space (to reduce/eliminate the fees to move data out of a cloud platform) there is no way snowflake will survive long-term with its current strategy
I love these! Please please keep it up. You had all the information that was super important to understand how a company runs through fundamental understanding of what makes a business great.
Thank you. This was very nice to read! That is the goal in every episode - to give enough of the fundamentals for others to understand the company, its strategy, and performance with just enough side commentary to introduce perspective without forcing opinion.
@@ModernMBA You did exactly that! Without succumbing to hype or clickbait tactics, you really did a great a job and I'm spreading the word for you. All the best!
Transitioning to snowflake is lengthy and costly. If customers are willing to do this then what's stopping them from doing it again for a competitor of snowflake or alternative technology? I don't get what the competitive moat is here.
Well they would be moving to cloud at this moment - and then they will build their applications around it. Then it get's harder. I think you're missing the point that many tools many companies have that not in the cloud have had a End of Life coming to them. Forcing them to upgrade - they happen to choose snowflake.
@@BachelorChowFlavour That's the argument for almost everything in business. Doesn't really matter about moat as much as the expertise and these guys have shown it
17.00 : This actually happened to my previous companies, where the management acquired a new vendor who provide worse service than the existing vendor, with a higher price.
I am an Oracle DBA for a fortune 500 company and now also a Snowflake DBA as well. In my experience Snowflake leaves a lot to be desired. So much so that I can't image our Snowflake implementation being considered a success long term. There is nothing we are able to do in Snowflake that we couldn't have done in our existing Oracle databases. It is costing us a fortune. We have a lot of resources assigned to the project and we are more than a year into it with no clear end in sight. The product is extremely limited in its feature set which, along with "cloud", is touted as an advantage. However, both the cloud and lack of functionality is actually quite problematic for our needs. I would not be surprised if the project is abandoned and some of the people responsible for that purchase are asked to resign. Snowflake probably has its place and there is a place in the market for it, but it's not a simple as "Snowflake better for reporting/analysis/data warehouse."
@@scj643, we did the same until some VP decided to buy Snowflake to replace a reporting tool (which makes zero sense given Snowflake is a database not a reporting tool). Now we are copying terabytes of data that is already in house over the internet to Snowflake. First we built all the processes to copy the raw data. Then it was deemed too hard/expensive to rebuild all of the warehouse stage to mart transformations natively in Snowflake, so now we are copying the in house mart level data (facts/dimensions) directly into Snowflake. So we have all the raw data and the data marts copied into Snowflake, but all of it is sourced from the existing in house data warehouse. That in house data warehouse can never be decommissioned because it is a source for the Snowflake "marts.". Snowflake data lags behind the in house DW for the same reason. So basically Snowflake is just a mirror of our Oracle DW in the cloud. Much of that is just poor planning and bad leadership making technical discissions about things they don't understand. However, a good portion of it is because we spent a lot of time early on trying to get Snowflake to meet our business needs and then ran out of budget to do it right (not that the budget would have been enough anyway). Snowflake and related tools just had a lot of limitations that made normally easy problems hard to solve. Over a year of work to build a copy of something we already had. Such a waste.
It seems to me like snowflake is a massive risk and no reward. Any company that’s large enough to use it probably already has a DBA and data science capabilities.
Yeah I thought the same thing when he showed the snowflake data architecture that's supposed to solve everything. It's straight up exactly the same thing as using any other product in a cloud environment. Snowflake is just another BI hype that will die out after a few years.
@Ryan Howe there are a massive number of incumbents in this space that do it better and cheaper than snowball. Heck, they’re competing against similar offerings from both Amazon and Microsoft despite also using cloud hosting offerings from those companies. I haven’t used snowball directly, but I watched some of their sales videos. Seemed too complex to use unless you already have competent data employees on staff. Products like PowerBI and tableau put a lot of effort on into making it “business person” friendly. And all the comments I’ve received about it from companies where it was rolled out have been negative. So in short, it’s a mediocre product that everyone hates competing against massive incumbents that it also relies on for operations. Big risk, no reward.
Love your channel and your content. Very useful and documented. Thanks for your effort in putting this together. PS. Small comment on the Snowflake video. The NPS score range is from -100 to +100, not 0 to 100 as said in the video. Nevertheless this does not change the bottom line message.
I found your channel yesterday and it is AWESOME. Right up my alley and super digestible. I do some op model work and snowflake is kinda involved but hearing you explain it makes so much sense
What moat are you referring to? Snowflake has no moat from technology standpoint. Any of the big techs can build a similar BI product. Their advantage is sales and marketing. That can be copied or eroded over time.
It was most likely Greg who made this bet not warren. The guys in the office are more tech heavy since it’s their circle of competence compared to warrens
The startup I work with is trying to fix the problem of having too much data in a database. Really at the end of the day our philosophy is that if you know the content your putting in the database, it’ll make retrieval that much easier
Agreed, but Snowflake (as all other cloud providers do aka Databricks) run on primarily the cloud. Microsoft plays the long game and they started investing on that territory. Will SF/Databricks be around in 5-10 years? Who knows.... But it will be interesting to expect a move from the actual wolves in the next couple of years.
I am working as a data engineer consultant. When we were designing data lake we immediately discarded snowflake because of its limited functionality. Instead we used our own lake design and implemented the entire project using multiple tools and platforms. Most of the consultants avoid snowflake right now as it’a just not mature right now. Maybe it will be in near future but their sales team is really good in terms of accomplishment. I would say on similar line to saleforce and oracle
Transactional databases are actually databases that have to update multiple rows or tables or databases successfully.. analytical databases drive business intelligence and business decisions.. the main two types of databases are SQL and NoSQL which both support the aforementioned solutions for whatever application..
If a transaction fails it rolls back every other update made in all the other places and acts like nothing happened to some extent.. but will log the error and notify the end user and some other stuff
2:12 I think he avoided all IPOs, tech or not. When he wants a private company, he can buy and keep it private. IPOs tends to inflate valuations and that's kinda the opposite of value investing.
I don’t think that, especially the way you conclude, Snowflake didn’t break his rules to invest in snowflake. Snowflake is just a real actual physical value. Data at this scale physically exists and physically needs to exist and be stored. So there is real actual existing value in exactly the same ways as Heinz or Apple.
Competitive moat? How? Am I the only one screaming PLTR here? Not to mention Azure, Google Cloud, and AWS’ own services atop cloud, as well as other smaller players such as Cognite, C3ai etc.
Well, the problem with the video is that it's not true. The Snowflake investment was made by Todd Combs coz Geico works (ed) on Snowflake platform and Todd was impressed with it. Frank Slootman, the Snowflake CEO, narrated that story
Great analysis and facts! As a data engineer working on a data warehouse snowflake blows every competitor out the water (Amazon and Google). Its just a better product by all aspects!
Never seen your videos before, not sure if you have a stutter or it’s the way you cut your videos, but I’m getting some crazy Rick vibes listening to you
AWS is a sprawling ecosystem where you have to figure out how to configure it. Snowflake gives you a way of interacting with AWS in a way that is fine tuned to your companies needs. Snowflake actually uses AWS, Microsoft Azure, or google as the physical infrastructure of the cloud.
One can overpay for a good business. the retention rate was measured against the backdrop of free money. We need to see how it fares this and next year.