Snowflake AI Data Cloud

Our main use case for Snowflake was to ingest data, use it for search, analytics, and data sharing with our customers; however, in reality, those use cases did not work.

When it came to our challenges with those use cases, Snowflake oversold their capabilities dramatically and told us we could simplify our Databricks implementation. In the end, this was confirmed by Snowflake architects that Snowflake ingestion would not work for our use case because Snowflake is not sophisticated enough. This resulted in us losing significant time, spending considerable money, and ended up with a product that was not a good fit for us after making a substantial financial commitment. When it came to search, we thought we could use it based on the initial conversations with the sales representatives, but after we purchased the product and made the capacity commitment, we were told that capability was not available. The end result was that we have a very expensive data lake that does not support things like Iceberg as it should.

Snowflake's data sharing capability with other enterprises is the best feature they offer, which is excellent for us as a software company. Everything else they offer can be done better with other platforms at a cheaper rate. Snowflake's data sharing capability stands out compared to other platforms because it is native to Snowflake. If your customer has Snowflake, it is valuable. You can send your data into Snowflake and then the customer receives it automatically.

However, Snowflake has been the biggest mistake we have made since we started the company. It was a drain on the team from a technical perspective and almost from an emotional perspective. A specific example of how it was a drain on the team includes that we were promised a significant amount of things pre-sale around access to solution architects, support to get us on the project, and technical abilities to simplify our infrastructure. All of these technical abilities were significantly oversold. When it came to setting up meetings and getting help, it took months to get meetings with people who could actually assist us. In the end, when we got the meeting with someone who actually had the technical ability to do the work, he confirmed that they could not actually help us. Since then, Snowflake has only held our feet to the fire for the capacity commitments we made on the commitments that they made to us pre-sale.

Lessons learned about our main use case or how Snowflake performed in those areas include that the sales representatives do not really understand the product, and that the solutions consultants basically read internal marketing materials and internal AI to tell you what to do without really knowing how the things work. A specific example is when they told us they fully support Iceberg, and then they have over three thousand words of bullet point text with a link to reference articles with all the limitations of their Iceberg support. This resulted in us having to build almost significant modifications in order to use Iceberg.

Snowflake can only be a data lake to share data with external people and external companies. Beyond that, I believe you are wasting time and money.

We have used Snowflake for three years.

Snowflake is stable.

Snowflake's scalability is such that customer support is an atrocious experience, filled with many challenges, whether it is getting a meeting to fulfill the promises they made, getting them to prepare for that meeting, or getting them to actually help or have the technical abilities to assist. We receive a lot of corporate marketing sent to us, for instance, claims that they fully support Iceberg, but ignoring the three thousand words of limitations that they do not actually support Iceberg. It was always a challenge to figure out what they really could do versus what they told us we could do. The process to use their technology was always one that was solely driven by sales and not by Snowflake's actual ability to deliver. We have encountered this repeatedly over the last three years.

Customer support is an atrocious experience, filled with many challenges, whether it is getting a meeting to fulfill the promises they made, getting them to prepare for that meeting, or getting them to actually help or have the technical abilities to assist.

We did evaluate a couple of other databases and data lakes before choosing Snowflake. I cannot remember what they were, but this was back when we were on Azure as well.

Deploying Snowflake in my environment is really easy to set up. If you want it to do sophisticated things for a SaaS company, it does not do them.

Snowflake's configuration process is straightforward and very easy.

In the end, Snowflake cost us more money. One of the things they said they could automate or improve for us was the ingestion of data and eliminate our Databricks requirements. In the end, we did not move to Snowflake to do ingestion, but they admitted that they could not actually eliminate any steps and would actually add steps to the process and add costs to the project. Snowflake undoubtedly cost us more money and provided almost no value to us.

We find the metering and billing experience for Snowflake to be expensive because there are many limitations to what it can do. When we compared performance-wise and cost-wise to Bedrock for some of the AI work we are trying to do, Bedrock was much more performant and better priced.

We found Snowflake's Cortex tools and some of their AI tools to be not as good as Bedrock and much more expensive and slower.

My advice to others looking into using Snowflake is to look at DuckLabs and other tools. I would not recommend Snowflake on any level.

I purchased Snowflake through the AWS Marketplace. I have given this review a rating of one out of ten.

Public Cloud

Amazon Web Services (AWS)

What do you like best about the product?

Ease of use, simple UI, simple integration.

What do you dislike about the product?

Sometimes queries can take a very long time in larger databases.

What problems is the product solving and how is that benefiting you?

Centralized database, integrating with systems that run on old integration architecture.

What do you like best about the product?

I'm in RevOps, so not 100% your typical user. I mostly use Snowflake to pull datasets and stitch things together in SQL: pipeline questions, customer data, whatever came up that week.

The main thing I like is that I can just do it myself. Write the query, share them, look at what comes back, access them via CLI, drop a quick chart on it in the worksheet to see if the trend is actually there. Before, that meant exporting to a spreadsheet or asking someone on the data team, which usually meant waiting a few days for an answer I needed that afternoon.

I don't need anyone else for the everyday questions anymore. For me that's the whole value.

What do you dislike about the product?

Nothing really, I struggled a bit with the new worksheet structure but I got the hang of it now

What problems is the product solving and how is that benefiting you?

Answered that in the first q: clear, structured, sharable, easy to use data access to revenue data, customer, product data etc.

What do you like best about the product?

The way snowflake stores data through clod is the best part in the entire application where all the beginners also can use this snowflake application in easy way without confusion. Also AI/ Intelligence which provides all most all the answers for the user query.

What do you dislike about the product?

The part I dislike most about the Snowflake application is that it doesn’t allow an on-premises option and is only a cloud-based storage platform. Also, cost-wise it feels a bit high because it consumes more credits, and I’ve noticed low latency as well.

What problems is the product solving and how is that benefiting you?

In Snowflake, most of our integration problems have been solved through enterprise automation. It has also helped us address data storage issues in the iPaaS domain within our organization.

What do you like best about the product?

Ease of use, Everything at one place, Good functionality, Great integrations

What do you dislike about the product?

Worksheets shouldn't have been replaced, apart from that everything is good

What problems is the product solving and how is that benefiting you?

Ease of integrations from other platforms for our data