Arctica's Unofficial Snowflake Pricing Calculator was designed to help you estimate your expected Snowflake costs. Whether you’re already using Snowflake, or merely looking into it, the calculator will allow you to understand what you’re paying for and adjust your Snowflake usage patterns to reduce it’s anticipated costs.
Using the calculator is fairly straightforward:
Define your use cases
As described in the calculator page, you are paying for credits used. To estimate how many credits you need, we recommend that you start by defining some common use cases.
For example: assuming you have a data engineering team who runs ETL processes, another team of analysts who sporadically query the data, and a squad of data scientists who execute complex queries.
Estimate you usage duration
To calculate Snowflake’s credits consumption rate, based on these three use cases, you first need to estimate how much time each team will be spending actively using Snowflake.
How many times a day do you plan to run your ETL jobs? For how long will it run? How many analysts will query the database? Do they work in various time zones?
Admittedly, the "how long" question is harder to answer, but as a ballpark estimation you may assume that it will run about x2 faster than your current cloud based processes or x4 faster than your current on-premise solution.
The following screenshot shows an ELT process that runs at every hour, seven days a week. The process run time (duration) is estimated at 10 minutes.
As you can see, the number of monthly consumed credits is 122 credits, which is 16.6% of the time.
The size of the warehouse
Another important parameter is the warehouse size — meaning, the computing power you need. This obviously depends on the volume of the data and the complexity of the ELT.
It is highly recommended to run a short benchmarking POC, to determine the minimal warehouse size that satisfies your ELT needs.
The X-Small warehouse may surprise you as it is powerful enough to handle hundreds of millions of records — there is nothing small about it! Rumors say that Snowflake was asked to release an even smaller warehouse.
Snowflake compresses data by default in an efficient way. As a rule of thumb, you may roughly estimate that 5TB of raw data will be compressed and stored as 1TB.
Note that you also need to consider some advanced data features such as "Time Travel" and "Fail Safe", that might be heavy storage consumers (mainly for backups and failure recovery).
In any case, the storage component is usually not a significant part of your Snowflake monthly costs.
Please note! The Cost per credit is different between clouds and regions. Please use the Snowflake pricing page here, choose the cloud, region then you'll see the cost per credit in you region\cloud. Use this cost in the text box below.
The storage cost is based on $23 per Tb for capacity customers and $40 for on-demand.
Please feel free to drop us a message to firstname.lastname@example.org regarding any pricing question you might have. We have a lot of experience in pricing usage estimation from many accounts we are managing on a daily basis.