Hello world! Welcome to the Mozart Data blog!
For our inaugural post, we have Mozart Data co-founder and CEO Peter Fishman discussing what defines a modern data stack, and how Mozart can assist companies trying to spin-up their own.
WHAT IS A MODERN DATA STACK?
Peter Fishman: A modern data stack is the set of technologies that enable a true data pipeline. A true data pipeline involves:
Extracting data from many siloed sources into a single data warehouse.
Transforming that data to make it clean, consistent, and accessible to the internal business user.
Connecting that data warehouse to a business intelligence tool to meaningfully visualize the data and guide business decisions.
Industry experts frequently refer to the first two steps as ETL (Extract, Transform, Load).
More directly, a modern data stack is about using world-class technologies that enable data analysts to quickly consume data from many different databases and SaaS tools, and make sense of it.
Data on its own has some value, but it’s clearly the case where the whole is greater than the sum of the parts.
When you’re able to combine data from, for example, your sales source of record, to your production system activity, you now know what your customers were doing as your sales changed. If you only have one of the pieces, you can do basic reporting — you can know the health of your business, how many sales your business is making, and how much customers are using your business. But what’s so much more powerful, is knowing what the probabilistic outcome is going to be in advance of a customer becoming unhealthy or a sale not closing. You get a lot of power out of combining data.
The tools that define a modern data stack enable people to combine data from multiple sources and visualize that data in a way that’s going to change how a company does business. Understandings like:
When and how to proactively reach out to customers.
How to better forecast what the company is selling, winning, or losing.
What the most effective marketing and sales channels are.
All of the things that people orient their business around by sometimes sticking their finger to the wind, now you can do those with a much better set of information and understanding.
A modern data stack is about being data-driven. Being data-driven is about actually deriving value from data, not just simply reporting it.
Very often, people want data so they can see a time series of their business going up and to the right. In practice, that never provides a lot of real value to companies. What does provide real value is when you’re able to understand specific happenings of your business, so you can double down on what’s going well and correct the course on what’s going poorly.
You need to be answering questions like:
Which of my ads channels are most effective? Should I invest more in LinkedIn or Google or Facebook?
Do I need to invest more in the product? Is the product not yielding the right conversions?
Are we not having the stickiness that we need?
The modern data stack needs to end where people are taking behavioral interventions. A lot of people think it ends with the reporting — it doesn’t. It ends with a change of behavior. A modern data stack is all the tooling that’s going to best enable people to do that.
HOW DOES MOZART DATA HELP ENABLE A MODERN DATA STACK?
Mozart Data is the easiest way to get spun up on a modern data stack. We manage your ETL, your data warehouse, and the scheduling of transformations.
Out of the box in one hour, you’re set up with that exact modern data stack that we’re talking about — the ability to get data from different sources into a central warehouse, clean the data up, connect to your favorite BI tool, and start automating reports, alerts, insights, in no time.
Thanks to our friends @MozartData who make having a data warehouse easy. We never could get over the hump and have professional data analytics for Gaia until they showed us the way.
— Andrew L. Johnson (@andrewljohnson) October 28, 2020
We’ve made a lot of technical choices for our customers that we think are top-notch:
We have a best-in-class ETL.
We’re leveraging Snowflake and Fivetran under the hood.
We’re doing all of the necessary researching legwork, negotiation, and technical set-up with each of the data pipeline service providers.
We also know something that’s a secret in the industry. People think data scientists and analysts spend most of their time studying and analyzing data because it’s in their title. That’s not at all what they do.
There’s a joke that I’ve always loved from my old boss, which is “95% of what a data scientist does is clean up their data. The other 5% is complaining about cleaning up their data.” Almost none of their work is actual data science.
With that kind of insight, we realized that data folks spend most of their time curating data, documenting data, cleaning up data, making tables accessible, and getting consensus on definitions. That’s the main function of Mozart. It’s the key place where you orchestrate all of this data. You’re cleaning it, you’re documenting it, you’re making it intuitive, and you’re making it accessible. That’s the place where most data folks are actually spending their time. The analysis comes at the very end.
It’s like building a house.
I’ve recently done some volunteering with Habitat for Humanity, and for almost all of the homes built, 90% of the work is prepping materials and placing the foundation. So little of it is actually building the walls.
The same is true with analysis. The pieces of analyses that become so key and critical to the business are only a small fraction of the end state. It’s just the tip of the iceberg. The foundation of the end state is having clean tables that are reliable, trustworthy, and give consistent answers. What Mozart is building is a platform that’s going to enable companies without a data engineering team to actually do that.
Could your company benefit from spinning-up a modern data stack? Request a demo to learn more about how Mozart Data can be the foundation of your data pipeline.
If you’re interested in improving the state of data tooling — we’re hiring! Check out the open rolls on our careers page.