For years, extracting, transforming, and loading data (ETL) from applications has been the modus operandi for feeding data into reports, analytics, and business intelligence (BI) tools.
Talk to any engineer, though, and you’ll soon find that traditional ETL tools are clunky and expensive. Connections need to be built for each data source, data modeled, warehouses setup. Some days you end up manually writing scripts and pushing out config files. Other days you’re Googling the hundreds of products that profess to do ETL for you, but end up being about as infuriating and byzantine as if you were doing ETL with a quill pen and carrier pigeon. Analysts end up exporting .csv files, formatting pivot tables in spreadsheets. Traditional ETL takes a long time and is, in all honesty, pretty boring to do “by hand”.
ETL ≠ Modern ETL
What exactly do we mean by "Modern ETL"?
Well, we should start by stating that "Modern ETL" isn't really ETL at all. Nor is it what tech pundits have dubbed “Cloud ETL", tools that pull data from APIs only to load that data in another location. Both are oblique euphemisms for data migration, for moving data from one endpoint to another.
With ETL, you're usually stuck dealing with a morass of different datasets. And with Cloud ETL, you're still tasked with writing for APIs. Which isn't easy. Each API has unique authentication rules, its own set of gotchas. At any time they can change and break all you've written downstream.
So if ETL and Cloud ETL are baroque terms for transferring data, then we can safely say that Modern ETL is everything ordinary ETL is not.
Modern ETL: Better, Faster, Stronger
Modern ETL is much simpler. From within a clean UI, you may select which data sources you would like to combine. Object mappings you can either customize or opt for the default. Then a data warehouse is built, stacked with tables whose database connects to any BI tool.
Modern ETL means the APIs for your SaaS applications are written into a unified schema. Developers can forget having to make sure every name has a capital letter, or that dates are in DD-MM-YYYY format. You have an entire data pipeline whose deep knowledge of objects and their relationships has been automated.
For analysts this is a dream come true. Just having disparate data sets merged is a huge win. But having like records matched and conflicts resolved? That’s magical. And because the data is in near real-time, anyone with a board meeting tomorrow can save reports for their top KPIs and hit refresh on their dashboards.
Modern ETL is the process by which Fusion operates.
You authenticate your SaaS connectors. The API to your connector recognizes Bedrock as a trusted partner for your database. Fusion replicates data from these diverse source systems, translates data types into a common format, transforms many data models into one, and loads this consolidated dataset into a cloud data warehouse so any BI tools can swiftly analyze and visualize all of your cloud data.
As a Modern ETL tool, Fusion supports a diverse range of data sources. It can handle large volumes of data at high velocity, and accommodate user types from CEO to sales managers. Anyone can change default mappings. If your system used c_name instead of first_name, you could change this mapping in the UI and brew a cup of coffee while your warehouse rebuilds. And for enterprises with an in-house system, soon developers will be able to write their own connector with the Fusion SDK.
Modern ETL is opens up a horizon of opportunities. You can:
- Get a trusted, single view of customers
- Create dashboards using data across the business
- Easily navigate closed-loop reporting
- Perform ad hoc queries for the analysts on your team
- Get insights on generating profitable & successful customers
- Analyze all of your Activity data in one place
Ready for Modern ETL? Then give Fusion a try.