Here's ..... Fivetran

Ever marveled at a cool dashboard in a meeting? Or wanted to know how data points from different tools relate to each other—like the likelihood of someone to buy from your online store after seeing an ad on Instagram? There’s a whole lot of work that goes on behind the scenes to get those visuals and insights to you. It starts with all the tools and platforms you use to generate the data. For example, if you are doing ROI analysis, you’ll need to bring in ad spend and impression data from all of your ad serving platforms like Facebook, LinkedIn, Google Ads, and so forth. 

You want to bring those data points into your data warehouse so that you can run fast, reliable analysis. This is the basics of ELT (extract, load, and transform - also known as ETL (see link below to "ETL vs. RLT: Choose the Right Approach for Data Integration") which flipped the last two around. Though, the industry is shifting to ELT as it allows for better, more scalable analysis).

In case you missed the Fivetran session at Great Data Minds' 2nd Annual Technology Matters Marathon on March 10th, I detail some key things you should know about this amazing technology. After reading this, let me know if you would be interested in either

• talking about Fivetran in more detail
• seeing a demo, or
• even learning more about how why a modern analytics program is critical to monetizing your data and how Fivetran fits into a modern analytics program.

Why Fivetran?

Getting the data out of those sources and into your data warehouse or data lake is complicated work most often done by highly skilled and highly paid data engineers. Yet, the work is often described as tedious and menial because the work is often deep in the weeds on API updates across so many different platforms. One small change on Facebook’s side, could, for example, break your pipeline and then all dashboards associated with that data. 

By outsourcing this work to Fivetran, companies can save massive amounts of money (see link below to "Neat Saves Engineering Resources, Amps up Analytics") on hiring new data engineers just for pipeline maintenance—not to mention high turnover of those valuable employees you do hire who want to work on more challenging projects. Freed from these manual tasks, engineers can do some pretty incredible data modeling (see link below to "Tide Drives Business Results with Customer Data") and dashboard building, allowing your company to find new insights and areas of efficiency.

What does Fivetran do?

That’s where Fivetran comes in. Fivetran is an automated data pipeline. What this means is that instead of a team of data engineers building and maintaining the pipelines, Fivetran does it all. The tool automates pulling data from your different sources and replicates it into your data warehouse or data lake. 

In addition to making the extract and load part of the ELT equation easy, Fivetran also has a host of free dbt packages that you can use for transformations. A recent example is the ad aggregation package (see link below to "Aggregate Ad Data With Our New dbt package) that takes all of your ad data and rolls it up for you so you don’t have to do all of the data modeling to find the impression, click, and spend data across all of your major advertising efforts. 

How does Fivetran do it?

With a lot of hard work and a big global team! But in all seriousness, there is a lot of thought and dedication that goes into building and maintaining data pipelines in a way that allows customers to take a really tedious and time-consuming task and make it one that is truly ‘set it and forget it.’

The building of their own connectors is really key. While some of their competitors claim a greater number of connectors, those are built and maintained by the grace of the community. Fivetran on the other hand, has a team of engineers and product managers building out the connectors in sync with the actual source companies. As changes are made to the data sources, Fivetran stays ahead of them and is constantly working to make the syncs faster and more reliable. In Q4 2020, for example, they achieved 99.9% uptime with their connectors. 

But it’s not just about pulling the data, it’s also about how it lands in your data warehouse. For that, Fivetran uses thoughtful, research-driven schemas and ERDs. This means you can easily see how the data will land and ensure you will be able to quickly find what you need to do your modeling and analysis. As schemas change in the source, Fivetran is ready for that with automated schema migrations. 

What makes Fivetran different?

What we’ve covered so far, honestly, is a lot of what makes Fivetran different in the market. Their breadth and depth of connectors along with their thoughtful approach to how the data lands makes them in a class of their own. Getting a connector set up is a breeze, taking as little as five minutes.

In addition, Fivetran provides a free 14-day trial (see link below) of its platform and for any new connector you add after your trial, you can test that out for 14 days of free usage (see link below). 

Fivetran’s CEO, George Fraser, puts it best (see link below to "Fivetran Doubles Revenue and Customers in 2020"): “Our strong business momentum is the result of two simple facts: we provide always-available data pipelines that just work, and we stand by our word to make access to data as simple and reliable as electricity. To put that in perspective, Fivetran manages more than 700,000 schema changes every month — events that commonly break legacy ETL pipelines and need to be fixed by a data engineer. This is the value of Fivetran — delivering accurate, timely data even when source systems change, which ultimately powers analytics teams to make better decisions.”

So, let me know if you would be interested in either

• talking about Fivetran in more detail
• seeing a demo, or
• even learning more about how why a modern analytics program is critical to monetizing your data and how Fivetran fits into a modern analytics program.