We’re still around that time of year when everyone is publishing articles on upcoming trends in data for the year ahead. One of the topics that seems to be getting some consistent focus is Reverse ETL (Extract, Transform, Load). So, what is it?
Reverse ETL is the process of taking data that has already been loaded and transformed in a central data platform and sending it to (typically) an operational system so that it can be acted upon.
An easy example to wrap your head around is in the world of marketing. You bring together a lot of different data about your customers into a warehouse, use that data to create a customer segmentation model, then send the segments to a marketing platform to drive targeted marketing campaigns.
Now if you’re reading that definition and thinking this isn’t something new, you’d be right! Reverse ETL has been around for a long, long time. So why is it coming back into focus? Well, I believe that’s driven by two predominant factors.
Firstly, it’s the ever-increasing maturity of the market. Business leaders understand they need to use data to gain a competitive advantage (or even just stay relevant). Firms such as Netflix, Facebook, Spotify and Amazon have all shown what can be achieved with data-led products and services. Just having some dashboards is no longer going to cut it.
Secondly, it’s technology. Tools such as Fivetran and Stitch have popularised simple connectors for bringing data into a platform, and now the same is happening for getting data out of a platform and back into those same sources. There’s an emerging pack of vendors that are focusing on this space and the key thing they all have in common is connectors – the bit of the solution that’s going to make it easy to plug in to your favourite tools. Vendors in this space include Seekwell, Census, Hightouch and Grouparoo. It’s still early days, so watch this space as regular ETL vendors put time and energy into extending their products to support these kinds of use-cases.
Ultimately, the (re)emergence of Reverse ETL is a good thing. It is one of the most effective ways to realise the value being created in a data platform.
Data that doesn’t leave a data platform can only go so far when it comes to enabling a business. While it might deliver a range of incredible insight, in practice most of the time that data will end up in a report that is used to support teams make informed decisions. The downside being those decisions then need to be acted upon by a human – there’s no automation and the time lag between insights being available and action being taken can be significant meaning missed opportunity.
With reverse ETL, insight being created in a platform can be shared with tools (and people) across the organisation to help level up their capabilities. Whether it be sending a prediction to drive a marketing campaign, providing up to date insight for a call centre representative or automating an otherwise time-consuming process in finance, by taking data out of the central data platform and into the systems that run the business its full value can be realised.
Sounds great right? And it is. Sharing data out of a platform is a genuinely great way to realise its value, but if mismanaged it can come at a cost.
As soon as you integrate your data platform into operational systems and make it part of business-critical workflows, you change its very nature. Done badly, this can manifest in a few ways:
As long as these kinds of considerations are thought through, they can all be managed effectively and shouldn’t stop you considering where Reverse ETL can support your use cases.
So, is Reverse ETL going to be one of the big trends of the year? I hope so! It’s one of the best ways to get value out of data platforms and is a great way to drive adoption. If you’d like to learn more about how this can help you then get in touch and drop email@example.com a line.