Reverse ETL - How To Get There
Reverse ETL allows you to pipe modelled data directly into the applications that your business users are already familiar with. It is a great way to activate your data and create actions such as churn flags, dynamic segments and dynamic emails. So how do you ensure you set up for success to leverage this technology?
Strategy - Asking the Right Questions
All too often personalisation projects are started from the point of view of a company looking to sell or upsell something and so they want to know:
- What are customers interested in buying?
- What data do we have on customers?
- What gaps in our data do we have and how can we fill them?
And so from the very start of the process, personalisation efforts have a narrow focus that cares less about the customer’s interests and more about selling them something. Companies are so guilty of finding out very little about their customers other than their basic demographic information and their shopping habits. Instead they should focus on key pieces of information that are really going to make a difference to the way you would talk to a customer.
- What brings them into my category?
- What are the key points of differentiation for my brand?
- What is the simplest question I can ask that will narrow down their interest?
Think carefully about the key pieces of information that are really going to make a difference to the way you would talk to a customer.
If you were a travel company, your messaging would be very different for a business trip vs a romantic getaway. It sounds so simple and straightforward, and yet think about the last time you booked a holiday, they asked you how many travellers but did not ask you your relation to them. Both a work trip and a romantic trip have some very clear opportunities for upselling for the travel company- hence the importance of thinking about those core questions and implementing segments as the first steps. Even by adding a simple tickbox for travelling for business or leisure, they can start to segment you effectively.
Lay out your customer journey and consider which questions would help get customers further down that journey the quickest. Consider how these questions can easily be added into that customer journey.
Start With The Data You Own
Modern customer journeys are omni-channel, multi-device and often sporadic in nature.
People are researching your brand whilst also paying their water bills, posting a selfie and cooking dinner. Mapping out all of this different activity and finding moments for intervention can be difficult, and with increasing privacy legislation may one day become near impossible.
Today, too many marketers are guilty of spending time, effort and money in trying to track every channel and every interaction instead of focusing on the data they already have: First-Party data. First-Party data refers to information collected directly by an organisation from its own sources or interactions with its customers or users.
By focusing on this data, you will have 100% control in the way it is captured, modelled and reported. You can ensure that this data is high quality and highly accurate.
A user id will allow you to have visibility and the ability to connect all touchpoints, e.g. purchases, subscriptions and email engagements etc. The key is to ensure the internal user id is passed to all systems providing and capturing the user activity, e.g. your CRM or sales systems. In order to facilitate this you need a centralised repository for all your data – a data warehouse.
“But I have this data in my CRM?” While our initial focus might be on the data you already collect, the ultimate goal is to get a complete picture. Data warehouses are designed to handle large volumes of diverse data, providing a centralised repository for an organisation’s data from multiple systems. They enable complex data modelling which can help identify and anticipate consumer trends like lifetime value signals or propensity to churn. So how do you set one up?
Setting Up The Right Infrastructure
A data stack should reflect the unique nature of your business. It is comprised of a number of different tools, each with their own considerations, but here are the core things you will need to utilise reverse ETL.
Data Dictionary: The first (and in many ways most important) step with any data project is a data dictionary which captures and defines KPIs across all business functions. It outlines the business and technical definitions for each metric along with other useful information. It is an iterative document that will grow with your business and analytics capability.
Data Sources: A data source is anything which produces digital information. Which in this modern world is pretty much anything! A data source could be a file, a programme, a website, etc. Every organisation uses multiple data sources everyday, and often they have to combine metrics from multiple sources to get the answers they need.
Data Extraction: Data extraction is the process of obtaining relevant data from your source(s) so they can be stored in a Data Warehouse/Lakehouse.
Data Warehouse/Lakes/Lakehouse: All of the data you extract has to live somewhere. Data Warehouses, Lakes and Lakehouses are all different ways in which to store your data, most often in the cloud but potentially on-site as well.
Data Modelling: Data modelling is the process of taking that extracted data and applying your unique business logics to help you organise and easily consume the data. So it might be for example matching all the customer ids which have visited your website but not placed an order. All data models should undergo a process of peer review and data reconciliation compared to its source of to ensure accuracy and data trust. The complexity of these models can vary depending on how many data sources are involved and what you need to model to show. Data modelling is most often done by data analysts, analytics engineers or data scientists.
Reverse ETL: Takes the data models and send them back into your applications. This could be as flags against users or segmented lists. Reverse ETL makes activation a lot easier but only if you have built solid data foundations.
Conclusion
It sounds like a lot of work, but Reverse ETL is a great tool that can help tie together data from across systems, makes automation easier and helps to personalise at scale. If you are seeking a partner who can help you implement a data stack and get started on this journey, look no further than 173tech.