Over the last year, COVID-19 has accelerated digital transformation
Retail is one of the most obvious places you’ll have seen this. Since the onset of the pandemic, traditional bricks-and-mortar brands have rushed to optimise their digital offerings, boost home delivery and launch ‘click and collect’ services as consumers have flocked online. In July, a landmark McKinsey report concluded that in just three months, we’d seen 10 years’ worth of e-commerce growth.
It’s a trend that isn’t likely to reverse. The digital customer experience is rapidly shifting from being a competitive differentiator to the key to a company’s survival in a new, digital-first economy. To deliver it, businesses need robust customer data and analytics – and that means getting the right set-up in place.
The power of customer data
As digital adoption has increased, so has the volume of customer data that companies now have at their fingertips.
When used thoughtfully, and with customers’ privacy and preferences in mind, this can enable some powerful applications – like delivering exactly the right content to the right customers at the right moment. Just think of Netflix, where over 75% of viewer activity is driven by film or show recommendations generated using customer data. This kind of personalisation can help ensure that a brand meets and exceeds customers’ growing expectations, as well as helping to build brand loyalty. In fact, recent surveys suggest that customers are more likely to buy from retailers who address them by name or offer them items based on past purchase history.
Some businesses are going even further, using machine learning, modelling and statistical analysis of existing customer data to predict future behaviour and buying trends.
In short, customer data is powerful not just because it tells us not only who customers are, but also what they do.
Designing a robust customer data architecture
Whether you’re building your data architecture from scratch or are revamping an existing data stack there are a few basics to keep in mind:
1. Take advantage of existing tools. To really take advantage of the data revolution, your business is likely to need a range of analytics tools that allow your teams to make sense of your customer data.
Building these tools in-house can prove a huge sink of time and money, so it’s generally better to opt for ready-made solutions. Doing this frees up your engineers to focus their time on your own digital products, and means you can chop and change tools as your company’s needs evolve.
Combine independently-made tools to create a bespoke data stack suited to your business. For example, you may want to try out customer analytics tools like Mixpanel or Amplitude or chat software like Drift or Intercom. As you layer on new tools, make the most of free trials so you can experiment with what works for your firm’s specific needs.
2. Build a strong data foundation. Whatever your tools, they’re only as good as the data that feeds them – so when building any data architecture, you need to pay attention to the foundations.
Customer data platforms (CDPs) are the way to go for this, as they centralise, clean and consolidate all the data your business is collecting from thousands of touchpoints. They coordinate all of your different data sources – almost like the conductor in an orchestra – and channel that data to all the places you need it. As a central resource, a CDP eliminates data silos and ensures that every team across your company has live access to reliable, consistent information.
CDPs can also segment customer data – sorting it into audiences and profiles – and most importantly, can easily integrate with the types of analytics or marketing tools already mentioned.
CDPs are often seen as a more modern replacement for DMP (Data management platform) and CRM (customer relationship management) systems, which are unsuited to the multiplicity of digital customer touchpoints that businesses now have to deal with.
3. Plan your data storage. When you have the basics in place, deep learning and artificial intelligence can allow you to go further. These cutting-edge applications learn from existing customer data to take the experience to the next level, for instance by automatically suggesting new offers based on past behaviour.
Taking advantage of these game-changing tools requires large repositories of data, and that means that effective storage like data lakes should be an essential part of your architecture. Data lakes store the raw, unorganised data that is essential for many predictive analytics tools to function. By providing deep access to historical customer data, lakes also allow your business to uncover year-over-year trends.
Data warehouses, which house structured, schematised data, are becoming increasingly outdated as the amount of customer data being generated continues to grow and customers expect more of a personalised experience based on years of loyalty. Although they have their place in a data architecture, data warehouses are not flexible enough to fuel AI-based analytics or martech tools with the data they need.
Ultimately, the best option for your business depends on how you intend to use your data.
Taking the next step
There are many ways to design a data architecture, but one thing’s for sure – yours needs to be built with your business in mind.
No matter the tools you intend to integrate, start with a foundation that gives you the flexibility to rebuild your entire data stack every few years. This will ensure you have a scalable solution that matches your priorities, and doesn’t lock you into tools that you just don’t want or need.
The key? Shop around, and make sure you’ve found ‘the one’ before you get to work.
About the Author

Tido Carriero is Chief Product Development Officer of Segment, a customer data platform. Segment helps companies harness first-party customer data. Our platform democratizes access to reliable data for all teams and offers a complete toolkit to standardize data collection, unify user records, and route customer data into any system where it’s needed.
Featured image: ©Monsitj