Replatforming is a large undertaking that shouldn’t be taken lightly. But if you’re asking yourself questions like:
- What percentage of web users make purchases?
- What is the impact of our paid advertising?
- How can we track the full customer journey?
Then it’s probably time for you to replatform, to unlock the potential of your data.
Getting business buy-in
You may understand the purpose of replatforming, but conveying its value can be a challenge. Data is often taken for granted as a by-product of running a business, and is rarely seen as a primary driver for transformation. This means data often accumulates over years, siloed in different areas of the business where its benefit is localised and limited.
Therefore, to get stakeholders to commit to replatforming, you’ll need to demonstrate its ROI and show how an effective data strategy could benefit different areas of the business.
During our recent digital transformation journey with Missguided, replatforming played a significant role in a series of changes made across their estate that culminated in a brand-new app. To find out more, tune into our webinar on March 22nd, How Missguided unlocked growth through modernising their digital strategy. Kilburn Club executive, Nathan Warner, will be joined by Missguided’s IT Director, John Rignall, who’ll share the steps we took to develop a composable architecture, helping them achieve digital growth across the business.
How do I know which platform is right for me?
There are many platforms available – each with their own perks and pitfalls – so navigating this arena can be challenging without the right guidance and support. When you find a suitable option, you’ll need to make sure it meets the individual requirements of your business. Key factors to consider are:
- Cost: Will the platform be cost-effective in the future? Are cloud services better value?
- Scope: Does the platform meet our objectives? Can surplus features be removed?
- Function: Does the platform support all the requirements of the product?
- Skillset: Do our engineers have the required skills to manage the platform?
- Support: What training does the provider offer to solve technical problems?
How will data be collected and analysed?
The next step in your replatforming journey is understanding your data and its ecosystem - the location, complexity, inputs, outputs and ownership. Some of the main data sources to consider are:
- Internal databases
- Archived data (data warehouses)
- Ecommerce data
- Stock inventory (SAP or warehouse data sources)
- Web traffic (Google Analytics)
- Look-up data
- Audit / metric data
Once you’ve identified your data sources, you can then determine the links between them – such as linking transaction details from web traffic data to order IDs in the sales data. Without these links, pockets of data are isolated from each other, so it’s important to build connections into your product for future analysis.
You can imagine this process like a cake. Each data source is a component of the final product. The icing, jam and sponge taste nice in isolation, but together, they form something completely different and significantly better. Continuing with this analogy, you can also create different combinations of data – different flavours or situations for your business and requirements. You don’t have to stick to the recipe!
Securing a safe location
As well as identifying the data and its links, you’ll also need to determine the location of the data and how it’ll be transferred to the central solution. Is data directly accessible? Will it need additional security or infrastructure to enable the link?
Security is another important agenda item for any replatforming project. You’ll need to ensure data is secure from unauthorised access and that sensitive data is protected throughout the process.
Finding the right replatforming solution
One of the most common replatforming solutions is the AWS Data Lake, comprised of:
- S3 (data storage)
- Glue (data processing and cataloguing)
- Lake Formation (data security)
The various data zones within the Data Lake can be formed using several S3 buckets – one for raw data ingestion, one for processed data, and a final bucket for refined data. Glue provides a mechanism for defining databases and tables as part of its catalogue. Unlike databases with fixed table structures that require complete data reprocessing to change, the Glue catalogue is a simple representation of the data in S3. This means, in order to change the data table structure, you can simply drop and recreate the table without affecting the underlying data it represents. Using the Lake Formation, you can then control user access, set permissions and define security policies.
An alternative approach is tag-based security. Databases, tables and columns are given a series of tags to group the resources. These tags are assigned to users along with security permissions. When tags are updated or further resources are allocated, the user permissions are updated automatically.
Transitioning to the AWS Data Lake
The AWS Data Lake approach was particularly useful for Missguided. They wanted to unlock the potential of their customer data to help drive sales, optimise advertising and improve the user experience. To do this, they linked various data sources together to identify patterns and harness valuable insight from across the business.
Missguided already had Google Analytics set up, but despite providing insights into web user behaviour, it didn’t track actual sales. By linking their sales and web traffic data sources together, Missguided were able to understand which marketing campaigns led to the most conversions, determining the value of each advertising stream. They also linked customer service data to better understand the user journey: which users were having problems, where, and how to help. They integrated sales and warehousing data to identify popular product trends, identify consistently returned items and manage stock levels. It gave them the insights to enhance online sales, drive digital growth and delight their customers.
