How new technology helped dunnhumby to deliver better business insights

Guest Author: Giorgio Ticinelli , Business Intelligence Manager, Corporate Systems, dunnhumby

dunnhumby: Creating a better world by putting the Customer first

dunnhumby is the global leader in customer data science, empowering businesses everywhere to compete and thrive in the modern data-driven economy. dunhumby empowers brands and retailers to interpret and act on Customer insight at scale and in real-time through more relevant and engaging experiences – ensuring the decisions that clients make encourage Customers to return, time after time.

The need for a better solution for Finance, Sales and HR

For a while reporting for sales, finance, and HR teams was  done out of Oracle Transactional Business Intelligence (OTBI) and Oracle BI Publisher. The reporting was mainly operational and for sales it wasn’t properly configured for our business, but even so it would still have missed the fact that we use additional sales systems in our media teams and that our client teams are tasked on revenue targets. They need to understand our revenue position at any point in time, not just our sales.

To solve this, the Commercial Insight and Analytics team (within Commercial Finance) built a solution in Excel to join several datasets together—from sales and finance systems—and surface the data and insights to our client teams for their daily use.  This provided basic access to pipeline, revenue, and forecast data for anyone that needed it.  

As we moved Oracle Fusion ERP Cloud beyond the UK to all markets, this Excel solution was moved into a desktop database as a data warehouse, with a BI desktop tool as a reporting front end. This worked well for our teams. It allowed us to optimise the processes and the reports, giving our users the business insights they needed to operate, drive sales, and deliver accurate revenue recognition and timely billing.

However, using the desktop database and OTBI, the commercial Insights team would consume an hour a day downloading the data and updating the BI desktop tool, then the sales teams had to download the file, spending even more time. 

Moreover, it did have several challenges:

• The solution worked, but it wasn’t “enterprise” as the process required a manual effort to update the desktop database (at least 1 hour every day).
• The reports were updated manually around 9 am GMT – not ideal from a process point of view, nor for a global business.
• The end to end relied on a desktop database specialist.
• The data eventually became so large that the desktop database couldn’t handle the volume and experienced multiple crashes.
• Manual fixes were necessary to accommodate data gaps.
• The BI desktop tool required file transfers for both updating data, and downloading to view data.
• There were version incompatibility issues between the different BI desktop tool’s file version/software installed onto local machines.

As a result, dunnhumby’s sales teams were pushing for a new intuitive user interface to measure and manage the commercial pipeline. In general, dunnhumby needed a more modern, reliable, and secure analytic environment for accessing its enterprise data from Oracle Fusion Applications.

To solve for these challenges, the Commercial Insights team kicked off a project in 2019 with the Corporate Systems Business Intelligence team to bring our global sales and forecasting reporting up to best practice enterprise standards for Business Intelligence (BI). This project became known as the “Commercial Insights” project.

A new solution with Oracle Analytics Cloud (OAC) and Oracle Autonomous Data Warehouse

Initially we looked to use Oracle Business Intelligence Applications (OBIA), the on-premises BI tool we had in house.  Despite good investment in terms of time and resources, in the summer 2019 the CST BI Team started conversations with Oracle to upgrade our BI stack to the latest, world class cloud services, in particular Oracle Analytics Cloud (OAC) and Oracle Autonomous Data Warehouse (ADW).

These technologies fit well with Oracle Fusion applications. They provide powerful features in terms of analytics, data visualisation and data integration, but they are also easy to use and flexible. Finally in June 2020 we brought OAC and ADW into dunnhumby. Initially we adopted OAC/ADW alongside OBIA, but the power of the OAC/ADW combination was rapidly apparent, and we were very quickly able to shift our Commercial Insights project to these tools, setting ourselves the goal of migrating from OBIA to these Oracle Cloud services. The OAC/ADW combination brought new speed to the project thanks to their power and ability for quick adoption.

The main advantages of the new solution are:

• Enterprise solution based on cutting edge cloud technology powered by Oracle.
• The data is scheduled to refresh automatically overnight and is therefore ready early morning for our APAC teams.  No manual intervention is needed.
• Cloud technology – there is no need to install any software on users’ machines or to transfer any files.
• Users can access information easily and in a secure manner.
• Access security is tailored by region/country (with row level security) and is applied automatically when the user logs in.

Going Live with OAC/ADW

We released our first OAC reports in a soft launch to our client teams in December 2020, with an official rollout of our three main dashboards in January 2021. The old BI desktop tool and desktop database were then decommissioned over the following six weeks, with OBIA dropped shortly after this.

We now have over 250 active users who use our dashboards and reports in OAC. These dashboards can be consumed very easily by anyone in our Client, Product, Tech or Finance teams. OAC is so intuitive and fast that training is very light touch, and adoption is good.”

In 2021 we used our learnings from Commercial Insights to enable a new set of dashboards in OAC to surface our Marketing Campaign and Lead data to our client teams. These allowed us to test connecting to raw CX data (Oracle Fusion CX Cloud) directly from ADW for the first time, and have allowed us to showcase how easy it is to design and build analytics in OAC.

OAC and ADW is a very powerful combination. I have 23 years of experience in BI and I think I've never seen such a powerful combination of an analytics platform that worked so well together with the database. You don't have to worry about installing any software on user machines or manage servers. There's no need to transfer files manually; it only takes one hour to provision with no manual intervention and security is very solid. Our users can access the system in a very secure way and access can be tailored based on role, with row-level security.

Administrators and Users

The client teams are very happy with this solution. It's not necessary to create any training, and we didn't have to shadow any training. We just created a one-page document with the link, so they have started to use it very rapidly. The look and feel of OAC is very attractive and we have configured three types of users:

• BI developers: (part of the BI Team) who build proper data structures and give access to hundreds of users – with limited interaction. Users will interact with filters but are not allowed to make changes.
• Power Users: a group of users with more freedom to connect to any data sources and build reports.
• Simple Users (Viewers): users with limited interaction capabilities (mainly filtering)

With BI developers, Power Users and Viewers it’s been necessary to clarify roles and responsibilities. The BI Team is responsible for maintaining the official data warehouse (on ADW) and deployment of any content (dashboards, reports) in production. Self-service content created by power users that requires transition into production must be managed by the CST BI Team who becomes responsible for its on-going maintenance. For example, at the end of 2021, a new report called New Retail Opportunities, which was entirely created by the sales team, is now live after the BI team’s validation.

Introducing Oracle Fusion Analytics

After going live with OAC/ADW we realized that a new solution, Fusion Analytics, could be very useful for us because of its pre-built data warehouse and subject areas on top of Oracle Fusion ERP Cloud, which is our main ERP system. A key advantage is that you don't have to worry about maintenance and upgrades because they are managed by Oracle. It's very useful to have pre-built subject areas on top of the Oracle cloud database, which gives you the potential to create dashboards very easily without the need for customizing the data structures.

We are now implementing Fusion ERP Analytics and Fusion HCM Analytics on top of our Fusion Applications. We have built a CFO dashboard that shows company performance by business unit, region and cost center. For HCM we are creating a positions report to help with hiring plans, and plan to replace the existing operational OTBI reports within Fusion Analytics, since users like the clean user interface. We have also created a main dashboard for procurement.

Our business users want to see the data in their own way, so we needed to customize the prebuilt solution a bit. Fusion Analytics offers you the possibility to customize it with the data augmentation and semantic extension; for example, we are working on customizations for accounts payable and procurement.  The look and feel is always very good and is consistent with OAC.

Summary

“We now have two main projects, commercial insight and marketing leads that are live and very stable with OAC, where users are very happy. And the deployment of Fusion ERP Analytics and HCM Analytics with business users where we are working on customizing based on business requirements. We anticipate that hundreds of users will use Fusion Analytics once we roll it out, with the potential to provide cross-functional business analysis for multiple departments.”