Data analytics is a tool used to better understand the end-user and ultimately enables businesses to optimise performance and achieve success. We hear from Petra Kasperova, Insights and Analytics Director, Dreams, and James Stewart, ETL/ELT specialist within the Dreams data insights and analytics function, to learn about how the Exasol collaboration has driven transformation for the retailer and enabled it to better understand its customers’ needs.
Can you tell us about your role at the company?
Kasperova: I am the Insights and Analytics Director at Dreams. Prior to this, I was heading up the Dreams finance department function. Over the last two years, Dreams has invested in analytical software (Alteryx/Tableau) and resources to start a journey of in-house analytical enhancement, centralising the analytical team and building the data analytics strategy. The function today represents six analysts and one primary technologist, James Stewart, who managed the implementation of Exasol in Dreams.
Our key goal is to move away from traditional static Excel reporting and towards building a fully automated suite of pre-defined self-serving cross-device dashboards, providing faster, more accurate and deeper insights to end-users; utilising advanced analytics tools and improving our analytical maturity through those as well as resource upskill. At the same time, the aim is to build, grow and maintain centralised data lake of verified internal and external data sources, overviewing the data capture processes and automation of any data-related tasks.
Stewart: I’m the ETL/ELT specialist within the Dream’s Data Insights and Analytics function responsible for enabling our three technology pillars: Exasol, Alteryx and Tableau. I’m responsible for administration and maintenance of all three platforms, as well as the ingestion of data sources into the Exasol database, and optimisation of ingestion from a varied range of platforms together with providing specialised technical expertise to the analytical team.
Why did you select Exasol as the technology vendor?
Kasperova: Historically, Dreams’ internal analytics has been produced within Excel and to a large degree on an ad-hoc high-level basis. As such, Dreams’ reporting functions didn’t have real need for an analytical database. However, with the changing landscape and technological advancement goals, and with greater volume and variety of data being generated and captured, this has certainly changed requirements for Dreams.
We continually strive to be better, more innovative, more technologically advanced in all that we do and this applies to the technology we invest in to support the business needs. Exasol has been recommended as the ‘go-to platform’ in combination with the other platforms that we use: Alteryx and Tableau. The impressive performance stats followed by a successful proof-of-concept trial have clearly identified the benefits we could gain post-implementation.
Stewart: We selected Exasol as we believe it to be the best-in-class analytical database. From a technical perspective, what we loved about Exasol is its fast performance, but with that it is also extremely important that performance continues to be maintained. Dreams doesn’t have the luxury of a large pool of experienced DBAs to call upon, so we are reliant on Exasol managing its own performance tuning, and so far Exasol has delivered on this aspect with no manual tuning required.
The use of commodity hardware and Exasol’s licensing options enabled a cost-effective entry point to an analytical database, which we have previously lacked.
With significant reliance on the ERP data, and limited resources, we also required a database that could easily integrate with our ERP, so we welcomed Exasol’s Virtual Schema. This has enabled efficient ERP data accessibility without the need for complex programming, resulting in self-sufficiency of the insights and analytics team and reduction of overall timeline and cost of the project, leading to democratisation of the data.
The post-sales support from Exasol has been outstanding. It has proactively maintained a strong relationship to ensure that Exasol’s solution is delivering Dreams’ goals, enabling us to identify and prevent any concerns early on and to enable mutual understanding of our Exasol implementation.
How important is having a robust data strategy in satisfying end-user demand?
Kasperova: All decisions at Dreams are based on data and analytics. From investment into our store portfolio, selection and production of innovative product, design of our website, improved delivery services, to offering a drink to our customers as they shop. All insights and analytics are aimed at improving our understanding of our customer and their needs, as well as improving the understanding of our business to provide excellent service and better sleep for all. We believe that democratised data and analytics empowers our colleagues and helps cross-functional effort, as analytics is enforcing communication among teams. While less effort is spent on data collation and preparation, more time can be invested in gaining and sharing the knowledge.
Data strategy is a key influencer of business performance, assisting in increased competitiveness as we identify risk and opportunity early.
Stewart: Through data lake build, we are implementing a strategy of a ‘Single Version of the Truth’. This is to provide more accurate and timely data, enriched for internal and external sources and to enable automation and greater accessibility across the organisation.
As result reporting and analytics is taken very seriously at Dreams, with accuracy, timeliness and accessibility being the most important elements. We are continuously striving for improvement and efficiency, and timely reporting and analytics adds value to everything Dreams does.
Can you give us an example of some of the data-driven decisions you make at Dreams and how Exasol’s technology helps with such decisions?
Kasperova: As previously mentioned, our analytics aims to support all decisions in everything we do. Our reporting is aimed at understanding performance, identifying risk and opportunity early across all business KPIs.
In the post-implementation of Exasol, we have focused on data ingestion of large-volume externally-generated data sources aimed at better understanding our customer’s experience, be it data from consumer surveys, call-centre telephony system, product quality data, or data generated behind Dreams Sleepmatch (proprietary mattress fit technology), as well as drive efficiency through all that we do (delivery vehicle telemetry, online Speed data, to name a few).
We have also implemented a web speed monitoring tool which stores speed test data several times a day and enables monitoring of key speed KPIs trends, and results in additional monitoring provided for our online development teams to review the impact of code releases.
We are in the process of developing consumer survey dashboards following an ingestion of these external data sources; aimed at providing end-users with a unified view of all key insights from all the relevant platforms where feedback and surveys are completed by our customers, to unify all reporting within one platform and make the end-user journey simpler. This will also benefit from further insights on customer experiences, improved accessibility among teams and providing further context to already developed analytics.
Stewart: Exasol has been a key component on the journey to automation and data ingestion from varied platforms, enabling easy access for our analytical resource, improving efficiency and releasing time to analyse rather than collate, prepare and ingest data.
As an example, we were able to provide the end-user with automated Sleepmatch analytics. Sleepmatch, a proprietary in-store mattress fit technology, is located in all Dreams stores and carries out live calculations to recommend the ideal product for each customer based on individual needs. Prior to ingestion, only limited high-level store usage stats were available, but with the ability to ingest into Exasol and analyse the large volumes of data, we can now understand the product recommendations made by Sleepmatch as well as better understand the typical customer preferences and requirements. This in turn will help us with product innovation and optimisation going forward. The added benefit of accessibility of data among the insights team results in data being easily blended and shared between projects, providing further context to all our reporting with minimal further effort.
What benefits have you seen since implementing the solution and how does this impact the end-user?
Stewart: The consolidation of the data within the highly available Exasol Cluster has resulted in the data just being available. Previously, where data was distributed across technologies and source systems, we had numerous single points of failure which often disrupted data availability.
Exasol has enabled more relevant analytics, with Insight and Analytics teams more agile and self-sufficient without being heavily reliant on IT development and code. This enables them to spend more effective time on analytics and visualisations to meeting the business needs and requirements.
Also, Exasol has been a key component of the data lake build, which provides the end-user with more relevant reporting as data is shared easily from project to project.
How has the solution allowed you to scale and future-proof operations?
Stewart: As mentioned, the cost-effective entry point enabled us to spec a cluster with significant head room, and Exasol’s licensing model enables a cost-effective path for growth. Exasol will enable us to move away from a model of Kill and Fill (aka Full Refresh) to an incremental model for refreshing our collected data marts/summaries reducing our processing and data refresh load times and making us more efficient.
