Today’s retail landscape is a tough gig for both bricks and mortar and online retailers, already a hugely competitive sector, where a shiny new store is no longer a certain path to more profit, as retail has been stood on its head by the COVID-19 pandemic. Tech-enabled innovation isn’t a nice to have, it’s a necessity for survival and sustainable success and In The Memory is a retail tech company with a difference, on a mission to unlock the vast potential that resides in data and turn it into a sales uplift for their clients. Coupling In The Memory’s AI expertise and agile design thinking methodology with Databricks means they can handle colossal, ever-growing volumes of data to build and run customized AI solutions for category managers, buyers and marketers delivering on their unique promise — that fully customized analyses can be answered in minutes, regardless of data set size, resulting in impressive sales growth for clients and 10x business expansion for In The Memory.
It’s the tech startup myth: Three founders — retail experts — working together on their first client project use Databricks as a quick-and-dirty solution on a small server to validate everything’s working together. Problem is, the solution delivers so successfully and rapidly that the client wants to scale up to 500 users, needing a bigger server and new hires. And so In The Memory is born, a company combining data, consultancy and technology, exploiting the market opportunity to deliver technology-enabled consulting innovation to retailers such as Intermarché and fast-moving consumer goods (FMCG) companies such as P&G.
“We’re a young, fast-growing firm. We needed a platform that could support our innovative data-based tech and consulting approach and handle huge volumes of data,” said Alexis Mau, President and Co-founder of In The Memory. In The Memory’s secret recipe is the blend of technology with design thinking and agile cloud-based solution delivery. “Competitors have an older technology stack and more traditional ways of working. Our approach is different. Rather than consult, recommend and leave to implement, our engagements are alive, evolving, perpetuating, thanks to the tech solution we implement, along our change management approach. We collaborate with our clients over the long term, giving them the best simulations so they can focus on commercials, relationships and human tasks,” said Alexis. Customer marketing, category management and local adaptation to improve store performance are In The Memory’s bread and butter. “We use data and AI to help our clients understand what products to stock and what promotions to run and when. What categories to increase and which to decrease, and how the pricing should vary,” said Alexis.
Teasing the actionable insights out of the vast volumes of client data isn’t easy, especially when turnaround needs to be in seconds rather than days. “Client data is in silos, in different clouds, with different rules, so our first challenge is to connect to various data sources, interconnect them and ingest the data daily so it’s ready for analysis quickly. We’re a small team so we need to be able to automate our processes and reduce the human time needed for recurrent tasks,” explained Alexis.
In The Memory uses Databricks and Delta Lake to provide a solution that is cloud agnostic and yet fully integrated into Microsoft Azure, and they’re already managing the data of several big clients with minimal human time spent on integration. “We can quickly make modifications when needed, specifically when we need to adjust data format or add new data,” said Alexis.
On top of this, client use cases can be very different — some require highly atomic use of the data, calculating the performance of one product in one store on one particular day, for example, through to use cases that require billions of rows of data to be crunched in seconds. Being able to create and customize clusters has been key here. “Depending on the use case, we can define the cluster size, and partition the data to quickly filter and query the data,” said Alexis.
For a small fast-growing organization working with some of the biggest clients in retail, efficient operations and fast upskilling are a must. Automated jobs are used wherever possible, from recurrent daily and weekly updates to automatically calculating requests from clients. And even new recruits can create a notebook in a few seconds to read data, run code and share it across the team. “Because Databricks is so intuitive, the delivery team works faster – we don’t need a specific team to work with Databricks. We’ve already saved the cost of one FTE as a result,” said Alexis.
“Without Databricks, In The Memory would not exist as it is today. Delivering our client value proposition is down to the technology choices we made. Databricks is so easy and intuitive we can design, test and deploy client solutions more rapidly, achieving a 30% better run time than the other tech available,” said Alexis.
And the commercial results speak for themselves. Category management clients experience a 2%–3% increase in sales in their large format stores and, in small format stores, rapid insights are delivering an 8%–10% sales uplift. “Databricks grows with us according to our needs, it supports our solution. It’s robust, quick, scalable,” said Alexis.
Unsurprisingly, In The Memory has very happy clients: 100% say that they are able to make better decisions and 99% say they have gained time, visibility and new insights into their business. In The Memory’s approach is freeing clients up to focus on innovation and strategic thinking. “Using Databricks means that we can deliver on our promise to our clients. Our AI tool and expertise mean clients spend 80% of their time on the highest value 20% of tasks. It’s all about enriching decision-making: using technology to make the most of human intelligence,” said Alexis.
Using Databricks means that we can deliver on our promise to our clients. Our AI tools mean clients spend 80% of their time on the highest value 20% of tasks.”
– Alexis Mau, President and Co-founder, In The Memory