But let’s go back to the beginning. Our four founders have lived and breathed retail data science and analytics for a combined 80 years. Before setting up HyperFinity, they repeatedly hit the same wall. They found retailers lacked connected data and decision making. Meanwhile, the available tech solutions just weren’t fit for purpose for modern, multi-channel retailers.
Our decision intelligence software is built on this first-hand knowledge and experience of retail analytics. We saw an opportunity to create something innovative and impactful. What if a tool could both speed up a data scientist’s job and give non-technical retail decision makers access to AI, machine learning and BI tools in a non-coding environment? What if it could connect data and insight, whilst also connecting decision making across multiple retail functions? What if this tool could ultimately power more intelligent commercial decisions, by putting customer data at the core?
Step in, HyperFinity.
The retail tech stack.
A retailer’s tech stack can be simplified into three main layers. Each needs to be implemented correctly so that organisations can set themselves up for success.
- The data cloud layer. Think of this as the engine room of any technology stack – Snowflake’s cloud data platform is a great example. At its core, the enterprise layer is a centralised repository of data that’s flexible and scalable to federate data or models across multiple business areas or applications. In the world of retail data and AI, it’s essential to choose the right technology – that’s powerful, connectable, and scalable, yet cost effective.
- The intelligence layer. Many companies skip the intelligence layer by jumping straight to activation or by filling the gaps with backwards-looking BI tools or cumbersome data science platforms. However, it’s an important layer, as it not only connects data but also joins up decision making. Lots of organisations have centralised data science or analytics teams that essentially act as the intelligence layer – by using data science platforms to create insight for decision makers. But herein lies the problem. Data science functions are pivotal, but they’re also finite resources. They cannot cope with the explosion of demand for data and insight – this is where HyperFinity comes in.
- The activation layer.The final layer activates insight to create value. For example, a CRM platform for personalised marketing campaigns such as Braze, or a CMS (content management system) for optimised online experiences such as Shopify. In the activation world, some tools claim to perform the intelligence layer, but endless configuration scenarios that are difficult to implement can prevent retailers accessing true value.
A tech stack that connects enterprise data, creates decision intelligence, then activates insight – whilst placing customer data at the heart – is the key to success.
Introducing our flexible product architecture.
We’ve positioned Snowflake at the heart of our tech stack. Why? Simply because its platform is built specifically for modern data stacks on the cloud. Snowflake also offers infinite scale, depending on the use case.
HyperFinity is both a managed app and a connected app. We’ve created a flexible application, depending on clients’ use cases and whether they use Snowflake as their enterprise layer. HyperFinity is built to be as connected as possible – by connecting directly to a client’s source data; by securely ingesting a client’s data into our environment; or by passing outputs to downstream activation platforms.
Our platform communicates with Snowflake through ELT to connect or ingest client data; through data querying to push data to our front-end for visualising in interactive charts; and through training machine learning models directly in Snowflake with Snowpark (to visualise model outputs and productionise for activation).
HyperFinity is both an exploratory data science platform and a data visualisation/BI tool in a point and click environment. Our architecture is built for speed to insight, but also to enable us to scale performance without out-of-control costs. Snowpark has enabled us to keep the platform as serverless as possible.
Our app is deployable from the moment we connect to clients’ data – in minutes and hours, not days and weeks.
Four pivotal data sets for customer-centric decision making.
Four core data sources fuel decision intelligence. These data sets help create foundational insights (affinity, attributes, customer need states and segments) that power a string of interlinked retail decisions.
Any information a retailer collects about customers, such as demographics and segments. Customer identifiers can be used to link together orders or web sessions.
A dataset of all the products a retailer sells, alongside product IDs, taxonomies, and corresponding features or attributes.
Every purchase a customer has made over an extended period – usually two years. eCommerce data can fill in any gaps from in-store data.
Web click data.
A treasure trove of data, detailing every click, interaction or purchase a customer makes when browsing or visiting a website (often found in Google Analytics).
HyperFinity in action.
Our decision intelligence software enables retail optimisation through the intelligent use of data and tech. For example:
Marketing and media.
Marketing teams need the ability to create audiences quickly, with any configuration of customer features. With HyperFinity, the possibilities are endless.
Our decision intelligence software generates foundational insights (such as product attributes and customer segments), so that marketers can create audiences. Let’s take this example: recent spending males, based in Yorkshire and over the age of 40, who often purchase coats (a description of our CTO Damon Bryan). Our platform searches for customers with the features listed, creates an audience, then provides additional insight. Similarly, HyperFinity can create audiences based on customers’ affinity with specific categories, products, or attributes. For example, it can establish the most relevant audience for a marketing campaign promoting women’s summer dresses.
Decisions can be created at both the individual and audience level. For instance, you create an audience for a summer clothing campaign, but show each individual 10 products most relevant to them through dynamic content. Meanwhile, you can understand which channel the audience will engage with most (i.e. email, SMS, app), then send communications at the best time/date for each individual. HyperFinity creates decision intelligence quickly and automatically pushes audiences to any activation platform.
Creating intelligent pricing strategies is crucial for retailers. Our platform creates and configures price-elasticity alongside pricing insight. For example, it can help identify which products are price inelastic but have a low margin – in this case, you might consider a price inflationary strategy to protect overall profitability. Alternatively, HyperFinity can help you understand which products should go on markdown – i.e., if sell-throughs have been low so there’s lots of stock left, but the product is high margin and highly price elastic. Here, a reasonable price deflation could result in a healthy sell through with good margin.
Our decision intelligence software allows retailers to create this insight for individual products or categories. Crucially, foundational affinity analytics help forecast and predict the overall impact on pricing strategies across full ranges and hierarchies – including the wider on your whole supply chain.
HyperFinity helps retailers sell the right products, at the right price, to the right people. Contact us to find out how decision intelligence could be your retail business’s superpower.