Mastering the space between data and insights

I’m often asked to speak or advise on the space between insights – that is, measuring consumer behaviour and psychology – and data itself. Business leaders want to know how to combine these two factors to deliver a superior data strategy or solution that drives customer, operational or commercial outcomes.

I’m of the opinion that good data practice boils down to a few key principles:

1.    Be clear on your use cases at the start of your data strategy

If you want to use data measurement, modelling, and analytics to get a business advantage, then you need to effectively capture the right data upfront. This way, you can measure, understand, and model those factors that will ultimately influence customer behaviour downstream.

Taking the time to identify all the business use cases that you want to solve will influence the upstream data that you capture. For instance, do you want your data to drive improved or personalised digital experiences, or do you want to gain a better, broader understanding of your customers? The success of your data strategy is dependent on understanding how it can support the range of strategic goals and objectives your business is facing.

Once you have identified what success looks like for your organisation, and what dynamics underpin this, you can work out the data points that will help you understand how your performance is tracking against these. There’s a common habit of businesses wanting to capture everything, but if not done in a purposeful way, this can muddy the waters and give confusing insights or result in misleading correlations.

Instead, I often say a good starting point is to ‘think about the customer behaviour that will drive your success’. This can provide a clearer path towards identifying what you’ll want to understand, measure or influence. These then form a central part of the things you should capture and measure (such as increased visitation, increased sales volumes or basket values, increased engagement, purchasing a broader range of products, etc.). Each data initiative should clearly link back to improving performance against your identified business goals. You should be able to articulate its intended value, along with how to measure if it is doing its job once deployed.

2.    Articulate how data insights can be implemented

Now you know what outcomes you are solving for and what data will inform these. This leads to the second challenge: understanding how insights can be used and implemented, and by whom. It’s only when data or insights are used within a product, customer, or commercial strategy that they can grow customer engagement and drive the business forward. Put simply, a model, or data collection more generally, is only as useful as its deployment. If the business doesn't know how to effectively deploy the data it has collected, then it’s not useful, trusted, or understood.

Clearly putting into words how business areas will use the data artifact within their normal day-to-day operations, including what decisions might be supported or processes automated, will help ensure that the end product will add real world value. It is critical to keep this at the core of your design, development and decision making, rather than focusing on the build approach or model accuracy – especially if you want to maximise the chances of downstream engagement and adoption from across the business. Too often data products and models get built, that never get effectively used once deployed. Largely because the build focuses too much on the ‘how’ and not enough on the ‘what it needs to do’ and ‘how it will be used’. In my opinion, you are better off building a less sophisticated product or model that exactly fills the business need and obviously adds value, instead of an overly complex solution that has limited applications or low levels of understanding around ‘real world usefulness’ within the business.

3.    Build and champion the change you want to see

The last challenge relates to driving business engagement and adoption. If a data solution is truly automated, and is zero-touch once deployed, then this step is of limited importance. But this is very rarely the case, with most data products or models requiring businesses to implement, support or review performance once deployed. In these situations, business engagement is key to the success of what you’ve delivered.

I’ve seen the exact same data solution deployed across multiple business units or teams with a wide range of success and value added. A primary driver behind this success or failure is level of business engagement in the build and implementation phase. Only with strong business involvement up front and throughout the build journey can you expect results in this space. Working with business units to truly understand how to weave the data solution into their business practices during implementation and helping them to measure use, and quantify value, will help demonstrate to them the value being delivered. This will keep them motivated when it comes to a sustained effort in supporting the strategy’s deployment and ongoing use within the business. As a data team, you should stay close to the business in these early stages of deployment. In this way you can support them to recognise the value delivered by the product – as you and your team already do.

In my experience, thinking about data and its uses upfront is the difference between a superior or an adequate data strategy or solution. For this reason, our Data Science and Insights team at Plexure works closely with our engineering team early on, to build things that will deliver the functionality that our clients want, and that contain the information we need to drive ongoing value and use.

It’s not that any one of the above shortcomings will kill a data programme in the water, but a lot of small decisions along the way on what we capture, use, how we define our scope, and how to support adoption can add up and prevent you getting the right result. When it comes to building a stronger data-led business, data needs to hold its weight – or it’ll sink you.