Insights transformation has the potential to accelerate business performance – but needs to be led from the business need, not from the technical solution. A shared definition of ‘insight’ and a focus on ‘action triggers’ will help to ensure that the commercial impact is realised.
Digital transformation is like the evolution of TV streaming. We used to be stuck with a few channels that we could watch when they were scheduled. But then, Netflix and other streaming platforms came along and changed the game. Suddenly, we could choose what we wanted to watch, when we wanted to watch it, and how we wanted to watch it.
In much the same way, digital transformation is changing the way businesses operate, and particularly the way that they seek to respond to emerging opportunities and threats as they happen. A future where important patterns are automatically detected, processed and routed to individuals in the business as personalised ‘action triggers’.
Data assets often represent significant investments, and yet Forrester reported that between 60% to 73% of business data is unused for analytics. This is perhaps unsurprising given how cheap and easy it has become to capture and store large swathes of data. However, data storage is only a start – if the data is to have any impact on the bottom line, it needs to be turned into ‘business fuel’, aka insight.
Dreams of insight transformation were born – where people would be unshackled from mundane analysis, whilst sales and marketing teams would be served with powerful insights that guide action. This requires a number of things:
- A shared understanding of what we mean by an ‘insight’
- Modern, shared data storage
- Appropriate methods to identify patterns of interest (‘discovery’)
- New approaches to how such insights are embedded in the workflows of leaders in the business
However, such ‘Insights Transformation’ has had something of a rocky start – with many early projects plagued with the challenges commonly associated with larger IT projects. In particular, the challenge of satisfying a broad range of stakeholders with the right insights, on their terms has proved to be significant.
Common Definition of ‘Insight’
One of the most overlooked issues with insight transformation is to define what we mean by an ‘insight’. The salience of the term has been eroded through mis-use and over-use – with FMCG manufacturers and agencies alike frequently interchanging the terms ‘data’ and ‘insight’. Insight transformation requires us to set a high bar for what qualifies as an insight. A good litmus test is to question:
- Is this a penetrating discovery, or something we already knew? (mind-opening)
- Does it relate to behaviour that we can influence? (behaviour-related)
- Can I act on this information in a way that drives commercial advantage? (personally and commercially relevant)
This definition will drive the scale of ambition for insights transformation work, in particular how to generate personally relevant insights for each individual across the business.
Storing Data for Advanced Analytics
The cost of storing and analysing data has fallen dramatically in recent years. Thanks to the growth of cloud computing, businesses no longer need to invest in expensive hardware and software to store and analyse data. Instead, they can use cloud-based services to store and analyse data on a pay-as-you-go basis, which is much more cost-effective and efficient.
A single, cloud-based data store is a common starting place for organisations embarking on insights transformation.
The way that data is structured and stored will be shaped not so much by the form of data as it arrives – but by the various use-cases for how it will be processed and applied to guide commercial action. Analysts predict that instead of big data, small and wide data will become more important. The rise of artificial intelligence, data fabric, and composite analytics solutions is allowing businesses to search for insights in a variety of small and large – as well as structured and unstructured – data by applying methods that look for useful information within tiny or even micro-data tables.
For example, although a conventional data source may have a column for the colour of an object, AI-friendly data might include numerous columns (often known as features) that inquire “Is it red? Is it blue? Is it green?” and so on. Because there are so many more potential columns/functions, these wide data structures necessitate a database engine’s particular attention.
According to Gartner experts, by 2025, 70% of businesses will have migrated from big data to small and wide data (or data that comes from a variety of sources), allowing for more contextual analytics and intelligent decision-making.
In FMCG, a key need is to be able to connect syndicated data (such as sales audit and panel data) with other sources that manufacturers collect or purchase (such as brand trackers, promotional feeds or retailer EPOS). Instead of being restricted to exploring each dataset within the ‘wrapper’ provided by the data supplier, companies need to pull data into their own single store as a foundation for more sophisticated processing and analytics.
IT can make mission-critical data more discoverable, widespread, and reusable across an organization’s various environments, including hybrid and multi-cloud configurations, by providing a unified data architecture that acts as an integrated layer between data endpoints and processes.
Owning Insight Discovery
When it comes to insight discovery, organisations have a difficult decision to make – where will the onus fall?
With numerous data sets stored across multiple silos, this discovery work has traditionally been carried out by a few specialists. This can work well if the team is sufficiently resourced and has the right skillset. However, it can also become a bottleneck if the team is unable to keep up with demand. In addition, the risk is that this expert team become bogged down in requests for ‘numbers’ for upcoming meetings and don’t have the space for more strategic work.
Steps such as providing dashboards and scorecards to visualise data can often be an initial step in allowing personalised insight discovery (‘self-serve’); however the onus of discovery still sits with the end user, who may be brilliant commercially – but may not be sufficiently versed in the best questions to ask of the data in order, nor how to interpret different data sets in combination, in order to get to action triggers.
In the future, there will be much more reliance on computers to augment the capabilities of humans in insight discovery – guiding them to act just as computers guide drivers on the best route to take. Artificial intelligence (AI), automation, and machine learning are revolutionising business across the world. AI is rapidly advancing, particularly in the field of data analytics, in which it not only extends but also helps to generate superior commercial value.
The pandemic and remote working capability have opened up new pathways for tracking and analysing information, in some cases building a stronger data-driven culture in businesses. More AI-based analytics are being funded as a result of this data culture.
AI has a wide range of applications, and it is likely to play an important role in the future. Some examples include increasing sales by anticipating demand and ensuring that warehouse storage is correct, improving client satisfaction by reducing delivery time, and automating procedures that would otherwise need a human.
Machine Learning (ML) and AI are not just valuable in pattern recognition – they can also be applied to actually direct the process of insight discovery. The findings from the first stage of analysis frequently produce new questions, but with tedious manual processes, this iterative approach is a luxury that few may achieve. Sophisticated ML-based approaches that can ‘see’ a pattern across initial analyses and use this to inform the next stage of automated discovery. This allows the latest software to generate action triggers for each manager autonomously.
It’s worth being aware of the hype-cycle. AI itself has become something of a buzz-word and it’s important that all AI applications are set up with the right level of domain expertise in order to avoid erroneous connections being made. Again, it’s about starting with the need and selecting the technology that will best address that need.
Embedding Insight in Workflows for Commercial Impact
Businesses are under more pressure than ever before to make fast and effective decisions. In an ideal world, businesses would routinely use the optimal blend of insight and intuition to decide on the next course of action. However, the challenge is how to generate and target insights to the right individual in a timely way, in order to drive commercial impact.
This has led to an increased focus on how insights can be embedded within workflows (making adoption inevitable) rather than requiring managers to go hunting for the insights that they need. For example, in range optimisation, a process that guides category managers through key stages, surfacing the ‘science’ so they can add the ‘art’ can cut Range Review preparation time by as much as 75%.
In a world where inflation and input cost pressures change fast, real-time input on pricing corridors and price elasticity can inform decisions about where and how to apply cost price increases. Success of insight transformation work would be to make the adoption of these insights second-nature.
A meaningful transformation
The challenge of translating data into action triggers for individuals has been with us for many decades. With market dynamics shifting faster than ever – and the volumes of data exploding – winning companies will be those that transform their approach to get the very best from both human and compute capabilities. No longer stuck looking at the same data that everyone else has. Instead, served with personalised tips and action triggers to drive commercial performance. Now is the time to explore how automation can help give managers the clarity and confidence to act fast.
If you would like to discuss any of the topics covered in this article, or if you would like to learn more about how you can turn swathes of data into personalised action triggers, please contact us today. We would be happy to discuss your specific needs and requirements.