The fast-moving consumer goods industry, like many other sectors, has ramped up investment in digital solutions in the last decade. ABI Research predicts that the net digital factory revenue by 2030 is supposed to reach US$24 billion. CPG (Consumer Packaged Goods) manufacturers are likely to be a sizable portion of this pie as they are expected to invest US$4 billion in data and analytics services. Compared to US$500 million in 2021, this is a massive step ahead. 

Some reasons for that are the ever-changing habits of shoppers, the explosion of e-commerce, the rise of new technologies, environmental concerns, and many others. The advent of machine learning (ML) methods and the holy grail promised by artificial intelligence (AI) generated an ‘AI race’ of sorts.  

However, this process has been convoluted for the most part of it. Mainly because two crucial aspects of a so-called ‘Digital Revolution’ have been widely neglected.  

First, many companies have engaged in endless experimentation with ML/AI without first building the foundations for the successful implementation of these advanced models. To capture the benefits promised by these wonderful technologies, companies need data, lots of data.  

We also need to define the appropriate data governance and data management. Very often, however, even when companies have plenty of data and the right governance in place, this information is not ‘ML-ready.’ This means a huge amount of work needs to go into cleaning the data, structuring it in data lakes, joining different sources, mapping, the list goes on. This can be doable with a defined roadmap, focus, and discipline in executing the plan. 

Secondly, we need to crack the adoption code. Even in the not so likely event when the models work and we have our ‘proof of concept,’ getting users to transition to a tool that will give automatic insights is an enormous challenge. We are all too busy with our daily jobs. How we approach the users of a new digital solution is crucial for a successful rollout. This, in my view, is the main challenge for digital product developers in our industry. 

Very often, the first versions of a product (minimum viable product) are clunky. The moment a super busy user sees something not working seamlessly, they give up testing the new tool.  

The ‘glass half full’ situation here is that at least we know a few necessary conditions to drive adoption: 

• Enlisting Leadership Support: We are navigating the unknown when building a, say, predictive analytics tool. If the company’s leaders do not believe in it, nobody will.  
• Selecting Early Adopters: Carefully select your early adopters (and stick with them) – go with the assumption that your users are busy. We are not launching a new toy, but a tool to help people do their jobs. The people who will test your MVP should be very excited about new technologies and willing to commit their time for the duration of the testing and hyper care stages (with support from their leaders!) 
• Agility: Iterate rapidly, but do not lose sight of your vision. The agile framework is great when navigating uncharted waters. Get the most value-added feedback and build it in the next sprint. But if you change the product too much and too often, you will lose credibility. Users will be overwhelmed with so many changes and a lack of direction – they will get disengaged. 

To conclude, the Digital Revolution is extremely exciting, but there are important challenges around data and adoption that we need to face with a structured approach.