Data built modern marketing, but AI is rewriting the rules

It’s hard to believe now, but there was a time when people collected data only when necessary. Photographs of a typical ’70s office, with rows of filing cabinets and index cards, spoke to a very different attitude toward data. You keep what you used to have, knowing full well that you will need to go back to it – and nothing else.

At this time, anything beyond a company’s core data was considered corporate waste. Data was a product, not an asset. This was largely driven by technology. Even as we moved from paper to the Internet, digital storage was slow, expensive and difficult to mine and analyze. Even when data was saved, it was often seen as a write-up, stored but never referenced. Data was a liability – expensive to store and potentially dangerous.

However, as technology advanced and analytical methods improved, things changed. Over the past few decades, there has been a constant change in the way we view the data we generate and collect. From being a piece of business, it quickly evolved into a marketing and business asset – the new oil, as we were often told.

How data has become the center of marketing

This change has pushed companies to rethink what data they collect and why. Even if you didn’t know how to use it, the obligation was to store all the data – even the smallest transaction data. Technology and data management techniques have advanced so much that data lakes, ponds and oceans have emerged, and all data is now clean and available for analysis. In theory, at least.

As our analytics and data science skills improved, we moved from being descriptive (“What did the customer buy?”) to predictive (“What is he likely to buy next?”). This type of understanding is very important to the company, allowing us to adapt our offerings and businesses to respond to consumer needs and improve efficiency.

But there was still another step to be taken: from the predictable to the implied. This step goes beyond telling the customer what to do next and instead tells us what to do next. Systems have begun to emerge that give us the next best thing to do – to actually do it. For the most part, this was limited in scope (that is, offering the following whatever discount you have to include), but it nevertheless gave us a powerful way to adapt to changing customer and market requirements. It’s all based on the data we collect.

All of the above depends on us treating the data as the property we return to. The purpose of advanced analytics – whether descriptive, predictive or explanatory – is to give us a better lens on the data we have and what that means for our business.

Why AI models are changing the role of data

We now find ourselves in another major technological shift, as LLMs and other AI-related technologies dramatically change the way we work. It can be tempting to think of these new methods and technologies as the best ways to work with the data we have – and in some ways, they are. However, if you step back and ask what role data plays in this technology, you will see that it is much stronger than cool new tools.

To understand this, we need to look a little under the hood. Most modern LLMs are built on a structure called transformers. They take your text input and process it using billions of parameters (mathematical rules) learned from the original data feed. The way they store this information can easily be compared to file compression.

The text “What is the capital of France?” it successfully produces “Paris” not because the model has a search engine inside it, but because its parameters are effective, such as compressed recall of the entire original training set. Although not perfect, this analogy is useful. As sci-fi author Ted Chiang says, LLM is like a “blurred JPEG of the web.”

The implication is that once the model is trained, it contains all the information it will store (with varying degrees of fidelity). When we use a model, we are not going to the source, but to an imperfect representation of it. If you think of the blurred JPEG analogy, our challenge is to supplement the model with a clear, hi-def image of our business, derived from our proprietary data.

Because the scope of today’s foundational models is now so deep, they are very good at the set part of the workflow, not just analyzing but telling what we do next. Along with your data assets, you now have the ability we’ve been working on — to go directly from data to action.

What this change means for your data plan

Another technology that is helping to drive this change in the way we use data is the Model Context Protocol (MCP) – a standard way to expose our proprietary data to models – effectively becoming a universal adapter that allows models to read your live data without forever swallowing it in their implicit memory. MCP is still in its infancy and will not be the final way of how data and models interact, but it shows how necessary it is to rethink the role of our data assets.

This means that we now need to rethink the role of our data. If the primary purpose of our data is to train or augment a model, does that change what we collect and when? Does it change its importance and role in our marketing and business environment?

Today’s challenge for anyone who collects business data, which is for all of us, is how to change our thinking to accept that data is no longer the primary asset? Companies that rethink the role of their data assets will thrive in this new ecosystem.

Important takeaways

• Data has changed from a stored commodity to something that feeds and shapes AI-driven decisions.
• The evolution from descriptive to predictive to predictive analytics sets the stage for today’s AI workflows.
• Large language models do not receive data in real time, they rely on compressed information that must be supplemented with proprietary data.
• The real benefit now comes from combining basic models with high-quality, business-specific data.
• Marketers need to rethink data strategy from collecting everything to making data usable for models and decision making in real time

List generated by AI