Should you let your data do the talking?

We all use AI apps, we all know that data is valuable and that data-driven decisions lead to better strategies - generally! But behind the scenes, how we look at data or extracting value out of it still seems to be stuck in the stone ages.

Take marketing insights as an example, questions are around efficiency of efforts, audience and pipeline metrics. It's the same 10 questions for most companies across most verticals - it's almost universal.

But beyond the charts and numbers, there's a less tangible but equally vital aspect of data that often goes unnoticed—tacit knowledge. This includes intuition, context, and the unspoken insights that profoundly influence our understanding of data. And this, my friend, could be your silver bullet(s)

So lets start with the basics, what is tacit knowledge?

Imagine the last time you had a "gut feeling" about something—like a creative leap or a strategic direction. That's tacit knowledge. It's the kind of understanding that comes from experience and intuition rather than concrete, quantifiable information. In data science, this encompasses everything from recognizing patterns and cleaning data to storytelling through visualizations. All these crucial skills benefit from a deep, often unspoken understanding of the subject matter - and require resources like time, data scientists and tools to extract.

Why should you care about this?

Because tacit knowledge helps us see beyond the obvious. While explicit data tells us what’s happening, tacit knowledge helps explain why. It allows you to dive deeper into the data, uncovering patterns and connections that aren't immediately apparent. The result? Richer insights that lead to more strategic decision-making. These patterns help you answer the why around what's working, so you can take that winning combo and repeat it for continued success.

But wait, tapping into this invaluable resource isn't straightforward...

By definition tacit knowledge is deeply personal and highly contextual. What works in one scenario might not apply in another, making it difficult to have repeatable success. It's the example I see so often of playbooks not translating across companies. This is the first problem.

Second problem is that traditional data analysis tools are geared towards quantitative data, often neglecting the nuanced insights that tacit knowledge provides. They will tell you campaign X is performing well, but not share the why so the learning can't be applied in the same way. And even if they were able to showcase these insights, you still have things like knowledge silos and just a general hesitation to get buy-in on an insight that isn't immediately obvious.

There-in lies the third problem. We have gotten so used to analytics that perpetuate confirmation bias, that a new insight that is anomalous causes a lot of raised eyebrows.

Imagine I told you that even though you generated 10X more leads in a month, the campaign that only generated 2 leads is the one that will actually convert to revenue. I would have a really hard time convincing you that the campaign that seems to be underperforming "on paper" is the one that actually got the job done. Now imagine how much harder my job would be to convince you to prioritize the efforts against that campaign because the data predicts that it's most likely to result in revenue. I imagine most marketers might tell me to take a seat - ha!

Here's another example, most of us use a series of filters to do first pass at creating an audience for our marketing efforts. These include things like company, title, 3rd party intent topics etc. And then we run experiments (but let's be honest, most of them barely reach any real statistical significance). But even if you get the matching right- meaning you delivered the right messaging to the right person and captured a lead - you still dont know what affinity made that pairing work. Did the person click on the messaging because it was a peer affinity? Was it a pain point affinity? That why is exactly what could emerge from tacit knowledge in your data.

On the surface this seems easy to map, but data systems are like ever evolving organisms. So the problem is hard by nature of the shape constantly changing.

We live in a world where data is plentiful but real insights are scarce - so recognizing and nurturing tacit knowledge becomes almost a necessity. It's important to understand the numbers but also the stories they tell.

What would our world look like if we let the data speak to itself? What if we have systems that shift from a question/answer paradigm to a show/tell paradigm? Show me what's working and tell me why it's working.

Thankfully, advances in AI, especially in cognitive AI, are beginning to address these challenges. It allows for tools that can interpret, learn from, and scale the application of tacit knowledge across different contexts.

I stumbled across this while trying to sift through data sets for my PhD. My thesis was in the pharma and smart city space, but I realized that the opportunity can be applied across other kinds of data as well.

I have been working on applying this approach to solve problems like:

• Providing defensible marketing attribution using a KNN-based approach.

• Accelerating customer journey mapping by rapidly identifying revenue-driving touchpoints.
• Dynamic audience discovery to detect shifts in consumer behavior as they happen, enhancing marketing intelligence.

And the results have been astounding (even for me!)

For example, one customer was able to double their market share in a key demographic by swiftly adjusting their digital marketing strategy in response to real-time data showing unexpected shifts in consumer behavior due to anticipation of bad weather - something you wouldn't think to marry.

Another customer was able to accelerate their buying cycle by INCREASING the number of touch points by 40%. They saw a trend around developer products where pushing for conversion too early or not having enough touch points was prolonging or killing their deals. Again, it's an insight that doesn't come naturally in existing tools.

I have seen impact around early risk detection (I know we are sweating about churn these days), proactive compliance management and cost reduction - all things that make the hearts of COOs and CFOs sing. All things you wouldn't expect to get out of traditional marketing analytics platforms.

All that is to say, the opportunity that data holds is pretty vast but most of it is lost if you aren't asking the right questions. Let the data tell you what it can answer for you, and then pick the things that are crucial for your business.

THAT is your silver bullet!

If you are interested in seeing these results, or just talking about some of these issues, shoot me a note and I would love to chat.