Big Data — sweet, rich, untapped data — was supposed to be the great equalizer. McKinsey declared it the “next frontier for innovation, competition, and productivity.” Every organization would have unparalleled insight into the opportunities and risks within their business.
A decade later and the signs of frustration are mounting. A recent report found that only 41% of firms are competing based on their analytics, while 39% are managing data as an actual business asset, and merely 24% claim to have built a data-driven culture.
There is no shortage of databases, dashboards or even alerts. Anyone who wants to visualize the state of their business can. But what do you do when those visualizations fail to show you what to do next??
All the dashboards and prediction algorithms in the world are ineffective unless they produce actual money either flowing into or kept by the business. Ostensibly, that’s what Business Intelligence and Analytics (BI&A) is for.
But stand up in a meeting armed with only charts that say “sales are down,” and you quickly recognize there’s no direct line to ROI. At the end of the day, analytics are about searching for answers to questions you’ve thought to ask, not about turning data into dollars.
That is changing. AI, particularly unsupervised learning, has emerged for the masses to provide a different way to analyze even extremely complex data. This is leading to the rise of a new category of analytics — Data Capitalization Management (DCM). Unlike traditional BI or analytics, DCM platforms are focused on the tangible, realizable value for the businesses using them by automating the analytics process.
But before we go deep into the definition of DCM, let’s talk about how we got here.
Let’s go back, way back. One of the earliest and still widely used analytic tools, the spreadsheet, was invented in the era of the floppy disk with a whopping 1.44MB storage capacity. Fast forward a few years to business intelligence reporting and dashboard tools that arrived when 25GB of storage was revolutionary.
The exponential growth of data makes those days seem quaint. How much data depends on who you ask, but let’s go with IDC’s projections that more data will be created between 2020 and 2023 than in the previous 30 years.
That kind of volume is hard to contemplate, much less analyze. Yet, volume isn’t the problem.It’s part of the bigger problem - data complexity.
As more channels and data sources have come online, the vast majority of the data generated isn’t well-structured or well-understood. Most of the world’s data is unstructured from text, audio, images which is incredibly complex.
Plus, it’s spread out. The average company is pulling from 400 data sources. That’s average. Many are spread out across 1000 data sources.
Buried in the data are the insights needed to not only run the business, but to change the business. You’d have to dig through an exponentially growing layer of complexity to find insights— one by one.
Using traditional AI and tools, this has largely been done by an overworked and under-resourced centralized data team. Most of their time is spent on preparing the data. As much as 80% of data scientists time is spent on massaging the data, which just so happens to be the part of the job they find least enjoyable.
This generates a frustrating scenario for most organizations: the data teams are frustrated with the high quantity of demands from various teams and business teams are frustrated with the lack of insights to act on.
That’s bad enough. As data complexity exploded, something else was happening.
It’s tempting to say the data complexity problem is a universal problem. The reality is that a select few organizations got ahead of the complex data curve.
These high-performing companies saw what was coming, and invested deeply in turning that complexity into their competitive edge within their respective markets. These weren’t companies with huge digital transformation projects in play; they were the ones transforming the digital world. And they’ve greatly outperformed the market.
Digital-first companies built a culture around uncovering hidden insights buried in the most complex data, permeating the entire organization — from which products they make and where they sell them to how they market and how they hire — with insights that guide human decision making, not replace it.
What gave them that leg up? Did they have better tools? Well, not really.
They created armies of data workers. There are more than 1,000 open data scientists roles at Facebook currently posted on LinkedIn. There are thousands of data scientists at Microsoft and Amazon, according to LinkedIn.
The digital-first advantage is not purely due to resources. It’s cultural. These are organizations that look at complex data as a business asset, rather than a liability where most data sits unused with the majority of the investment going into managing, storing and securing it.
The investments in data science are difficult for even large enterprises to match. But it’s the culture of ensuring each team— sales, marketing, supply chain, finance, customer experience, etc. — have real-time access to truly actionable insights that holds even the best intentioned organizations back.
There needs to be a different way. And this is where Data Capitalization Management comes in.
Let’s start with a definition:
Data CapitalizationManagement, or DCM, is the practice of transforming an organization’s data into monetary value. This is accomplished by identifying, prioritizing and tracking insights based on the ROI impact for the business.
The words used here are important. If BI is about providing a system for managing the business, and analytics is about probing into ad hoc questions about the business, DCM is about automating the discovery of insights based on the key performance indicators (KPIs) business teams are trying to improve and providing a workflow for tracking the monetary value of acting on those insights.
If this sounds vaguely academic, let’s think of it in terms of the process for how insights are funneled through an organization.
There are six basic stages for how data analysis works within an organization:
Like all funnels, we start with a lot at the top and the output is constrained through each stage. The main constraint, though, is where we apply human and artificial intelligence. At a majority of organizations, human energy is spent at the top of this funnel, leaving teams to do the heavy lifting of preparing and sifting through an overwhelming amount of data in search of insights. AI is applied at the bottom of this funnel where human decision-making is most critical.
There are all kinds of inefficiencies in this model including lengthy cycles to get to insights, lack of specificity in the insights due to potentially valuable data left out of the analysis, etc.
But most importantly, the people within the business — from how products are marketed and sold to supply chain and risk management — have surprisingly little access to the data-backed insights as they continue to make the most crucial decisions for the business despite the increasing growth of automation.
The DCM version of the funnel flips the roles of AI and human intelligence. AI is applied to the vast and complex work of preparing and enhancing data, and finding meaningful patterns within that complex data. Human time and talent is focused on strategizing and executing on those insights, resulting in increased money into the company in the form of topline or bottomline revenue.
Specifically, there are three steps on this path to becoming a “data profitable” organization.
Dashboards — the most prominent form of data management within business over the past two decades — exist to provide a macro understanding of the business via data visualization. At their most helpful, the dashboard can inform you of existing or emerging challenges. If sales are down in the southwest, it’s fairly easy to see that in your dashboard.
Understanding the why is when challenges can be addressed head on and opportunities uncovered. Most organizations (indeed most leaders) struggle with getting from the what to the why. And using dashboards to get to that level of granularity usually involves several challenges including:
Using the DCM process, the focus goes from diagnosing what is happening in the business to capitalizing on why. This essentially moves the business units from collectively working against high-level goals to executing on granular, often nuanced, insights. The goal is to put these insights at the team’s fingertips rather than requiring deep dives and pulling on data science or analytics resources for lengthy diagnostic cycles.
This isn’t to say dashboards aren’t useful. It’s recognizing what they’re useful for, and that to compete effectively in today’s market requires your business teams to have access to a more granular layer of insights.
Complexity is often seen as the enemy of progress when it comes to analytics. Traditional BI and even AI often can’t handle the inherent complexity of the data a business collects. So tradeoffs are made.
The amount of your data left on the cutting room floor could actually lead to new, fresh, and surprising insights.
To understand why, it's helpful to think of our knowledge of the business in three spheres of increasingly growing sizes.
Most organizations only probe the first two spheres. Tools like BI are available to help visualize the first sphere.Analytics and data science experiments are good for probing the second sphere.But most only scratch the surface — at best — of the third sphere.
Yet, it’s only when the business is able to analyze data across all three spheres that it is able to fully transform their data from a liability to an asset. When the full picture across all spheres is revealed, the business secures a competitive advantage. Not only that, it’s a proprietary competitive advantage as attributes within the data, particularly the complex data, often is only knowable to the specific business.
In a DCM business, complexity is embraced rather than simplified.
This is the last mile of analytics. Frustratingly, even when insights are painstakingly revealed many (some would argue most) never get acted on. After all that hard work of finding the insight, it simply sits in a backlog and becomes irrelevant to our future progress.
In the DCM process, moving from the insight to an actual opportunity is as important as finding the insights themselves. Data profitable organizations ensure there’s a clear workflow for prioritizing insights based on their monetary impact, moving them into action, and tracking the realized return over time.
Businesses do not get rewarded in the market based on how many insights they successfully identify; they are rewarded for the opportunities they find, act on and successfully execute. The organizations that outperform their competitors are the ones that expertly manage opportunities unearthed in the data.
Research from analyst firm Gartner predicts that we will soon see the convergence of Analytics/BI, data science and AI. And that this convergence will allow companies to be more agile and will transform the analytics ecosystem.
In the meantime, what’s become clear is that the competitive difference between the winners and losers in markets will be those who can turn their complex data into monetizable insights...and those who can’t.