I recently had a pleasurable discussion with a Dutch gentleman who had decided to take a sea change and renovate the French chateau where we were staying.

He had recently sold his business that had been in his family for three generations. The business manufactured simple concrete collars for reinforcing bars.

It had withstood the challenge of plastic substitutes and is still a healthy business today.

Improving Productivity

He related that when his grandfather started the business, all the collars were manually cast. On average, a single worker could make about 5,000 collars each day.

When his father took over the business, he introduced hydraulic machinery to assist the workers. A single worker was then able to make 25,000 collars per day.

When he himself ultimately took over the business, he used his electrical engineering training to introduce electronic control systems into the process. With this in place, each worker could now make 125,000 collars per day.

The Case for Business Intelligence

There is something nice and symmetrical about the manufacturing processes. When humans are acting as a piece of the larger suite of machinery, everything appears simple, logical and impressively linear and predictable.

It’s little wonder that the processes and practices learnt through generations of manufacturing experience have pervaded our thinking.

Our human staffed call centers are encouraged to behave like the concrete collar maker and follow the standard procedures and take advantage of the automated aids made available to help them be more productive.

Our hospitals and care centers are staffed by people who are now having the time they spend on tasks compared and benchmarked against scientifically created "best practice" benchmarks.

It’s the enterprise business intelligence (BI) systems that are responsible for collecting the data to assess the performance of the "machine" in the same way that our Dutch gentleman’s electronic control systems were inevitably collecting data from the concrete collar production line and feeding it up to himself and his operations managers to react to any unexpected variances in performance.

The Service Sector

The problem is that services now employ something like 85 percent of the workforce, and they have grown largely at the expense of the manufacturing sector.

Services are human centered. Some services indeed lend themselves to codification as standard manufacturing like processes, but it is these types of jobs that are most at risk to automation.

What will remain are the higher skilled jobs, most of which require interaction and collaboration with others.

So if the future workforce is to be dominated by independent thinking and judging individuals working in collaboration to create unique, high value outputs, what place does a traditional hierarchical, activity monitoring enterprise business intelligence system play in our future?

Growing Workplace Complexity

I recently published an article that contained some compelling data showing how traditional activity measures exhibited absolutely no association with how the organization was collaborating.

In this article I challenged the traditional target for enterprise business intelligence being limited to line management i.e.. those charged with managing others.

There is an implicit assumption here that those managers’ having identified what they believe to be a performance issue will be able to instigate a simple intervention to put things "right."

The organization of the future is likely more comparable to the complex ecosystem of a Brazilian rainforest than the concrete collar production line my Dutch acquaintance had developed.

Visualizing Business Intelligence

The following diagram shows a typical structure of a traditional Business Intelligence system that has hardly changed since the 1970s.

In essence they are reporting tools for managers. No doubt the small number of end-user managers that access these reports are learning something, otherwise they would not have made the investment.

The big question is whether in the organizations of the near future, even if the intelligence is valid, is able to execute the change interventions they believe are required.

typical structure of a traditional Business Intelligence system

As organizations strive to achieve the flexibility and agility of the new waves of disruptive businesses entering the market, the above structure looks positively archaic.

While "end user" reporting from BI systems has been a catch cry from BI vendors for at least the last 30 years, in reality the proportion of staff that actually access such systems could be embarrassingly small, given their current complexity.

Some commentators are advocating addressing the issue through more user-friendly "search" mechanisms or self-service BI.

This might address the access issue but not the content issue. The vast amount of data stored in the enterprise data warehouses are historical, operational and activity centered. If the organization of the future is to live or die based on the interactions and relationships that they build, where is that data? Who is looking at that?

Linear vs. Non-Linear Growth

We are getting an early indication of how linear methods are failing to predict performance with the rise of platform and holocracy modeled businesses.

Learning Opportunities

Both models prioritize interactions and relationships over process. Both have demonstrated non-linear growth in performance beyond what could be expected from a conventional supply chain.

Platform businesses like Uber and Airbnb now outperform their traditional competitors several times over, after just a relatively few years in business. Dutch community care organization Buurtzorg has shown similar stellar growth and performance using a holocracy organizational model.

In less than a decade it has grown to the market leader with more than 6,000 nurses, outperforming the competition on not only cost but also quality.

The competition still focused on classifying service elements to the extreme and allocating performance targets to each element — much like what would happen in the concrete collar making production line.

But people are not machines. If our business intelligence systems are simply capturing data on activities conducted in businesses like these, then they are missing the point.

It's interactions and relationships that underpin the success of these businesses, not the pure activity metrics commonly found in traditional BI systems.

Rather than starting with line mangers as the target, a disruptive business intelligence system needs to focus on the customer facing staff first.

Rather than viewing aggregated data first, staff needs to see data that show how they personally are working and interacting. When you can see how you are working, you are more able to change the way you work.

Rather than "drilling down" we need to "aggregate up" to teams, business units and the enterprise as a whole.

Networked businesses change from the individual up, not the other way around. Granted, the senior management still need to set directions and business goals, but "process" and "compliance" now need to be replaced by "empowerment" and "accountability" in the language of performance management.

What would a disrupting business intelligence system look like?

business intelligence model

Front line staff is the target of the disrupting BI system. Their interactions in the form of on-line and off-line conversations are captured via interaction mining of online collaboration systems e.g. enterprise social, email, instant messaging, video conferencing and periodic social networking surveys into an ‘interactions/relationships’ data warehouse.

Relationship structures that form the basis of work teams are surfaced as "social graphs" from this data and fed back to staff as a self-help mechanism for enhancing collaborative behaviors.

Enterprise vision, mission and goals are injected into the conversation space to be contextualized and owned at the individual and team levels. Performance/outcome data is fed back into the conversation space for collective reaction.

The relationship intelligence feedback loop is used to provide individual reflection and assessment and then collective action regarding performance outcomes.

As with other disruptive business models, one could anticipate this form of BI system providing more actionable insights, being far simpler to use and far less costly to build and operate. To summarize the nature of this disruption of traditional enterprise BI:

traditional vs disruptive bi

It’s time to shine the digital disruption blowtorch on legacy enterprise business intelligence systems and unlock the true potential of BI to the whole enterprise.