In Part 1 of this series, I spoke about the different types of data regimes.  In this post, I will examine the pros and cons of data democracies and provide guidelines for how you can be successful.

Customers often ask me, “Is there an approach or data regime that is generally superior?”  It would seem like a data democracy is the obvious choice for most scenarios. Vendors have been touting “BI for the masses” for years – they must be right. Right?

Wrong.

Firstly, it’s confusing because concepts like,  “single version of the truth” and “golden copy of data” have no place in a democracy. These concepts belong to dictatorships. Democracies have multi-faceted and personalized versions of the truth.  For e.g. A metric like “Revenue” may have multiple – yet correct – definitions based on the context of the observer.

Secondly, there are negatives, including a perception that decision making slows down. Analysis paralysis sets in. Have you ever heard of committees taking decisive action?

Another negative is that democracies emphasize quantity over quality. More data in the hands of more people does not imply better quality decisions.  Sometimes organizations can get lost in minutiae, like when Google decided to experiment with 41 shades of blue. Group think, and second guessing are also “traps” that data democracies fall into.

Data democracies are hard to create and even harder to sustain. But despite all the drawbacks,  there are a few significant long term advantages.

Higher corporate IQ & sustainable innovation

For organizations, it’s not important how clever individuals are; what really matters is how smart the collective brain is.

Creating a data-driven democracy is a great way to raise the IQ of the collective brain. Because of the way information flows in a democracy and connects people with shared interests, it is analogous to how the neurons in your brain work. More interconnections are good – data can flow through multiple pathways. This in turn provides the ideal substrate for ideas to have sex and innovation to flourish.

In other words, data democracies make for smarter, more innovative organizations.

Highly engaged & motivated teams

Democracies help galvanize tribes. People with similar interest can collaborate to solve problems like never before.  The idea of self-assembling groups, working together to resolve problems based on shared interests, is powerful.

Being a part of a tribe also contributes to a sense autonomy and purpose, which according to Dan Pink, are keys to solving the puzzle of motivation.

These advantages are critical to long term competitiveness. However, you can’t mandate a democracy – you have to let it blossom.

Some guidelines that may help:

1) Lower the bar

Allow decision making to take place at the lowest level in the organization possible. Employees closer to the  ground have tacit information and context that senior managers, more removed from the situation, simply don’t have. While this sounds like you are lowering the quality of decision making by allowing lesser qualified people make decisions – you end up with not just higher quality but faster decisions and higher morale.

2) Give rights

Provide “decision rights” for all employees. A data democracy is meaningless without “decision rights.” Those with the most complete knowledge of the situation at hand should make the decision. In the past senior managers had more data and could second guess lower level decisions. Providing “decision rights” will speed up decision making and prevent the trap of “second guessing.”

3) Share

Share access to corporate goals, strategies, priorities, and policies. This will help provide front line workers and managers a better framework for decision making. Otherwise, they are just cogs in a machine and will not benefit from having more data.

4) Diversify

Avoid group think by adding diversity. Not biological or racial diversity for the sake of it, but diversity of background, perspective, experience and thought. This will prevent an echo chamber of like-minded people taking poor decisions.

5) Let new leaders emerge

Just as the finance organization manages money, and HR is responsible for intellectual capital, a new information organization will emerge, outside of IT, led by the Chief Data Officer (CDO). Gartner has recommendations  and estimates that there are currently about 50 established CDOs out there already and expects this to grow to 20 percent of enterprises worldwide.

There is no one size fits all. The best run organizations need to find a balance. But for companies who care about motivated employees and innovation more than short term results,  getting a healthy dose of data democracy is the way to go

It has been said that democracy is the worst form of government except all the others that have been tried. -Winston Churchill

Democracy has changed the world. It is widely acknowledged as a positive and uplifting force – an ideal to be emulated.  Small wonder that people talk about liberating and democratizing data. Public data sets are easier to democratize. But what about corporate data for employees?  I find it ironic that many large cor­po­ra­tions have more in common with dictatorships than they do with the democracies.

Consider this. Corporate leaders are not elected. Organizational structures are rigid and  hierarchical. Information flows up the chain of command and orders flow down. Discipline and coordination are imposed from above to ensure compliance with plans. Insubordination gets you fired!

In summary, large corporations are dictatorships.  Benevolent perhaps, but dictatorships never the less.

This irony is that dictatorships work. Or at least, have worked well historically.  The looming question is – will they continue to work?

Flat and matrixed organizations are more democratic. Decision making authority is distributed. Information flows horizontally and vertically. But flat organizations have a hard time scaling. And matrixed organizations have their problems too.

This leads me to wonder, are hierarchies, and by extension dictatorships, just a part of human nature? Can you have a real “data democracy” when organizational structures impede the free flow of information?

No.

However, change is around the corner. The rise of “data-driven decision making” and “democratization of data” is causing a shift in power. If you believe the analogy – data is the new oil – it should be obvious that those who control data will have enormous clout. With the right training, skills and passion, even a modest data analyst can become a power center.

This shift in power is giving rise to the idea of  data regimes, loosely defined as: organizational structures and behaviors caused by the desire to control or harness data. While data regimes are not a consciously acknowledged construct, they will influence organizational structures for years to come.  Organizations have already experimented with a variety of data regimes with varying degrees of success.

Some of the approaches that I have observed include:

• Data Dictatorship –  IT or some group controls access to data and is restrictive about its use. There is asymmetric distribution of information primarily based on your position in the pecking order.
• Data Aristocracy – Data scientists and PhDs are needed to do anything meaningful. Power concentrates in the hands of these data priests and their benevolent masters.
• Data Democracy-  Everybody gets timely and equitable access to data. Line of business users are empowered and “own” the data. Executives and IT get out of the way.
• Data Anarchy – Business users feel underserved and take matters into their own hands. They create “shadow IT” systems and work around the “unresponsive” IT group. This type of scenario arises when executives don’t provide clear guidelines for interaction between business and IT.

It is noteworthy that these regimes don’t exist as discrete entities. They are part of a continuum. Each type of regime is subject to various flaws and pitfalls. A healthy data-driven culture must be constructed from elements of each.

In Part 2 of this series I will talk about which type of data regimes will dominate, why, and how you can position yourself for success.

This blog was originally written by me here.