little data, little differentiation turns bigger

==========

“Waiting for big ideas is intellectual escapism.”

===================

‘Institutional inertia’ needed to be countered. There is no doubt we are at the moment is on a journey, and that journey needs to arrive at a place of inclusion further on than we are at the moment.

What matters to me in terms of my own responsibility and my own advocacy is that we don’t settle for second best, that we keep trying to move the organisation forward.”

—–

Paul Bayes

==================

“But the brain does much more than just recollect it inter-compares, it synthesizes, it analyzes, it generates abstractions. The simplest thought like the concept of the number one has an elaborate logical underpinning. “

Carl Sagan

============

Work the system, culture will follow. As a corollary, work the data, knowledge will follow.

These thoughts are related if not intertwined in today’s business world.

Currently business gathers gobs of information, knowledge and data and generally don’t really know what to do with it (effectively). That’s why people lean in on generalizations which drive towards superficial decisions. They know they have something important but, more often than not, just aren’t sure what they have.

That summarizes maybe 95% of any company gathering data.

That summarizes maybe 95% of any company discussing culture.

Because of this a business typically does one of two things:

  • is as smart about it as possible and have humans use what they can as well as they can, or,
  • delegate to machines/tech/algorithms (which will always be flawed because it cannot accommodate human irrationality).

Frankly, this is the wretched in-between. They neither optimize the data or utilize even the best data.

Even worse this typically means the system is ignored. This is worse because culture typically mirrors systems and when systems are being used, or ignored, or even used improperly (including the use of good data), some wacky things happen to the culture. Wacky in that it loses its moorings and behavior gets tethered to some dubious ‘values’ which come and go depending on how the system is used. This happens because the moorings are to data, and what data tells us, rather than critical thinking on what is the right, and best, thing to do.

This is where little data comes into play.

Little data can make disproportionately big differences to culture. For while it’s true a business truly only needs a smaller amount of relevant data to make big learnings and big differences, small uses of data used without any true rigor, also creates some big consequences – sometimes even some big differences. In fact, I could argue small amounts of data can create a really big shitstorm.

“I am 100% certain that I am 0% sure of what I’m going to do.”

Parks & Recreation

Interestingly, one of the greatest examples of this shitstorm is the concept of programmatic:

——————————-

Programmatic:

“X company” activates 87% percent of media spend in programmatic channels, with traders trained on multiple platforms. The “company” executes complex technical strategies, like retargeting shoppers from display to YouTube, or designing a pixel recency strategy that ensures retargeting bids are in balance with a shopper’s site history.

Using custom machine-learning algorithms, the system optimizes programmatic bids to drive lift against business KPIs. Different algorithms optimize to different KPIs, whether that’s user acquisition or sales.

All this helps find efficiencies while still seeking creative optimization and sequential messaging to personalize campaigns for individual consumers.

————————

What a bunch of technical baloney wherein they dance on the heads of a bunch of pins suggesting that instead of deflating imaginative balloons to uncover ideas the business their techniques will inflate ‘big ideas’ by pin pricks of reductive tactics. Smart bullshit, but bullshit nonetheless.

Once again. Systems are people and people are inherently impacted by systems.

And, yes, data can play a role in some small ways, that lead to big differences, but in general things like programmatic dwell on the superficial insignificance of the overall needs of things. And maybe that is my point. Anyone can make small things look big & important, but not everyone can discern the small things that actually make big important things.

=====

“You’ve got to think about the big things while your doing small things, so that all the small things go in the right direction.”

Alvin Toffler

===============

Look. I admit this isn’t easy.

Business doesn’t really make it easy for us. It kind of encourages us to be sure to focus on big stuff, and little stuff, at the exact same time <all the while encouraging us to focus on one thing and do it well>. That said. It rewards ‘big’ at an exponentially higher level while penalizing at an exponentially higher level for small (even if you do it well consistently).

At exactly the same time business asks us to squeeze as much as is feasible into the smallest increments of time as possible. AND, at exactly the same time we are encouraged to think big, have big ideas and do big things.

Its nuts.

In addition. I would argue that most of us get “big” wrong <because the truth is that identifying the ‘big’ for us is closer to 0 than it is to 100>.

Let me digress and discuss the relationship between big & small a bit before I get back to data.

Big things are seemingly easy to do because we all see them. And they appear easy to check off our “to do” list. But, inevitably, it seems like everything really does come down to the little things, which unfortunately, we have a nasty habit of overlooking <not necessarily on purpose>. That’s bad <that overlooking thing>. It can be a big mistake <which is a big thing by the way> and a large creation of stress in our life <another big thing>.

This mistake gets compounded by a truth that small things most often have a large impact to alleviate some of these potentially negative ‘big’ things in our lives. Take a moment and think about it. Stress can be partially managed by doing a couple of really small things like ‘show up on time’, ‘listen’, well, you get it. This is true with even small amounts of data used well. Smartly used data can highlight impending problems which can then be head off at the pass.

Yeah.  Sometimes small important things sound small.  But do big things. Make a big impact.

Anyway. Getting back to little data.

Man’s shoulders full of dandruff

For the most part little data creates little differentiation. This isn’t because little data is useless, but rather because the little data we grab most often is the stuff that just keeps us chugging along without having to think too hard or have to ‘overthink’ too much. We use just enough of that oil (remember – data is oil) to keep the existing engines running smoothly.

“The belief that the commonplace is really worth looking at, and the courage to look at it with a minimum of theorizing”

John Szarkowski

I would also suggest that data, little or big, requires making decisions.

Aw shit. Decisions.

Suffice it to say decisions, in general, are hard to make <even by people who are quite capable of making a good hard decision>. Decisions are hard mostly because you are guessing or judging probabilities on consequences (even with data in hand).

Decisions, in and of themselves, are easy.

Consequences, in and of themselves, are a real sonuvabitch.

And this is where I come back to data.

Consequences are such a sonuvabitch we overwhelm ourselves with data attempting to make a probability a prediction. Okay. Maybe let me say it this way. We overwhelm ourselves with the data we want to make probability a prediction. I say that because “overwhelming” can be done with lots of data poorly used and little data misused. Either way, in doing so we, well, we lose sight of everything else.

He blew on one of the dandelions, and the whole world disappeared.

Which leads me to how we mitigate our risk.

Over fitting. While overfitting is a data analysis term, it’s also a human bias. I would also suggest overfitting is the greatest enemy of conceptual thinking (which would then make it the greatest enemy of an organization).

This may be one of the few times I compare humans to machines. At the core of machine learning are programs that learn from past experience. This actually means machine learning can arc toward increased performance or poor performance as it navigates underfitting and overfitting.

Simplistically, this is exactly the situation human decisionmakers in a business are faced with. In fact, I would argue because of the feelings of uncertainty that business complexity inherently creates, combined with business’s general penalization of risk to employees, the battle between underfitting, appropriate fitting and over fitting will define who wins the future war in the future of work.

Overfitting is one of the banes of business – it is a reflection of the ongoing battle between certainty & uncertainty. it is actually the response to people who do not understand probabilities and fear explaining uncertainty, randomness and the sometimes quirkiness of anything to do with people. Overfitting, in business, is the response to survival in the corporate world.

And, once again, I would point out that overfitting does not demand a shitload of data. In fact. I would be remiss if I didn’t point out many many businesses do not really invest in a lot of useful data therefore the small data they have, when overfitted, only increases the likelihood of bigger mistakes. But. More often than not someone will bludgeon you with a dull axe of data points believing ‘more mitigates risk.’

That said. All of this leads me back to little data.

I will suggest focusing small. Focus on small amounts of data, data that shows the underpinnings of patterns, avoid data exceptions, and make the decision.

The truth is Data, and information, has diminishing returns in any given context. In other words, cramming more data or information in adds value, but in diminishing returns. You keep thinking ‘if I add this one more little thing it will finally uncover the bigness thing I am searching for’. Well. It will not. I would argue if you gather a small amount of useful data well, anything big that could be discovered will be increasingly likely to be found there.

** note: read my piece called data decipherer or my piece called data, music and stories

In the end.

Most often it is the little things that are the truly important things. Little data showcases cracks in differentiation which can be exploited to a business’s benefit. Which leads to end where I began – work the system, culture will follow. This piece may seem like it is about data and data analysis, but it is actually about business culture. How you use data is part of the system of how a business works which inevitably begets a culture. Risk, critical thinking, understanding connectivity and complexity, all of those things are embedded in how a business uses its data. All of those things, are, well, culture. Ponder.

, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
Written by Bruce