Translation Error

It is estimated that our data deluge doubles every year or two. Given all of this information, you’d expect better decisions … but there is a translation error that prevents this from happening. Why?

There’s a difficult conundrum. If you lack essential information, akin to knowing where you are on a map, you don’t have what you need to determine the best route to your destination.

Information on current conditions, like traffic reports and the amount of gas you currently have, affect those decisions on which route to take and how long you can travel.

Yet having too much information overloads the decision maker, hindering their ability to make a decision. Too much detail on traffic conditions means that the information is less useful, due to obsolescence. Too much information hurts you when it is inaccurate by the time you make a decision, because it is no longer as good or even wrong by the time you’ve finished analyzing it.

This is already well known and its implementation is best scene when attorneys respond to information requests by sending everything related to it. We can’t sort through all of the information fast enough to find the relevant nuggets in a timely fashion, and sorting through the deluge while trying to assign value means we’re less likely to find the critical, and in this case, incriminating, details.

The result of the deluge of information for many people is an immediate paralysis due to analysis.

In some cases, people sit waiting for the next few pieces of information, hoping that they’ll get an indicator of what to do. It is easier to sit there and wait for more information, because it seems like the intelligent choice. That choice is made easier by the greater fear of making a mistake than the relative safety of, “I’m waiting for more information.” In some cases, the reaction is to react based on the first critical piece because waiting for what may be another relevant piece makes the first information obsolete. Unless it is a time critical issue like dealing with someone bleeding to death while trying to determine the cause of the injury or wanting to get home before the kids’ scream even louder for dinner, the default solution in the deluge is to try to avoid sinking in the storm and waiting for some sign of what to do. This is the translation error – the gap of trying to translate all of the information into a clear indicator of what to do.

The beauty of SPC charts and the reason for their popularity is that it avoids this deluge. The key metric to pay attention to is identified. And more importantly, people have clear rules as to what triggers action and how to act. When the trend line follows a trend, call in a quality engineer if it isn’t obvious that the equipment is in need of repair. When the unit starts generating way more outliers, shut down and call in someone to fix it. When it is trending up and out of spec, recalibrate or fix it. You know both when something requires action and how to act.

The translation error that occurs with today’s information deluge is two-fold. What requires action? And how should you act?

We have more data, and we get more notifications than ever on something crying for attention or seemingly out of specification. The question in these cases is: what should I act upon? And which one, in the increasingly cluttered field, should be investigated further?

Regardless of which question, the next translation error is determining how to act. We often see this become the source of a bogged down analyst. Financial metrics look wrong or aren’t in the green. Now what? The customer satisfaction statistics aren’t good but you don’t know how to make them better. What should you do in response? Asking for data is a common “solution”. This is done in the hope that more data will help you determine why, as well as answer the silent question, “What now?” And asking for more information, more analysis and new reports provides the immediate answer to others who ask, “What are you doing in response to the anomaly?” After all, analyzing the situation is a step toward trying to solve it. That you may not be able to translate the information into the necessary action is irrelevant – you’re doing something. That it takes up time and may lead you to doing the wrong thing is irrelevant, too.

What is the solution to the problem of translation error?
• Focus on the few key metrics that are actually important, so that people aren’t distracted by unimportant reports or overloaded trying to determine what is important.
• Set up simple rules on how to identify when something is an outlier versus when something is truly wrong.
• Teach people which indicators they can react to immediately and which ones to escalate. If they know what they can do themselves, the problems are taken care of more quickly and you don’t waste time analyzing it.
• Create a simple plan of how to escalate these identified problems to a problem solver when the monitoring party can’t fix it themselves.

Advertisements