Category Archives: Data Management

More data doesn’t always equal better decisions

There is currently a lot of interest in Big Data. We can now capture and store immense amounts of information about almost any process, include how humans behave. Buying and browsing habits drive internet advertising at a personal level. Face-recognition software allows companies and authorities to track people’s movements and activities. GPS tracking on cell phones shows where we have been, and when we were there.

In the supply chain world more data has often been seen as an advantage. The thinking is that the more we know, the better our decisions will be, since we have more facts to use as the basis for our choices. In general this is true. However there are at least 3 factors that can reduce the value of the data.

First, in many cases we have reached the point where there is more data than we can effectively analyze. Time constraints often require that decisions be made before all the data is available. So rigorous analysis has become a luxury. And standardized reporting has replaced ad hoc reporting.

Second, more data can hide valuable data. It’s not more facts but the right facts properly interpreted that allow for good business decisions. This is where experience can trump data: the numbers may look good, but if your gut says the decisions being made are not right, it’s worth taking another look at the numbers.

Finally, data is subject to interpretation. An instock figure of 99% may be good for a commodity item, but not for a seasonal item that is approaching the end of its annual cycle. And the way data is presented can skew how it is interpreted. If a presentation looks too good to be true, ask to see the raw data and have the presenter to walk you through the process used to arrive at the final data. It may not be as objective as is appears. And too often politics rather than data drives decisions.

In short, more data will not automatically lead to better decisions. Sound decisions require data, clear thinking and time. And in many cases these are shortchanged. It’s no wonder then, that people are often disappointed in the results of decisions made only on the volume of data available.

Analysis must be action-able

Managing a supply chain successfully requires gathering data to help monitor the performance of the various links in the overall chain. Much of this reporting is routine, such as sku performance, shipping accuracy, on-time delivery and key receiving accuracy. These reports are useful for monitoring, but they often don’t tell us how to improve performance. For this we need more analysis.

In my view all analysis must be designed to support change, that is, it must be action-able in a way that adds value to the process it evaluates. We could also say that all proper analysis always seeks to answer a question: how can I improve this process? Where are the bottlenecks? What players need to know about this so that it can be resolved? There is no need – and rarely time – for analysis that simply restates what we already know.

Out of stock reporting is a good example. That a product is out of stock is not helpful. We need to know why a product ran out. Was it unexpected sales? Incorrect inventory counts? Short-shipments? We can only correct a problem if we know the root cause.

So when you analyzing data, remember that the goal is to envision how to improve a process, and not merely to impress others with elaborate spreadsheets or fancy presentations. Provide value by showing a better way to manage the business.

Data doesn’t always equal intelligence

Data drives many modern supply chain processes. Measurement of the time, equipment and costs involved in these processes allows for optimization and improvement. What can be measured can be improved.

It is easy to collect lots of data. What is hard is to know what data is valid, and how to interpret it so that it can be used to improve the processes. More data doesn’t always equal a better process. In fact, there is a point where collecting more data actually increases the noise in the data and makes it harder to interpret.

Data properly validated and interpreted equals intelligence. And intelligence must be action-able. It must tell us what we need to do differently.

For example, collecting lots of data on vendor lead time variability might seem like a good idea. But by itself this data won’t tell us the whole story. We need to know about all the factors that could impact this data, and decide which factors are significant and manageable, and which factors are outside our control. Penalizing a vendor for poor lead time performance based solely on data measurements may hide the fact that our own processes are keeping the supplier from improving. I know of a case where the vendor’s lead time increased dramatically because the company that was ordering product from them was also arranging the freight pickups from the vendor, and was not scheduling these correctly. Not the vendor’s fault here.

So remember that data by itself frequently won’t tell you how to improve your business. Only the intelligent interpretation of the data will allow you to see where you can take action to improve your processes.

People – not systems – make supply chains successful

The supply chain world today is increasingly driven by data and systems. This makes sense as the goal of every supply chain is to find the lowest operating cost structure. And systems allow us to track and measure many more cost variables than we ever could in the past.

But merely accumulating data doesn’t necessarily lead to improvement. Someone has to validate and interpret the data, as not all data is reliable. So no matter how sophisticated our systems may become, there will always be room for intelligent people to play a role in supply chain success. In fact, one of the curses of a complex supply chain is that it is very brittle. Small changes cause disproportionate disruptions. Changing a cost or a ship point becomes an Olympic event. Rerouting shipments requires extensive manual interventions and levels of approvals. Shifting customer demands frequently require manual overrides.

Ever try to change an online order with Amazon? Better be quick.

Some of these disruptions are unavoidable. And only people fluent in supply chain tactics and systems can make these changes without tearing up the system or reducing its effectiveness. We have invested a great deal in our systems. I hope we will be willing to invest a similar amount of time and money is developing the talented people who will allow our systems to save money while at the same time remaining loose and flexible enough to respond to the changes required to truly serve the business needs.

You are only as good as your data

Supply chain and inventory management decisions can be complicated, and the results of a bad decision can be very expensive. We rely on data in many forms to help us make good decisions. But how often do we question the reliability of this data, and how often is it updated? Good data that is outdated is not much better than bad data.

It takes time to validate data, but if you are serious about your success, you owe it to yourself to take the time required to check your data’s accuracy. You don’t want to find out after the fact that your made your decision based on bad information. Go back to the team that compiled the data and ask them to explain how they got it. More importantly, ask them about any assumptions they made when interpreting the data.

As an example, I recently discovered that a report I was using was no longer supported by our IT teams. No one had validated the data for several months. Shame on me for not checking. The result was that I had been making recommendations that were completely misleading. So going forward I now have a policy that I won’t use a report that has not been validated within the last 30 days.

Get in the habit of questioning the data that everyone else is assuming is foolproof.

%d bloggers like this: