These days we hear a lot about analytics, metrics, data driven decision making and so-called big data. While the theory of making decisions based on data makes intuitive sense to most managers, it is challenging to know where to start from a practical standpoint. Data collection and analysis can be very challenging and managers can consume copious amounts of time getting overwhelmed. In this post I provide a quick overview of the rationale for analytics followed by a set of practical steps to gather and use analytics to improve your business operations.
Kai What!?
Kaizen is Japanese for a philosophy of continuous improvement. After World War II this was implemented at several manufacturers in Japan as a way to boost the industrial production. It later made its way into the mainstream of business as part of Toyota Motor’s famed manufacturing system. At its core, Kaizen refers to a relentless pursuit of improvement in engineering, manufacturing and other business processes. The reasoning goes that by continually improving anything, just even a tiny bit every week, month or year, organizations reap huge rewards in the long run as a result of the compounding effect. In addition to the magic of compounding, it seems to me that the other secret to Kaizen’s success is its emphasis on incremental improvement. Because it is accepted for the improvements to be small by definition, they seem much more attainable and therefore often are attained. Sounds simple right? In fact it is not that difficult to implement Kaizen whether you call it that or just plain old continuous improvement as long as one crucial component is present: the ability to measure, to measure if a process, system, operation or team is getting better or worse along a certain dimension. And this is the crux of the matter. It is not that most managers or workers don’t aspire to improve the way they solve problems, deal with uncertain situations or gets rid of bottlenecks; it is just that they often cannot tell if things are in fact getting better! In the absence of this crucial feedback loop we can hardly expect managers or their teams to steer things in one direction or the other. And this is where simple understandable business analytics play a hugely important role.
Analytics Are Everywhere
If you ever watched the movie “Moneyball”, you’d see the importance of analytics, numbers and data to any operation, even to the success of a baseball team. Yes, it sounds bizarre that a game with endless variables can be predicted and improved but this can be done and is now common practice for all major sports teams. Or how about Nate Silver – this was the guy who predicted the outcome of the 2012 U.S. presidential election correctly for every state! I will not even attempt to explain the methods he used for his pinpoint accuracy, but he did it with analyzing and tracking statistics.
The Road to Success is Paved with Tracking and Analysis
Constant improvement is a formidable challenge for any manager. But this is especially true for those who oversee distributed operations involving multiple job sites, plants, facilities, people and equipment. Examples of these folks include industrial operations managers, facilities construction managers and equipment field service managers. Analyzing and tracking issues to see where the reoccurring problems are, where the bottlenecks are and where the bulk of the expenses lie is even more difficult for this breed of managers when their operations are strewn across vast distances, even multiple time zones. For these folks it is not easy to hold regular meetings with staff or personally walk around and get an overall sense of their operations as it would be perhaps for a manager overseeing a group of office workers housed in the same building pushing bits and bytes around. As one example, for a plant manager, a trivial yet recurring production issue on the plant floor may be a top contributor to production costs. For a field service manager an issue with a poorly designed cable harness or connector may turn out to be the root cause of a large proportion of customer complaint cases. But how would these managers become aware of these seemingly mundane issues taking place out there away from the HQ? Without consistent data collection, organization and analytics bringing these little facts to their awareness, they are at best relying on hearsay and anecdotal evidence which is often skewed by people’s perceptions and biases.
As a manager wouldn’t it be great to know exactly how your business units are doing, how your teams are doing, and whether the operations are improving? You might say: we have weekly meetings to catch up on our progress and operations and as far as I can tell if we are improving by looking at the financials at the end of the month. Everything looks groovy, but underneath the high level numbers, a lot of rich, subtle detail is yet to be uncovered, analyzed and put into action improving your performance. This is the world of analytics.
How to Get Analytics to Work for You not Against You
Based on our experience of developing and deploying analytics below are some key recommendations to help you implement and leverage analytics to improve your team’s performance without getting mired in a ‘data :
a) Pick the right high level metrics: A good way to think about this is to imagine your operation, your team, your plant as your favorite muscle car, now imagine which three gauges you would absolutely want to have on the instrument panel? And I am not talking about the obvious revenue-in, expenses-out. These would be the equivalent of the speedometer or electronic compass. They give you an overall sense of where you are headed and how fast you are moving, which are important but they don’t really give you actionable insights: things you can actually do to improve things. When thinking about metrics you want to dig a bit deeper below the surface and think of indicators or warning lights you want to have in plain sight to tell you and your team what’s happening inside the engine. From our experience talking to customers and developing such metrics in context to our Scoop platform, we homed in on three primary metrics: volume of issues by category, issue closure speed and team-wide activity. When selecting your ‘gauges’ it pays to be a minimalist. Like a busy instrument panel that distracts a driver and fails to convey the main signals, too many metrics to monitor and improve also makes it difficult for managers to benefit from analytics.
b) Collect the necessary raw data and keep it coming: to light up those gauges we talked about, you need the right electrical signals to be collected and fed into them. Likewise to present the right analytics to managers, the right underlying data needs to be collected out there in your operations. This is largely a function of two elements: tools and training. The right information technology is needed to allow collection of the data at the place and time when the action is taking place. For instance if the data is related to a service call, the technician needs to have a handy, easy to use tool to document different aspects of the call such as type of equipment, type of issue, images of the faulty component and the like. If the tool he is using is not accessible at the service site e.g., because it runs only on his laptop and which is not convenient to fire up on the spot or requires a WiFi connection that is not available at the time, or doesn’t enforce collection of the right data fields etc. then the process begins to break down. Some people will collect the right data, some won’t, some intend to collect it but forget and other reasons abound. Now, although the right IT tool is a necessary component it is not on its own sufficient. You still need to train the people and this is not just training in the sense of steps to use to operate the IT tool but also a cultural adjustment to a metrics collecting, metrics driven organization.
c) Ensure your data has structure: one of the primary challenges faced by industry today is what is referred to as ‘big data’. Despite its somewhat positive sounding name, this is basically a big hot mess. Thousands of Terrabytes of data is being collected by organizations around the world every day. Most of this data has little or no structure. Examples include–you guessed it, email, documents, image and video archives that are not organized or sorted in any specific way and simply get stuffed into data warehouses and backup archives. As you can imagine, it is very difficult to draw reliable conclusions from a collection of emails or a huge blob of files. In our service manager example, it would be very difficult if not impossible for her to sift through thousands of emails and documents exchanged by her team in the past quarter. So when it comes to data collection, it pays to choose a tool that guides the data along the right pipes and sorts it into the right buckets from the get-go. A number of electronic form-based systems (including our own Scoop platform) accomplish this by allowing users to gather a minimum set of data fields and then storing these in a relational database. Additional fields can be introduced over time to add to the richness of the data collected. The end result is that the data is much easier to analyze, sort, filter and use to drive actionable conclusions.
d) Use analytics to form theories then test & repeat: of course all of the data collection, summarization and presentation is useless if it doesn’t inspire action, action to change things for the better. Taking a methodical, measured approach to using analytics is important. It is important to remember that the gauges are just reflections of what is actually happening in your operations not the absolute truth. Therefore you want to use the metrics provided to first form a theory or ideally a number of theories about what is causing a certain change in your business. For instance if you are seeing a large volume of issues related to a specific model of the equipment your team installs in the field, you don’t want to jump to conclusions and blame the problem on the equipment-that’s just one theory. You also want to form a number of other theories about the root cause of the problem including installation procedure, level of training and other potential causes. Then use the existing data or collect additional data to rule in or rule out each theory.
What type of analytics and tracking systems have you used to monitor and improve your operations? What are the pros and cons?
About the Author: Babak Sardary is a veteran of field engineering and founder of Trusterra Technologies. Trusterra develops the Scoop™ mobile software platform helping thousands of users in distributed teams record, alert, collaborate and resolve issues across field and office.






