Posts Taged data-driven

The One Year, Thirty Minute Challenge :: Week 36 :: Technology :: Big Data

According to Statista, in 2010, the total of amount of data collected worldwide was 2 zettabytes – or to use a unit of measurement you might be more familiar with, that’s 2 trillion gigabytes. In 2024, that number is projected to hit 149 zettabytes. All that data isn’t kept, so IDC predicts that by 2025, the world’s accumulated datastore will be 175 zettabytes. According to Forbes, we (collectively) generate 1.7 megabytes of new data per person, per second. And here’s maybe the most interesting fact of all, according to IDC, less than 5% of that data will be analyzed.

So, why is any of that important in the world of the One Year, Thirty Minute Challenge? Companies who capitalize on the data available to their organization by –

  • Identifying what parts of that data directly impact their financial performance
  • Making meaning of that data with expert analysis
  • Turning that analysis into actionable insights
  • Changing organizational behavior based on those insights
  • Measuring the financial impact of those changes
  • Making additional changes based on those measurements

are seeing results. Here are a few examples from Tech Republic.

  • Supply chain safety and theft detection enables companies, with help of item-placed sensors and business intelligence, to reduce in-transit theft rates of supplies from 50% to 4% and to detect when the environmentals or seals on shipment containers have been compromised.
  • Logistics tracking and routing using business intelligence and machine-based data/sensors optimize delivery routes and driver habits creating fuel savings and better service.
  • Collections work at companies is avoided by learning more about customers who are behind on their payments through big data aggregation and business intelligence that can predict who in good faith can pay their debts with a little help–and then helping these customers keep their purchases and keeping companies from having to write off defaults.
  • Buying habits and preferences of consumers are better understood and lead to increased sales.
  • Predictive maintenance enables urban tram systems to stay online, reroute traffic where necessary, and flash adviser alerts to customers over their mobile phones while repair crews are dispatched to replace faulty components before the components actually fail.

Big Data doesn’t just refer to just the volume of data available today, it encompasses the “4 Vs” of Big Data –

  • Volume – Certainly volume is an important part of the equation. We have internal data from our CRM and ERP systems that tell us about vendor performance, product performance, customer behavior, employee performance and a host of other things. We have external data from social networks, online review sites and more. Because of the Internet of Things (IoT), we have data that originates not just from the actions of our employees or customers, but from inanimate devices connected to the internet. So, we can know the number of times a door opens and for who, the temperature inside a shipping container and when a client’s copier is low on toner.
  • Variety – This data comes at us in multiple ways. Structured data from internal systems where we’ve controlled what is collected and how and unstructured data from external sources. We might get a text from a customer with a video of the dishwasher we just fixed showing us that it’s still doing what it was doing before we “fixed” it, a Google review, a Twitter DM, a reading from a sensor on our delivery van alerting us to a tire pressure problem and the list goes on.
  • Velocity – If the previous two aren’t enough, maybe the most daunting is the speed at which it comes at us. Last minute’s data reporting that all is well, is superseded by this minute’s data reporting a problem on the factory floor or a customer unhappy with your product or service. Multiply those by the number of inputs (customers, employees, vendors, sensors) and it can seem overwhelming.
  • Value – In actuality, this is the one that matters most. Of all the data collected by your organization, what really impacts financial performance, customer experience and employee wellbeing (ability to do their job effectively and efficiently)?

One more thing before we jump into this week’s exercise. Big Data requires different skills and tools than the traditional reporting you’ve pulled from your internal systems. First, because of the mix of structured and unstructured data, you’ll need a data management infrastructure that can manage both. Second, you need someone who can help you navigate this new world. You can hire a data scientist or you might opt for outsourcing this part of your work to a vendor specializing in Big Data Analysis. The most rudimentary Big Data analyses are looking for trends (each month for the past six months, distributors of our product in the Southeast have reported a stock out. Each month, it’s been earlier in the month than the month before), patterns (customers whose first purchase from us is product X never make another purchase, but customers whose first purchase is product Y have an 80% chance of being a repeat customer), and correlations (in the Fall, the first time the temperature dips below 50 degrees canned soup sales double and stay at that level until the first time the temperature hits 60 degrees in the Spring). A Data Scientist can help you start thinking in this vein. Third, in addition to the infrastructure tools to do the heavy lifting, you need visualization tools that help you easily see what this large amount of data is telling you. Even if the data scientist tells you everything you need to know, you want to roll this information out to everyone in the organization who can benefit from it. Good visualization tools will allow them to consume large amounts of information (and make sense of it) more easily.

For this week’s exercise, I want you to identify some problems or opportunities in the organization where Big Data-generated insights might make a difference. Here are some thought starters –

  • More callbacks on service calls – Is it the same technician? Are they working on the same brand of equipment? Are replacement parts from the same vendor failing at a high rate?
  • Inventory management is more challenging than it should be – Can you get access into distributor data so you can see when distributors are most likely to place a reorder? Is a single vendor slowing production with late or defective products?
  • Customers seem uninterested in a new product or service – What is the factory defect rate on this product vs the defect rate on its predecessor? Have customers who purchased the product commented on social media regarding the product? Is it especially unpopular among your customers who purchase another product from you?
  • We have too many employees during some shifts and not enough during others – Can you examine sales by hour for the same day of the week last week or the same week last year? Can you examine the nature of sales during each shift – selling a hand-dipped ice cream cone is more labor intensive than selling a bottled soft drink?

Take your list and contact a Big Data company for a consultation. See if it makes sense to do a pilot project.

Big Data is the foundation for technologies like Machine Learning – the improvement of computer algorithms through experience (people who bought this book also bought this book, powering your Amazon recommendations) and Artificial Intelligence – when a system “perceives its environment and takes actions that maximize its chance of successfully achieving its goals” (think Big Blue playing chess against a Grand Master, examining the chess board and making the optimal chess move).

Of all the technology assets in your company, data is the most important. It catalogs the past behavior of your employees and customers. And the best predictor of future behavior is past behavior. Don’t neglect the power of this asset to solve problems that have puzzled you for a long time.

The One Year, Thirty Minute Challenge :: Week 29 :: Governance :: Decision Making

Over the course of a day we make hundreds of decisions. Many, in the great scheme of things, are inconsequential – blue shirt or yellow shirt, mustard or mayo, checkout aisle 6 or 11. However, when we’re at work, some of our decisions might have a bit more impact – this new region or that new region or both, this new employee or that new employee, abandon this product or invest a bunch of money into marketing it for another quarter or two. These types of decisions affect the lives of people, the trajectory of our company and the amount of money we make or lose in upcoming quarters.

So what if we could get better at decision making? Let’s agree up front that every decision carries risk. We can’t “good decision” our way out of every fork in the road and remove risk. Most of us fall victim to, what those who study decision-making call, “resulting”. We believe if we get good results, we made a good decision. If we get bad results, we made a bad decision. Let me illustrate. The odds of winning on any given number at the roulette table are 1 in 37. If you walk up to the table, place your chips on 5 and the little ball goes into the 5 on the wheel, you might believe you made a good decision. In reality, you made a bad decision (the math was against you) but got a good result. Conversely, if you hire a salesperson with experience in your industry, stellar credentials, a history of strong sales and equip them with every resource they need to sell your product and they fail miserably, you more than likely made a good decision, but got a bad result. Just one of the foibles that we, as humans, struggle with as we evaluate our decisions.

So how do we up our decision-making game? In this week’s One Year, Thirty Minute Challenge, I encourage you to spend your thirty-minute exercise piecing together a decision-making framework that you’ll use when your organization is faced with a decision. I’ll give you some thought starters and you can grab what works for you and add your own.

  • What empirical data can we bring to bear on the decision? It’s easy to fall in love with people, products, places and processes. Can we put our hands on data that will give us objective information – sales numbers, number of defects, number of returns, sales by location, sales by hour, sales by salesperson, production per assembly line, bounce rate for the landing page.
  • How can I remove my ego from the decision? It’s tough to divorce yourself from a person or project that you’ve poured yourself into. In reality, you are not what you do. You still have worth and you’re still smart, even if the object of your affection is looking questionable. Recognize this for yourself and recognize that others in the organization will have similar feelings towards the people and things they’ve invested in. Step away – and help them step away. Two more things on ego. First, we love our own ideas and struggle to see how they might have a couple of holes. Second, we love information (both empirical and anecdotal) that supports our position and tend to discount information that opposes our position. Be on guard against both of these things.
  • Enlist the collective genius of the people most affected by the decision. If I could list the most frequent management screw-ups, this would be close to the top – people not familiar with the intimate details of the work, trying to improve the work. In reality, the people who do the work are most qualified to improve it. Get input from employees, customers, and vendors – whoever can help you assemble the largest body of knowledge on the subject about which you are making a decision. One important thing – an outside perspective does help because people are occasionally so blinded by the forest, they can’t see the trees. But I’d err on the side of getting lots of input from those in the know.
  • Get help from someone who’s made similar decisions. The Israelite King Solomon said, “There’s nothing new under the sun.” True when he wrote it 3000 years ago. Still true now. Find someone who’s faced a similar situation and pick their brain.
  • Propel your self forward and look back. As much as you can, transport yourself to the end of every fork in the road (all the possible decision options) and look backward. Things might seem much clearer – after all, hindsight is 20/20. What would have to go right to get here? What could go wrong on the way to here? Can I live with the consequences of the things that might go wrong? What are the probabilities for each of these things going right or wrong? Conduct a pre-mortem – in your head, jump to the end of the decision, assume it failed miserably then ask, “What did we screw up that caused this?”
  • Would you put money on this? I wish I could claim credit for this idea, but it comes from Annie Duke’s brilliant book, Thinking in Bets. She encourages her readers to ask themselves, “would I bet on this?” This moves the discussion from theoretical to financial. Before we bet on something, we contemplate the probability (run or pass, cover the spread or not cover the spread). Our emotions (we love our team and hate the other team) are eclipsed by the reality of what could happen to our wallet.
  • Find a contrarian. Seek out someone to poke holes in the decision you’re narrowing in on. They can be inside or outside the organization. Encourage them to pick it apart personnel-wise, strategically, operationally, and financially.
  • Festina lente. Caesar Augustus adopted this motto – Make haste, slowly. Make decisions quickly but deliberately. Don’t fall victim to paralysis by analysis, but don’t fire from the hip. Good decision-making is thoughtful and complete but with a bias for action.

 

Following these steps, or any others for that matter, won’t result in perfect decision-making. There’s no such thing. We’ll still be duped by the poor decisions that have good results (thinking they were good decisions), puzzled by the good decisions that have bad results and feel smug about the good decisions that yield good results. Our best hope is that we optimize our methodology.