Voice of the DBA

Tag: analytics

  • Big Data Downsides

    Companies often want more data to help them make decisions on how they run their business. There has been this quest to gather and analyze as much data as possible to increase the efficiency of their operations to help reduce costs or increase profits. This has led to the importance of data as an asset, and the need for more data professionals in many organizations.

    That’s good for many of us that work with data.

    However, using data to try and improve your efficiency has a downside. It can lead you to a very narrow focus in your approach. That can be good in narrow, well-defined areas, such as minimizing the distance driven or packing containers. For less focused tasks, such as telling a story or writing code, this can mean you get stuck in a rut and limit your opportunities to improve.

    There’s an interesting article about big data and Hollywood, specifically looking at the types of products produced. Big data analysis leads companies to aim for the most effective types of movies that make money. Good for a company, not so good for society. Arguably, not even good for a company over time as people will tire of the same story, or type of story over time. Eventually, making simple decisions based on past data will start to fail.

    I can see the same thing in other industries as well. Using Big Data to drive decisions can help, but many of the areas where we use these techniques will evolve and change over time. The way we solve problems with code change over time as we develop new tools, techniques, platforms, languages, etc. There isn’t a perfect way to design a database or write a CRUD app precisely because new capabilities or new possibilities emerge. You could say the same things about marketing, manufacturing, medicine, and many other endeavors.

    This isn’t to imply big data and complex analysis isn’t helpful or useful. It’s just not everything. We need to balance human input, with some creativity, some instinct, some diverse thought, and some guessing. Most importantly, we ought to experiment and learn, not only from what machines might extrapolate, but from how humans change their thinking over time.

    Find a balance, accepting some imperfection in your process and in the world at large. Hopefully that will lead you to some success.

    Steve Jones

    Listen to the podcast at Libsyn, Stitcher, Spotify, or iTunes.

  • Staying Successful

    This is an interesting piece on data predictions. There’s a college that’s using data to help students, but not in the way you might think. Rather than customizing learning plans or finding more efficient ways to teach, the college is analyzing data to find students that might be struggling with college life overall and intervening early to help them cope with college and continue to move forward in their studies before they drop out.

    This is similar to something HP has done, trying to predict if employees might leave and intervene to see if something can be done. I heard an interview with the HP managers, who praised the program, and found they were among the most likely candidates to leave the company. However HP also realized that identifying people likely to leave allowed their HR people to try and find creative ways to retain their employees.

    I’m not entirely sure this is a process I’d like to be a part of, but it does seem like HP is doing well in their experiment. They realize they can’t solve every issue and keep every employee, but they can work proactively to try and reduce the losses. However privacy is a big part of this data analysis and I hope HP keeps that in mind as they move forward.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 1.8MB) podcast or subscribe to the feed at iTunes and LibSyn.

  • Managerial Moneyball

    I really enjoyed reading Moneyball. Its a book about baseball and data and how information should be used to choose baseball players for the Oakland As. It’s an interesting approach, one that has rarely been used in the sport in the past, though it is gaining traction. It does seem that this approach has helped the As to high level of success given the constraint of their limited payroll. There’s even a great movie if you don’t want to read the book, but the book is really much better and goes into more detail on data points and how they are used.

    The idea of using data to make decisions has been applied to other areas, with “The Moneyball Effect” being talked about in other industries. Recently I also ran across an opinion piece on bad managers that also referenced Moneyball. The piece notes that most people make poor managers. They lack the skills, and more importantly, they really lack those innate qualities that motivate, inspire, and engage employees. Whether you agree with that last part, I think most of you agree that most managers are poorly chosen, trained, and certainly not qualified.

    The idea of using data to identify people that would make good managers, and perhaps even move people out of managerial roles. The premise of the piece is really the bad managers make their teams perform worse, so if you’ve got one of the seven-out-of-ten people ill suited to the work, you should move them out of that position. Then identify, promote, train, and support the others to manage your employees and help them to perform at their best.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 2.3MB) podcast or subscribe to the feed at iTunes and LibSyn. feed

  • The Value of Data

    How valuable is your data? It’s a good question, and certainly the type of data your organization has along with the business in which you are engaged will make your data more or less valuable. More and more we find the differentiation between companies is in the way they collect, manage, and use the data available to them. So much in business is based on guesses, but more and more the guesses have some basis in data. We are starting to see those who make decisions in business feel some need to justify or support their choices with data.

    Is data the new oil? Oil was arguably the most important commodity of the twentieth century (and perhaps still is). The SQLRockstar wrote a piece with that same title, with the idea that knowing more about how valuable data can be will make you more successful in business. The post is based on the review of Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, and talks about some of the challenges of using data to make decisions.

    I certainly believe in the power of data, and that more data often gives us more insight into how the world works, as well as allowing us to draw some inferences about the future. Not necessarily better insight, but certainly more. I do think that computer extrapolation of patterns to the future is vastly overrated as most of our algorithms are far too simple, using too little data and discounting the increasing effects of small variables as scale increases. In short, I don’t think we’re anywhere close to a Foundation-like computer that can help us predict the success of new products, much less the future of a country.

    However I think that using analytics to make small decisions, and help guide our directions is important. We will still need humans that apply their internal supercomputers to interpret data, and continue to evolve the algorithms, and I hope that more and more of you are gaining deeper industry insight in your particular field. After all, many of us data professionals will be needed to help guide analysts in gathering, transforming, interpreting, and displaying data in ways that allows us to make decisions with more confidence.

    Steve Jones

    The Voice of the DBA Podcast

    Listen to the MP3 Audio ( 2.6MB) podcast or subscribe to the feed at iTunes and LibSyn.