Although analytics is a valuable specialty that every large organization needs, it’s not always full appreciated. We analysts hate that, but sometimes, it’s our own fault. Our own bad habits can hold us back.
Be honest about these seven terrible habits and take corrective action to have greater impact and get more appreciation in the workplace.
Ignoring management priorities
Analysts don’t get to run the operation. We’re staff, and our purpose is to provide information that supports management decision making. That means the highest priority obligation for any data analyst is determining what’s important to management, and why. If you don’t have a clear understanding of your management’s reward structure (in other words, what has to happen for the boss to get maximum pay), business goals and the scope and limitation of responsibilities (what the manager can and cannot control), you won’t be able to assure that the information you provide is a good match.
Corrective action: Spend a few minutes each day learning about the general state of your organization. Actually read some of those emails from management, see what’s in the newspapers, ask questions when you meet with managers. As you work, consider how the project and your approach to it fits in with broader goals, and look for ways to make the information you provide more relevant, clearer and easier to use.
Disregarding complementary business functions
It’s no fun to discover, after you’ve sunk a lot of time and energy into developing a predictive model, that operations has no way to incorporate it into everyday business practice. It’s pointless to recommend a course of action that requires resources that will be unavailable when you need them. It’s embarrassing to announce remarkable findings from your latest analysis, only to hear someone explain that the data doesn’t mean what you thought it meant, or worse, that you’re using data in a way that violates the law or contractual requirements.
Corrective action: Get out of your cube. Reach out to your peers in IT, operations, sales and other roles, even if they do not seem closely related to what you do. Look for opportunities to meet and get to know your colleagues in low-stress situations, perhaps over lunch. Ask about their jobs, what they do and how it helps the business. Take the time to learn why things are done the way they’re done and not done the ways they’re not done). You’ll know more going into every project, and establish a network of partners to help you develop more effective multidisciplinary approaches to analytics.
Neglecting process
What’s your analytics process model? Do you use CRISP-DM or SEMMA? Do you have a process of your own, and if so, is there a written guide that explains it? Does this all sound strange and unfamiliar?
Corrective action: A good analytics process model guides you through the many tasks that you should perform and the details you should cover in each of them, so you avoid leaving things out, repeating work or misunderstanding what has and has not been done. If your organization already has a standard analytics process, learn it well and follow it. If not, get familiar with CRISP-DM, the most popular analytics process model. It’s an open standard that’s available to everyone, free to use and flexible, so you can use it in any industry, with any tools you prefer.
Failing to document
Remember that project you completed last year? No? Sure you do. It was the thing, the thing you did right after that one thing, and before the other thing. We need you to get right back to that and pick up where from where you were when you went on the other thing. You documented all of that, everything you did, right? It’s important, because we’ve got to move fast on this, there’s no time to waste.
Corrective action: Using a good analytics process model helps, because the process model describes the tasks you must perform and the required documentation for each of them. (The trouble is, many analysts ignore most the parts where you write down the details of what you did. You know, write down what you did, and the results, what worked and what didn’t. In words, sentences, paragraphs and so forth.) Be fastidious about documentation, to make the most of your time, provide hard evidence of what you’ve done, and make it easier for others to carry on as necessary.
Speaking “Geek”
Upper management is very interested in numbers, as long as those numbers represent dollars, Euros, yen or some other form of money. Otherwise, math is not of much interest to them. Technical terminology, details of your models and other mechanics of your daily work do not interest them. They are busy people who don’t answer to you, so if your presentation doesn’t interest the Big Boss, you’ll quickly find that Big Boss doesn’t listen, or even remain in the room.
Corrective action: Review the discussion of management priorities above. Focus on what’s important to your executive management, use plain language and get to the point quickly. It will not hurt to take classes, or read some books, on technical communication.
Underpreparing for scrutiny
Executives are allowed to ask questions. Some use that power for good, others for evil, but either way, if you can’t provide a simple, direct and useful answer, you’re going to look like a fool. This is no secret, yet it’s awfully common to witness presentations where analysts can’t answer questions that are reasonable, and which should have been expected.
Corrective action: Review all the discussions above. Review every bit of material you intend to present and yourself what supports your conclusions and what may be missing. Run it by your new friends in other business areas. Prepare supplemental slides, bring along your notes, and learn you own material well. Don’t do a big information dump, yet be ready with resources you may need to respond to probing questions.
Suffering from delusions of grandeur
Statisticians are sexy. Data scientists have skills as rare as a unicorn. Full-stack programmers can leap tall buildings in a single bound.
Corrective action: Please get over it.
