As someone who hasn't worked for an analytics firm, but rather in the analytics and business improvement functions within a company, this is, in my view, one of the toughest and, unfortunately, most critical elements to the life of a data analyst.

Whether they are willing to admit it or not, Execs can be fickle beasts (not a complaint, I plan to be one soon). Financial dashboards and short term decisions are all too common. The challenge is continually proving the worth of Analytics, before a short term decision crops up and results in budget reassignment and goodness knows what else.

It’s probably fair to say that the data and tools at your disposal will be sub-optimal, and you can't expect that to change immediately, so it’s down to you to make the data sing and demonstrate to the business how they can spend smarter to save more.

Right now I'm going to put my hands up. I don't have a magic, one size fits all solution; but when I do, rest assured the patent will be lodged before anything else happens. What I will do is look at a suggested approach to get it all across the line.

Firstly, education – and I mean education for the analyst, not the exec. Results and communication are the things you’ll live and die by in your career. You know how to generate the results, but the conversations you have with those that decide on your results are a different matter. I have met plenty of data analysts that can engage a business person well, but I’ve met just as many who can’t. In his June 2011 article, What Data Miners need to learn, Dean Abbott suggests that business education/learning is almost more important to a data analyst than anything concerning stats and mathematics, and he’s right. To be successful in a business setting, you need to know how to communicate with your bosses, that means, you need to know how to relay information in a way that invokes the action you want. It’s incumbent upon the analyst to impart, not the exec to understand.

Now, let’s talk about the process itself. For simplicities sake, I’m going to work through “an idea on a project you’d like to do” and you can pick/modify etc the parts that suit your scenario. I will more or less ignore the actual analytics part, it’s not my strong suit and you all know how to analyse data. I will focus on the “business” steps that help get everything ticking along.

Step 1. Early Engagement. Once you think you have an opportunity to improve the company's position with Analytics, then start speaking to whichever executive or senior manager you think will get the most out of the opportunity. Speak to them about what you've seen, what your idea is and what it could bring. Don't commit to numbers unless you are certain of them; just ask for time and support. You'll need this person every step of the way, so make sure they've got clout. It's also important that you are speaking to them at every step along the way. If you can get them emotionally invested here (and by that I mean excited about the possibilities) they will move the business equivalent of heaven and earth to make sure you have the best chance of success.

Step 2. Source your data. Hopefully your new sponsor will already have what you need, or you'll be able to source it from somewhere. Worst case scenario, you need to set up a collection mechanism. If that's the case make sure you get the smallest data set possible that will still enable you to succeed. In some cases that'll be a week's data, sometimes a month. You aren’t going to have the time to work on this for 3-6 months without getting a wider involvement from the company and that tends to breed interference.

Step 3. Set realistic, qualified expectations. It is important and might even be part of step 1. The key words here are realistic and qualified. Depending on the size and performance of your company, you may not have money or time to burn so make sure you are realistic about what you can achieve and clear that with more investment in tools, data and time that much more is possible. You need to do this part as early as possible and tactfully bring it up as often as possible.

Step 4. Use a suitable tool – not an all singing and dancing one. I'm going to work on the basis here that you don't already have a top shelf Analytics tool available and have no budget to get one. I'll also assume the open source ones like r are out of the question. Look for trial copies of tools that you can run on available hardware (generally your desktop) either as an add-in for Excel or as a standalone product. I’ve used Tableau in the past, it’s intuitive and easy to get going quickly.

Step 5. Make the data sing. I shouldn't need to tell you how to do this, and if you want some tips, keep an eye out for IAPA's education packages.

Step 6. Deliver on the findings. This involves a couple of things. Writing a killer report on what you've found and how the organization can use the information to improve and then working with the stakeholders in question to deliver. Make sure you measure the changes and report back an ROI. When that number comes back good, you're in the box seat. It’s simply a matter of making sure that word spreads and again, that’s what your exec is for. It should result in questions and discussions around what else is possible and what investment (read better tools and more time) is required by the business to get more.

Sounds easy doesn't it? Let me know when you win...


  • Tim Chewter

    Tim Chewter

    Manager Cunningham Lindsey
  • ·
  • 4 years ago
Make the data sing - thats it!

Comment on this article

You must login or signup to comment on articles