Who else wants to get a job in analytics? Part three – how to frame your experience
Hiring managers typically agree that they would prefer a candidate with real world analytics experience that created value rather than someone with the right education but no track record. However, that does not mean that all experience is equal. There are certain characteristics that hiring managers look for. Here is how you should talk about them in the interview.
Interviews come in all shapes and sizes. Some are conversational while others follow a script. Regardless of the format, how you prepare should not change. Below in bold are the questions you should have an answer for when discussing your experience.
Can you tell me about the most impactful or challenging project that you have worked on? This is how behavioral interviews start. When you craft your answer you should keep in mind the follow up questions that may or may not follow. How well you answer the all questions differentiates between good and great interviews.
Note: I should preface that not all of the questions I pose below will be asked in an interview. An inexperienced hiring manager may not know that these are things they should be asking or looking for. However, if you drive the conversation along these story lines your experience will look great.
How did you frame the problem? Framing an analytical solution to the problem is half the battle. Managers are looking for evidence of hypothesis driven problem solving. Consultants would say they use this method of problem solving to avoid “boiling the ocean.”
Sometimes the organization will frame the problem for the analysis team and they have no choice in the original framing. If you are in this situation you will have almost always need to adjust the framing of the problem. That change is what you want to highlight.
How did you define success? This is a similar and related question to framing the problem, but there is a difference to highlight. Many analysis programs fail to deliver value to the business (150 data scientists and no value). One of the many reasons for this is a failure to define what success looks like beforehand. If I ask you this question I am looking to see if you understood that issue and sought to mitigate it. Measuring success is also key to sustaining success. Did you produce an interesting and cool analysis that was useless, or did you produce results that translated into actions whose impact was tracked and measured?
Metrics for success can be given at the outset just like framing of the problem. If you were able to discover that the metrics were irrelevant or suboptimal and designed new ones that were useful, that would look great in the interview.
How did you choose your method of analysis? There is more than one way to answer most problems; I am looking for why you chose the one you did. There are lots of good ways to answer this question. If you realized that one method was far more likely to achieve results than others, it will sound great when you explain why. If outside constraints such as data availability only allowed for one course of action, the fact that you recognized that looks good as well. If you figured out a way to remove outside constraints to allow a more powerful analytical solution, that is rock star level stuff that you should mention.
What did you do when something went wrong? Every good analyst has had multiple times when a great plan did not work. Sometimes the second or third try did not work either. If you highlight how you overcame difficulties by re-framing the problem, identifying new data sources, revising your assumptions, etc., it shows you are the type of person who can handle difficult environments.
How did you handle problems with teammates or stakeholders? This may not be relevant to every story. However, if you do not have any stories about people issues and how you overcame them it makes you seem like a liar or unaccomplished. Difficult problems cause issues in teams. People get stressed; they might misunderstand their part of the work. Stakeholders may feel threatened by the results. People may fear losing their jobs. Analysts often work in teams, and they always depend on outside stakeholders or for something. Hiring managers want to see that you have the soft skills to handle working in that environment.
What were your results and how did you communicate them? The best analysis in the world is worthless if it is unable to inspire action. Poor ability to communicate results is not always a deal breaker. Many analysts are poor communicators. However that is a deficiency they need to make up in other places. If a candidate is a strong communicator that sets them apart from those who are not. Teams cannot afford to be without a strong communicator, so highlighting these skills may be the difference that gets you hired.
What impact was made? The best answer is one where you show what your analysis found, why it is important (often referred to as the “so what”), and how you developed and implemented a plan to do something about it. There are times that great analyses and plans are not acted on, and that is ok. If your work was great, but nothing came of it, it is best if you can talk about why without sounding bitter or angry.
The best answer is one where you overcame initial objections. Re-framing the analysis or adjusting the recommendations to address objections is often needed. No one gets it right the first time, every time. How you react to that situation is what I am interested as a hiring manager. If you can do that, and get stakeholders to agree, that is a great story.
The first and last questions above, how to frame the problem and what impact was made, are the most important for you to answer. Those two questions are the minimum acceptable answer. After you answer those two, the more other questions you answer the better.
There is one final point to make, and that is about the level of detail. If you are telling the whole story uninterrupted, your level of detail should be deep enough to cover all questions in around 90 seconds. Do not go into much detail if not prompted. However, if you are asked for more details, you must be able to talk to each of these questions in depth.
But why is experience better than education?
I am sure not everyone will agree with me, but in general most hiring managers favor experience over education due to proof of impact. The major difference between the business and academic approaches to analytics is the mindset of the individual conducting the analysis. Mindsets are driven by incentives:
In academia professors are incentivized to keep publishing. No one will publish an article that says we did not find anything conclusive. Similarly, no one will publish an article that says, “it looks like X is true even though our results were inconclusive.” In business, both of those results could be valuable. Here are a couple of ways that could work out in the real world:
“We thought discounting drove our sales, but this analysis shows that the evidence supporting that hypothesis is weak. We should test the effects of reducing discounts.”
“While the lift in sales due to our new loyalty program was not statistically significant, the increase in gross margin was large enough to justify rollout to other regions.”
Both of those situations can be highly valuable to a business, but an academic approach may not identify them. Of course that is not always the case, but that is often the perception.
There is one other reason that experience is often preferable to an academic pedigree. In a classroom setting the data is often clean and structured in a useful manner. In the real world, data cleaning and preparation can take up most of the total analysis time. Academic research can have some of the same problems as real world business data though.
The most amazing experience in the world will not get you hired if you are unable to show hiring managers why it is amazing. Not all managers will ask these questions, or even know that they should, but they will all be impressed if you answer them well.
Next time we will discuss what to do if you do not have any experience in analytics. The catch-22 of needing experience to land a job but not having a job where you can get that experience exists everywhere. Fortunately I believe it is easier to counteract for analytics jobs than in many other areas.
Did I miss anything? Do you disagree about the importance of these questions? Leave a comment below if you have anything to add or this spurs any questions.
 If you wanted a cup of boiling water you could pour one cup of water into a pot and boil it or you could boil the ocean and then get a cup. Hypothesis driven problem solving allows us to save a ton of time by being specific rather than trying to solve huge and broad problems aka “boiling the ocean.”