“Today, big data is about business disruption. Organizations are embarking on a battle not just for success but for survival. If you want to survive, it’s time to act.” – Capgemini and EMC² in their study Big & Fast Data: The Rise of Insight-Driven Business.
It seems like we’re in the midst of a data analysis crisis. Although organizations spend millions of dollars on collecting and analyzing data with various data analysis tools, it seems like most people have trouble actually using that data in actionable, profitable ways.
This might sound a little dramatic. However, consider the following statistics pulled from a Capgemini and EMC study that surveyed over 1000 senior decision makers in nine regions:
- 56% surveyed claim that their investment in big data over the next three years will exceed past investment in information management.
- 65% surveyed admit they risk becoming irrelevant and uncompetitive if they do not leverage data. This is especially true given that non-traditional providers, like startups thriving on big data processing, are moving into their industries.
- Although companies realize they desperately need to dig into data analytics to maintain their business position, 45% surveyed think their current internal IT development cycles are not sufficient for new analytics and don’t fulfill their business requirements.
- Making matters worse, over half (52%) of those surveyed see the speed of their organization’s insight generation from data analytics as constrained by its existing IT infrastructure.
Data Is Only As Good As The Questions You Ask
Based on this survey, it seems that C-level executives believe that big data is the ultimate cure for all their business ills.
However, the truth is that no matter how advanced your IT infrastructure is, your data will not provide you with a ready-made solution unless you ask it a specific question.
To help transform data into business decisions, you should start preparing the painpoints you want to gain insights into before you even start the data gathering process. Based on your company’s strategy, goals, budget, and target customers you should prepare a set of questions that will smoothly walk you through the data analysis and help you arrive at relevant insights.
8 Tips for Asking The Right Data Analysis Questions
Here at datapine we have helped solve hundreds of data analysis problems for our clients. All of our experience has taught us that data analysis is only as good as the questions you ask. Additionally, you want to clarify these questions now – which will make your future business intelligence much clearer.
That’s why we’ve prepared this list of data analysis questions – to make sure you won’t fall into the trap of futile, “after the fact” data processing.
1) What exactly do you want to find out?
It’s good to evaluate the well-being of your business first. Agree companywide what KPIs are most relevant for your business and how do they already develop. Think in what way you want them to develop further. Can you influence this development? Identify where changes can be made. If nothing can be changed, there is no point of analyzing data.
Consider what your goal is and what decision-making it will facilitate. What outcome from analysis you would deem a success? These introductory data analysis questions are necessary to guide you through the process and help focus on key insights.
Let’s have fun with a little imaginative exercise.
Let’s say that you have access to an all-knowing business genie who can see into the future. This genie (who we’ll call Data Dan) embodies the idea of a perfect data analytics platform through his magic powers.
Now, with Data Dan, you only get to ask him three questions. Don’t ask us why – we didn’t make the rules! Given that you’ll get exactly the right answer to each of them, what are you going to ask it? Let’s see….
Talking With A Data Genie
You: Data Dan! Nice to meet you, my friend. Didn’t know you were real.
Data Dan: Well, I’m not actually. Anyways – what’s your first data analysis question?
You: Well, I was hoping you could tell me how we can raise more revenue in our business.
Data Dan: (Rolls eyes). That’s a pretty lame question, but I guess I’ll answer it. How can you raise revenue? You can do partnerships with some key influencers, you can create some sales incentives, you can try to do add-on services to your most existing clients. You can do a lot of things. Ok, that’s it. You have two questions left.
You: (Panicking) Uhhh, I mean – you didn’t answer well! You just gave me a bunch of hypotheticals!
Data Dan: I exactly answered your question. Maybe you should ask better ones.
You: (Sweating) My boss is going to be so mad at me if I waste my questions with a magic business genie. Only two left, only two left… OK, I know! Genie – what should I ask you in order to make my business the most successful?
Data Dan: OK, you’re still not good at this, but I’ll be nice since you only have one data analysis question left. Listen up buddy – I’m only going to say this once.
The Key To Asking Good Data Analysis Questions
Data Dan: First of all, you want your questions to be extremely specific. The more specific it is, the more valuable (and actionable) the answer is going to be. So, instead of asking, “How can I raise revenue?”, you should ask: “What are the channels we should focus more on in order to raise revenue while not raising costs very much, leading to bigger profit margins?”. Or even better: “Which marketing campaign that I did this quarter got the best ROI, and how can I replicate its success?”
2) What standard KPIs will you use that can help?
OK, let’s move on from the whole genie thing. Sorry, Data Dan! It’s crucial to know what data analysis questions you want to ask from the get-go. They form the bedrock for the rest of this process.
Think about it like this: your goal with business intelligence is to see reality clearly, so that you can make profitable decisions to help your company thrive. The questions to ask when analyzing data will be the framework, the lens, that allows you to focus on specific aspects of your business reality.
Once you have your data analytics questions, you need to have some standard KPIs that you can use to measure them. For example, let’s say you want to see which of your PPC campaigns last quarter did the best. As Data Dan reminded us, “did the best” is too vague to be useful. Did the best according to what? Driving revenue? Driving profit? Giving the most ROI? Giving the cheapest email subscribers?
All of these KPI examples can be valid choices. You just need to pick the right ones first and have them in agreement company-wide (or at least within your department).
3) Where will your data come from?
OK – so far, you’ve picked out some data analysis questions, and you’ve found KPIs to measure them. Our next step is to identify data sources you need to dig into all your data, pick the fields that you’ll need, leaving some space for data you might potentially need in the future, and gather all the information into one place. Be open minded about your data sources in this step – all departments in your company, sales, finance, IT, etc., have the potential to provide insights.
Don’t worry if you feel like the abundance of data sources makes things seem complicated. Our next step is to “edit” these sources and make sure their data quality is up to par, which will get rid of some of them as useful choices.
Right now though, we’re just creating the rough draft. You can use CRM data, data from things like Facebook and Google Analytics, financial data from your company – let your imagination go wild (as long as the data source is relevant to the questions you’ve identified in step 1).
3.5) Which scales apply to your different data sets?
WARNING: This is a bit of a “data nerd out” section. You can skip this part if you like or if it doesn’t make much sense to you.
You’ll want to be mindful of the level of measurement for your different variables, as this will affect the statistical techniques you will be able to apply in in your analysis.
There are basically 4 types of scales:
Statistics Level Measurement Table
- Nominal – you organize your data in non-numeric categories that cannot be ranked or compared quantitatively.
– Different colors of shirts
– Different types of fruits
– Different genres of music
- Ordinal – Graphpad gives this useful explanation of ordinal data:
“You might ask patients to express the amount of pain they are feeling on a scale of 1 to 10. A score of 7 means more pain that a score of 5, and that is more than a score of 3. But the difference between the 7 and the 5 may not be the same as that between 5 and 3. The values simply express an order. Another example would be movie ratings, from 0 to 5 stars.”
- Interval – in this type of scale, data is grouped into categories with order and equal distance between these categories.
Direct comparison is possible. Adding and subtracting is possible, but you cannot multiply or divide the variables. Example: Temperature ratings. An interval scale is used for both Fahrenheit and Celsius.
Again, Graphpad has a ready explanation: “The difference between a temperature of 100 degrees and 90 degrees is the same difference as between 90 degrees and 80 degrees.”
- Ratio – has the features of all three earlier scales.
Like a nominal scale, it provides a category for each item, items are ordered like in an ordinal scale and the distances between items (intervals) are equal and carry the same meaning.
With ratio scales, you can add, subtract, divide, multiply… all the fun stuff you need to create averages and get some cool, useful data. Examples: height, weight, revenue numbers, leads, client meetings.
4) How can you ensure data quality?
Insights and analytics based on a shaky “data foundation” will give you… well, poor insights and analytics.
Remember – your data analysis questions are designed to get a clear view of reality as it relates to your business being more profitable. If your data is incorrect, you’re going to be seeing a distorted view of reality.
That’s why your next step is to “clean” your data sets in order to discard wrong or outdated information. This is also a good time to add more fields to your data to make it more complete and useful. Yes, this is annoying, but so are many things in life that are very important.
When you’ve done the legwork to ensure your data quality, you’ll have built yourself the useful asset of accurate data sets that can be transformed, joined, and measured with statistical methods.
5) Which statistical analysis techniques do you want to apply?
There are dozens of statistical analysis techniques that you can use. However, in our experience these 3 statistical techniques are most widely used for business analysis:
- Regression Analysis – a statistical process for estimating the relationships and correlations among variables.
More specifically, regression analysis helps one understand how the typical value of the dependent variable changes when any one of the independent variables is varied, while the other independent variables are held fixed.
In this way regression analysis shows which among the independent variables are related to the dependent variable, and to explore the forms of these relationships. Usually, regression analysis is based on past data, allowing you to learn from the past for better decisions about the future
- Cohort Analysis – it enables you to easily compare how different groups, or cohorts, of customers, behave over time.
For example, you can create a cohort of customers based on the date when they made their first purchase. Subsequently, you can study the spending trends of cohorts from different periods in time to determine whether the quality of the average acquired customer is increasing or decreasing over time.
Cohort analysis tools give you quick and clear insight into customer retention trends and the perspectives for your business.
- Predictive & Prescriptive Analysis – in short, it is based on analyzing current and historical datasets to predict future possibilities, including alternative scenarios and risk assessment.
Methods like artificial neural networks (ANN) and autoregressive integrated moving average (ARIMA), time series, seasonal naïve approach and data mining find wide application in data analytics nowadays.
We’ve already explained them and recognized as one of the biggest business analytics trends for 2017. Your choice of method should depend on the type of data you’ve collected, your team’s skills and your resources.
6) Who are the final users of your analysis results?
The last but not least significant of your data analytics questions refers to the end users of our analysis. Who are they? How will they apply your reports? You must get to know your final users, including:
- What they expect to learn from the data
- What their needs are
- Their technical skills
- How much time they can spend analyzing data
Knowing the answers will help you to decide how detailed your report will be and what data you should focus on.
Remember that internal and external users have different needs. If the reports are designed for your own company, you more or less know what insights will be useful for your staff and what level of data complexity they can struggle through.
However, if your reports will also be used by external parties, remember to stick to your corporate identity. The visual reports you provide them with should be easy-to-use and actionable. Your final users should be able to read and understand them independently, with no IT support needed.
Also: think about the status of the final users. Are they junior members of the staff or part of the governing body? Every type of user has different needs and expectations.
7) What data visualizations should you choose?
Your data is clean and your calculations are done, but you are not finished yet. You can have the most valuable insights in the world, but if they’re presented poorly, your target audience won’t receive the impact from them that you’re hoping for.
And we don’t live in a world where simply having the right data is the end all, be all. You have to convince other decision makers within your company that this data is:
- Urgent to act upon
Effective presentation aids in all of these areas. There are dozens of data charts to choose from and you can either thwart all your data-crunching efforts by picking the wrong data visualization (like displaying a time evolution on a pie chart) or give it an additional boost by choosing the right data visualization type.
8) How can you create a data-driven culture?
In order to truly incorporate this data-driven approach to running the business, all individuals in the organization, regardless of the department they work in, need to know how to start asking the right data analytics questions.
They need to understand why it is important to conduct data analysis in the first place.
However, simply wishing and hoping that others will conduct data analysis is a strategy doomed to fail. Frankly, asking them to use data analysis (without showing them the benefits first) is also unlikely to succeed.
Instead, lead by example. Show your internal users that the habit of regular data analysis is a priceless aid for optimizing your business performance. Try to create a beneficial dashboard culture in your company.
Data analysis isn’t a means to discipline your employees and find who is responsible for failures, but to empower them to improve their performance and self-improve.
We just outlined an 8 step process you can use to set up your company for success through the use of the right data analysis questions.
With this information, you can outline questions that will help you to make important business decisions, and then setup your infrastructure (and culture) to address them on a consistent basis through accurate data insights.