How is data analysis useful to your business?
Raw data by itself is not that useful. However, with the right methods and analysis, you can use it to tell a story. You can use data to help tell stories about employee performance, learning outcomes or why some of your advertising campaigns are performing better than others.
Your HR team may gather people data to measure employee performance. Consider 360 feedback, self-evaluation and a checklist. The extent to which your data can be used to inform your business strategy is limited by the experience of your data analysts and HR team.
The more experienced your data analysts, the more they will be able to suggest outcomes based on data predictions. You can do this via machine learning, or AI.
How to measure data science
Data science gives you the tools to extract good data from the noise. Your data analysts should be able to clean up this data. Consider having a specialist team of data analysts and data scientists to help inform your decisions. Communicate your objectives to them, and why you need to collect this particular kind of data. This is because to collect the best data possible, you need to make sure that you are clear about why you are gathering the data in the first place. Good data will have the following characteristics and be:
- Reliable;
- Accurate;
- Timely;
- Relevant;
- Complete.
Data scientists can use algorithms, scientific methods, and other systems to source good data. There are different measurement levels that you should be aware of as well when collecting data. The best way of analysing your data depends on what kinds of data you are gathering. These measurement levels break down into two categories – qualitative and quantitative:
Qualitative data
Qualitative data is either nominal or ordinal. Nominal data includes any values that are organised by a label or name, such as different brands, hair colour, or gender. These are therefore different groups that do not have any numerical ordering. You cannot rank or order hair colour according to a numerical value, so it is considered nominal data.
You can organise this data visually in charts or graphs, looking at the frequency of category distribution. A chart can help you to visualise and describe data. It can show things like how many customers buy a particular brand over another brand. Ordinal data, on the other hand, is any data that can be organised by category but has a numerical ordering. This kind of data is positioned according to a ranking system. If you are surveying customer income levels, this data is ordinal, as it can be organised according to how high or low the numerical value is. You can organise this data according to quartiles and percentiles, and the mode and median values presented in the data.
Quantitative data
Quantitative data separates into intervals and ratios. Both types are represented by numerical values . However, ratios have a true zero while intervals do not. A ratio therefore includes measurements like length or time. This is because these kinds of measurement have a true zero. You can have a measurement with no length, such as zero feet. Here, the zero value reflects that there is no length to measure here. Therefore, the zero value here is meaningful. The difference between ratio data and interval data can be seen in temperature, where there is not usually a true zero. 0 degrees Fahrenheit and Celsius do not reflect absolute zero temperature, which is -273.15 Celsius. Here intervals are more meaningful. We can say that 20 degrees Celsius is higher than that of 15 degrees Celsius. Therefore, the interval here gives us more meaningful information rather than the point of zero. In some cases, you might find it useful to bring together qualitative and quantitative methods. This is because it can give you a better understanding of the phenomenon you are looking at.
When should I be thinking about using data triangulation?
Data triangulation can be time and cost-intensive. Therefore, it is useful to consider if you have a good reason for putting data together in this way. Firstly, you should be thinking about what sorts of data you have available, and how you plan to use it. You need to think about what data points are available to be combined. Perhaps you are looking to tell a story about a product you are selling by collecting data from different customer or user surveys. Data triangulation can be a useful tool when looking at learning outcomes and employee development. You can compare data-points from test results with employee learning feedback surveys. This feedback can help inform new learning outcomes, or new strategies for employee development.
What is data triangulation
Data triangulation gives you multiple angles on one aspect of a phenomenon. It is most often used in social sciences. It is therefore useful in performance management , if you are collecting data on employee performance and need to predict future outcomes. There are different reasons why you might consider data triangulation to be a useful tool in your data analysis strategy. You can use it to corroborate evidence or provide multiple contexts to a question you have regarding your data. Data triangulation can involve several research methods to study a particular phenomenon. Therefore, you can use data triangulation to combine:
- several observers;
- different theories;
- empirical/scientific methods.
With these in mind, you should then consider what sorts of data you need to analyse. You can use data triangulation to analyse the following data-points or phenomena:
- Your different target groups/demographics;
- Different points in time;
- Study or learning behaviours of employees;
- Multiple surveys;
- Levels.
You can use data triangulation for cross-verification, which involves using multiple sources to help validate your findings. You may have two different theories with some crossover, which you can combine to help clarify a phenomenon. Two different data sources with different aspects can come together to reveal similar things about a phenomenon. You can use this to help affirm the validity of both of these data sources. Consider combining focus groups with surveys, or a needs assessment. You can also combine different points in time to see employee learning development over a specific period, which can help your HR team provide more effective training strategies.
Conversely, if your data sources contradict each other on the same phenomenon, this can give you a new set of questions about that phenomenon, such as where to go next. Another way of looking at this is that it gives you different viewpoints on the same phenomenon. Therefore, data triangulation can help increase the credibility of your results. So you will find that it can be a useful tool in tidying up data. It can be used in both qualitative and quantitative studies, and so therefore can bring versatility to your analysis of data.
Disadvantages of data-point assessments and data triangulation
Although data-point assessment and data triangulation can be useful tools in your arsenal, you may find that they are not always helpful. It is a good idea to avoid treating these tools as if they have to be used. These tools are simply one option to choose from when analysing and working with your data. Disadvantages of data-point assessments and data triangulation include:
- Data triangulation can be time consuming and expensive, which can have particularly negative consequence if you do not have a good reason for using this approach;
- It can be difficult to bring together qualitative and quantitative findings using this approach;
- It may not be useful if you have limited data to work with;
- It may confuse findings if you do not have a good reason for combining different methodologies;
- If you do not have an idea of what outcomes you are looking for in your data triangulation, you may also confuse findings.
Further reading:
