menu
Last updated on
Plutora Blog - Business Intelligence, DevOps, Digital Transformation, Release Management, Software Development, Value Stream Management

What Is Business Analytics? A Comprehensive Definition

Reading time 13 minutes

In today’s world, knowledge is power. Therefore, if you have it, you have the power to make informed decisions. Those will change the way you build a product or how you deliver a service. They can even change the way you currently make your business work. Knowledge can help your business to improve efficiency, reduce errors, and increase your planning capabilities. It can also aid you in forecasting scenarios and predictions. That’s why it’s important to understand what business analytics is.

In order to generate knowledge, you need to be able to obtain data from all your business processes. These can be logs, commits, releases, test results, planning, etc. Then you have to extract meaningful information. Afterward, you’ll have the tools to make informed decisions. It’ll become clear what direction to follow. In the end, you’ll repeat the cycle, feeding new data, creating new insights, and adjusting the direction. In the end, the result will be the achievement of every one of your business goals.

Sound interesting? This is what business analytics is for, and in this post, I’ll show you how to leverage its power.

Definition

In the last section, I briefly described what business analytics is for; however, we still need to define what it is. If you search the web looking for a definition (hence you arrived here), you’ll see that there isn’t a clear view of what business analytics is. That is because sometimes it may look similar to the definition of business intelligence. So, let’s make it clear what it is:

Business analytics is the set of tools to process data in order to answer the question “why,” and then creating action items based upon the answer.

Much better, isn’t it? Now let’s explain the concepts business analytics encapsulates a bit more in the following paragraphs.

Tool Set

Initial Tools

Business analytics doesn’t have a predefined set of tools. Tools can be as simple as spreadsheets, in conjunction with some formulas, different types of calculations, and graphics. These will help you visualize the information. You can see these as starting points to begin defining some metrics you may already have. Similarly, you’ll get some conclusions, and in the process, discover other metrics you need.

As the data collected increases, a spreadsheet won’t be enough. You can start grouping related information, give it more structure, remove duplicated data, and infer relationships. Now you need a database manager. This way, it’ll be easier to search for specific data. Then, filter it according to your desired criteria, aggregate it, and summarize it. Finally, a database manager will help you to increase the reports’ complexity.

With structure, you’re able to include more data sources, and you need to adapt them to the structure you have. Your toolset will increase to include data transformation tools. For example, a new logging tool needs to transform its records into one or more rows in your database tables. Once there, you need to analyze them, find patterns, and fix or improve the process related to the tool that generated the logs. Afterward, you add the results again to your database.

Advanced Tools

You can see that the process is incremental. Your tables may transform into dimensions and facts in a data warehouse model. Reporting becomes even more difficult since the variables to analyze increase, and the correlations are harder to build. But the resulting information will add more than enough value to justify the effort.

Add statistical models and analysis, and you’ll be able to predict some outcomes and act upon them. Take a look at the concept of predictive analytics for a wider look at tools to predict the future.

You’ll avoid errors, find bottlenecks, and then find what processes to improve.

There may be a point where not even existing tools can help you to perform analytics. Still, not a big deal—you can build your own. The advantage of this approach is that the tool will fit like a glove.

Data

As you may note from the previous section, the tools you use are extremely coupled to the amount and type of data you have. You usually start small—just one or two data sources with a simple structure. Although it’s a small set, you still have enough to help you obtain valuable information. As your analytics mature, you need to add more data sources. Therefore, with each data source, the amount of data also grows, and the tools to process it will change too.

The Slow Developer Example

An example of how to use data begins with a list of issues tracked on Jira. You can review how much time it took developers to fix them. You can see the date from their detection up until they were deployed. In the same fashion, you’re able to figure out how much time passed in each development phase and if it was solved on time and by whom.

Let’s say that you find a developer who barely delivers on time. You may want to know how many issues they’re assigned to and what the complexity is. If you can’t figure out a plausible cause, consider adding another data source to your analysis.

Take the source code repository as your new data source. How often do developers commit their code into the repository? Is the code working correctly? For our slow developer, how many commits a day do they make? Are all commits for the same issue? Let’s assume that their numbers are more or less the same as the other developers. Then the issue isn’t here. Time for another data source.

Now let’s add Jenkins to the mix. How often are test environments created? Are they created fast enough? How many resources are provisioned with each one? Are there any errors when provisioning the environments? It may turn out that provisioning is failing often, but the team in charge hasn’t found a cause yet. How might this relate to our slow developer? The test environments that are failing are all related to the issues that they were working on. Now the DevOps team can put their effort into reviewing why this developer’s environments are having issues and help them release on time.

This example proves how important data is and how to use it to find and solve issues. I used three data sources, but many more can be added. However, each source saves its data in a different way, and you need to implement a transformation process in order to consolidate and correlate the information. In the end, that information will help you to arrive at a conclusion.

An Answer to “Why”

This is the key element in the business analytics definition. With enough data and the right tools, you’ll be able to discover the answer to the questions “what” and “how.” Many business decisions are based solely on these answers. But, by answering the question “why,” you discover a plethora of options hidden within the data you already have. The knowledge that you gain will have a strong impact on future business decisions. It’ll also empower you to ask more questions, make assumptions, and play with different scenarios. It may even help you see what other data you need to collect that otherwise wouldn’t seem useful.

In the slow developer example, we wanted to know why that developer was always releasing late issues. By asking this question, we increased our one data source (Jira) to three data sources (Jira, source code repository, and Jenkins) and correlated the issues the developer solved with the test environments created with errors. Why were they created with errors? That’s where the DevOps team helped, but we already gave them a hand by pointing out that those environments were related to the issues the slow developer was fixing.

Taking Action

The final step in the definition is to create action items. To me, it’s useless to spend resources on getting the tools and collecting and analyzing data if, in the end, the answer to the “why” question doesn’t result in an actionable item. It may seem obvious, but a case may present where an answer is all that’s needed. Then, for that specific case, evaluate if the information is worthy of the effort to get it.

In the slow developer example, we analyzed the data provided by the business analytics. Then we could identify an undesired behavior (the slow developer not delivering on time). By asking “why,” an issue was found. Finally, the action was sent to the DevOps team to fix the issue associated with the slow developer.

Business Analytics Advantages

Now we’ve defined business analytics and provided a small implementation example. It showcased how, by analyzing data, you can get more business insight. Still, you’ll realize that building the blocks to make useful analytics requires a good amount of effort and, consequently, a large amount of time. However, in this section, I want to highlight some of the rewards you’ll reap by doing business analytics.

Improve Efficiency

Having business analytics will enable you to better visualize all your business processes from start to end. Every step in your processes generates data. When it’s integrated into your analytics engine, data showcases variables to be measured. Those variables will translate into process metrics. After analyzing those metrics, you can generate indicators that will raise alarms when a process is performing lower than expected. In the end, you’ll decide if the process or the indicators need adjustment.

Let’s see an example for a company that creates a software product. In the next image, you can see some insights.

What information can we get from this image?

  • The image shows data averages for five months.
  • The longest phase in the product development process is development, with a duration of 23 days on average. The shortest phase is deployment, lasting five days on average.

But you can also ask the following questions:

  • Why does planning take almost half the time of development? Upon answering, you may learn the seniority of your developers. Perhaps sponsoring a specific course will increase their proficiency for developing product features. Or you may learn that the requirements aren’t well defined in the first stage, leading to duplicated work or rework.
  • Why was November’s feature cost so high compared to other months? Maybe an incident occurred, and your developers needed to work additional time. Next time, you may need to consider the risk of those incidents happening again.

Answering these questions will lead to actions to improve the process. Then you’ll have a baseline reference to analyze a similar period in the future. Finally, new questions will arise, thus adding efficiency to the process.

Reduce Errors

It’s extremely difficult to have a zero-error process. However, you can always aim to reduce the amount to the minimum possibility. Analytics gives you an overview of all the errors presented during a specific process. When asking the appropriate questions, analytics will provide guidance on where the error source may be. Thus, the time to identify errors is shorter. It can also help identify paths that end up generating unexpected errors.

Here is an example dashboard to visualize error information:

PR-SS-Report

In the image, you can see a summarized view making a comparison between errors that occurred in the last week compared to the current week. For the current week, the metrics show that errors have increased by 13% compared to last week. It also shows that developers are fixing 13% fewer errors. Although it seems the errors are getting worse, the chart indicates that there are more test cases than errors. It means tests cover most scenarios, and not all of them resulted in errors.

Increase Planning Capabilities

Analytics helps us to measure the time it takes to execute every step in any business process. With enough information, time patterns emerge, and you can use them to define thresholds. That knowledge will help you predict how much time future tasks may take and have the confidence that those times will hold true.

Let’s take a look at another report example.

Release Reporting

In the previous image, you can see a list of planned software releases, who the owner of the release is, and the status of the tasks involved in each release. Furthermore, in the following image, you can see a release calendar together with a risk profile.

Release Calendar

Notice that the release risk profile is showing a high release risk. With that information in mind, you have the evidence that the release most likely won’t happen on time, and you can alert stakeholders about it. Another solution could be to include fewer tasks per release, mitigating the risk of an overdue release.

Again, proper visualization makes it easier to notice aspects that require additional attention and act upon them. The actions are immediate. Just after implementing them you’ll notice the results. Now you can continue planning and act if any other event jeopardizes the release.

Predict the Future

By using predictive analysis, you have the power to analyze past data. In doing so, you gain insight on what outcomes may occur in the future. It’s not magic. The tools required for predictive analysis include machine learning, data mining, statistics, and artificial intelligence over large sets of data. Combine it with all the other tools, and your data will tell you a very reliable story.

Decisions, Decisions

The whole point of business analytics is to empower leaders to make informed decisions and take action to improve their businesses. There are many options for selecting what tools to use for analytics, so any tool you choose will be a good investment. Acquiring data won’t be a fast process; however, there are tools that ease the effort by connecting to a large number of data sources. Those tools also help suggest a predefined set of metrics to get insights from said data sources. Afterward, you’ll visualize the status for the metrics you consider worth measuring. It’s at this point that you have a clear view of your business processes. You see which aspects you’re doing well and in which points you need to improve.

Additionally, you can use your data to make predictions and forecast scenarios that include risks and ways to mitigate them. Now it’s up to you to have the final say in where you want to take your business—your data has your back. For more information about business analytics platforms, including the images used in the examples above, check out the Plutora website.

Juan Pablo Macias Gonzalez Juan Pablo Macias Gonzalez

Juan is a computer systems engineer with experience in backend, frontend, databases and systems administration.