Self-Service Analytics: is it about to self-destruct?

Let’s begin by explaining precisely what it refers to:

Self-service analytics is a form of business intelligence in which line-of-business professionals are enabled and encouraged to perform queries and generate reports of their own, with nominal IT support.

In theory, less technical business users will be able to carry out self-service analytics tasks (with little help from IT), and savvy business users will enjoy the tremendous flexibility of self-service functionality.

Why self-service analytics are growing in popularity

Self-service analytics tools, such as Power BI, are a rapidly growing requirement for most organisations, mainly due to several factors:

• The technology available these days lets users create rich, insightful and powerful data visualisations. Better insights mean better decision making.
• Who understands the business and data the best? The users themselves do, so why not put them in the driver’s seat.
• IT departments don’t have the capability or the capacity to create relevant and timely dashboards. Even the process of generating these can create bottlenecks. The onus is therefore on the line-of-business user.
• According to Gartner, the number of data and analytics experts in businesses will grow at three times the rate of IT experts. Organisations who want to stay ahead are rethinking their organisational model and skill-sets.
• In the next 2-3 years, Gartner also stated that businesses who offer users access to a governed, curated repository of internal and external data will get twice as much business value from analytics investments as those that don't, i.e. If you’re looking at implementing governed self-service analytics, now is the time to do it. Not only are the results immediate, but ROI is much easier to achieve.

Simple self-service?

While there are many benefits, it’s not as simple as it sounds. It’s not just about installing the right self-service tools, providing some training and allowing users access to the data. Even if you do that, you might still end up with data chaos.

Before we continue, it would be wrong not to mention one of the most significant overarching risks - poor data quality. This often leads to the development of multiple data models and, all with the same or similar content, and all showing differing results. In this blog, we focus on other problems related to self-service (as opposed to data quality – an entirely different issue to be addressed).

What are the risks?

The risks are present in various places when generating self-service analytics.

Here are the risks you should be aware of, and what it could result in if you don’t do something about it.

1. No clarity on the source of truth

You might have numerous people creating supposedly the same report, and all with different outcomes. Who is correct? Or maybe no one is!

2. Unusable data models with flawed business logic and metrics

Users who have access and permission to create their own data models can very easily embed business rules in these models. These business rules (e.g. how to calculate net profit, or calculating the total number of customers) could deviate from business rules for these metrics that are used elsewhere. This has the potential to create distrust about the numbers that stakeholders are seeing and using, with the flow-on effect of people not knowing which dashboards to use and eventually reverting to using Excel again.

3. Business decisions emerging from bad or incorrect data

Following on from the first two points, if the reporting is inaccurate, then the business decisions based on the reports may also be the wrong decisions.

4. Audit failures in case of data verifications

Very often, reports and dashboards used to provide key metrics to business users are subject to audits. If these reports and dashboards do not have the necessary metadata to support their metrics (such as data lineage, the use of data sources that are trusted and the use of approved business rules) then these reports and dashboards will not pass audit, are possibly producing erroneous numbers and will not be able to be used.

5. Compliance failures and regulatory penalties

There are secondary effects of reporting the wrong information, including penalties. If you’re in an industry that requires metric reporting, you need to be 100% sure that the reported information is correct.

6. Analytics & BI system maintenance nightmares

You could end up creating more work on the maintenance side of things. Think of it like when Excel was the tool of choice for reporting. Finding the right spreadsheet to use, understanding the macros and V-Lookups, updating the spreadsheet without causing it to break. If you do not have good governance processes embedded into your self-service capability, no matter which tool you use to visualise your data – you will land up back in Excel world.

7. Huge data privacy and security issues

A lack of control means a lack of control. If you haven’t got all bases covered when it comes to security, you’re leaving open huge security risk gaps.

Steps to better governance

To allow business users to become truly self-sufficient, you need to have governance. This includes assigning roles, responsibilities, ownership, policies and procedures. It will enable you to have a great experience using self-service analytics while mitigating the risks.

1. Manage who is using your data

By nominating Power BI admin roles, and with the use of automation, you can quite easily set up permission-based and role-based access enabling you to understand and manage who is accessing the various data sets and dashboards being created

2. Protect how sensitive data is being used

Use watermarking and versioning (bronze, silver, gold) for sanctioned data sources. This lets you add unique tracking records into your database and monitors how your data is being used.

3. Keep track of changes

Data version control allows you to recall specific versions further down the line if you need to.

4. Share data in a secure way

Depending on your current setup, you may or may not already have safe ways of sharing data. An example of this would be people emailing a Power BI (pbix) file to someone else. This could expose your company to serious data privacy and data security risks

5. Have a clear display of data over time

Data lineage will give you the lifecycle of the data – the origin, where it moves and what happens to it over time. Make sure you are seeing the full picture of your data.

6. Share insights with others

This might be internal or external sharing. Collaboration is one of the most effective ways of working and sharing insights. Using data visualisation and analytics tools that include functionality for seamless collaboration enables organisations to identify, understand and share insights quickly leading to more proactive and efficient decision-making processes

7. Decide who the administrator is and what their obligations are

The administrator can manage users, groups and other admins. They should know what the required setup for the business and control the customisation features.

Once you have your governance in place, self-service analytics provide a vast amount of value.