Yesterday, my 65-year old neighbor sighed, ‘I have to do everything myself these days’. Before he could go to the hospital for a check-up he first had to enter a lot of data on my.hospital.com. He was asked every question imaginable about his identity, place of residence, and physical condition. He continued, ‘The hospital employees used to do that themselves!’
Somehow, I understood him. The downside of self-service for the customer is that now, the customer has to do more. What is the other downside? In other words: who benefits from this and when? And can that be extended step by step to BI-self service? And how does it work? What should an organization do with it?
One can no longer avoid self-service in our society
Organizations thus end up outsourcing a portion of their business processes and do this to the customer or the supplier. You can fill out your own police reports online for certain things, you put together Ikea furniture and at the supermarket you fill your own shopping trolley, and nowadays you even pay the bill yourself.
In many cases, you no longer even need a supermarket employee. Specific technology makes self-service possible. With reports, it is the internet and integration technology, among other things, that enters the reports in the police system behind the scenes. In the supermarket, this includes the bar codes and the self-service cash register systems.
Self-service and intelligent organization
Self-service is an unstoppable trend that seamlessly fits into the concept of the intelligent organization (Van Beek, 2014), as self-service can enable:
- Cheaper services or products or more added value: dependent on the strategy of the organization, they will translate to either a greater number of marginal, cheaper products, or a higher quality of service.
- Faster processing: there are usually fewer links required in the process, by which it can, in theory, be carried out faster.
- More flexible and less dependent: the customer can largely determine for himself where, when, and what is done. Chances are, however, that mistakes are made because the essential knowledge is lacking. Who has never put together an Ikea cabinet and afterwards been left with one puzzling screw remaining? Which means that the cabinet is not stable on its base and has to be put back together again.
- A more active relationship: the customer becomes more closely involved in the implementation of the process so that a more active involvement with the organization can emerge.
Self-service can also go horribly wrong
In practice it shows that self-service can also go horribly wrong. Employees of a bank that have implemented far-reaching self-service cannot even advise the client on what type of account he should open on the internet! An expensive repair man must eventually be hired for the Ikea cabinet. In short, self-service without service. It is therefore conceivable that the customer will then just develop an aversion to the company.
Can the self-service concept be translated into Business Intelligence, and if so, how? Do the same principles apply there? Broadly speaking, this is the case above all with regard to the aspects of speed and flexibility. The following table shows where the greatest advantages of BI-self service are.
|Cheaper or more added value||The IT department can focus on the infrastructure, the supply of ready-made components, and more complex data sets. The simple tasks can now be outsourced to the users.|
|Faster||Users can gain new insights much faster. At the same time, they can experiment, and in doing so, help IT to develop and validate new data sets in a relatively short time.|
|More flexibility||The users are flexible and much less dependent on IT.|
|More active relationship||Depending on other factors, including the role of the BI manager.|
The preferred BI of the future
Many of the BI vendors are trying to promote BI self-service. They are marketing this principle almost excessively and are presenting it as the preferred BI of the future. A part of the market now has the perception that end users (have to) be able to do everything. A complex report or a dashboard can be compiled with a few clicks. But is that perception justified?
Definition of BI self-service
“The concept with which BI end users and analysts can quickly create data sets themselves based on controlled or uncontrolled, internal and external data sources, and visualize the data therein and can share the resulting insights with others.”
The suppliers address a common problem with self-service BI: IT is not always able to quickly satisfy the acute information needs of end-users or analysts. This is especially true when the organization already has access to a data warehouse and then sometimes it takes weeks or months before IT has made the requested IT data sets available.
The business needs the insights now, because the world might look completely different tomorrow. Users sometimes choose to download a BI tool out of desperation or to activate it in the cloud and then take a shot at it themselves. Does this not cause the organization to return back to the pre-BI era?
This form of agile BI causes the BI manager lot of headaches:
- How can self-service users be prevented from producing nonsense? After all, linking tables is usually not simple and very complex for the average user. Let alone the need to link internal and external (big) data sources.
- How can it be prevented that multiple versions of the truth will be created together?
- How can a patchwork of BI tools be prevented from coming into the organization?
A difficult dilemma
The experienced BI manager experiences self-service BI as a difficult dilemma. He will, on the one hand, not want to lose control of BI and also not want the carefully constructed single version of the truth to be exposed to ‘outside attacks’. He fears that users can now not only arrive at new reliable insights with lightning speed, but also at terribly wrong insights.
On the other hand, he will not be able to deny the existence and benefits of BI self-service: to quickly be able to experiment with potentially valuable data sets and react more quickly to developments. He is wise not to throw the baby out with the bath water. It can be thought to develop a BI self-service policy that takes into account the accountability and complexity of data sets and the type of user.
|Controlled data sets||Uncontrolled data sets|
|Simple task||Complex||Simple task||Complex|
|End User||Allow||Do Not Allow||Do Not Allow||Do Not Allow|
|Key User||Allow||Warn||Do Not Allow||Do Not Allow|
|Analyst||Pass to key user||Encourage||Allow Condition.||Allow Condition.|
This policy can force the BI manager arrange a process with the main users, analysts, and IT. And by commissioning the proper authorizations and workflow. And where necessary, to increase the knowledge and skills of main users and analysts. The exact interpretation of the policy can vary by organization, since the level of users and analysts differ, for instance. And there must also be room for experimentation and learning.
The ideal BI environment with self-service?
What does the ideal BI environment with self-service look like and who takes responsibility for what?
- IT ensures a solid, reliable information infrastructure with controlled data sets and ready-to-use, reusable components. And they provide a single BI platform that enables both standard reporting, dashboarding and self-service.
- Analysts can use self-service to experiment with controlled and uncontrolled data sets. They share new insights only when there is a guarantee that the information is correct.
- Main users can, based on ready-to-use components (KPIs and reusable insights), compile dashboards and simple reports on their own. This can be commissioned with a Business Intelligence Framework.
- The BI manager, together with the business managers, ensures that the experimental data sets are validated by domain experts. They will eventually want to include specific data sets in the organization-wide information infrastructure.
- The end user ‘only’ consumes the information.
BI self-service has important advantages such as being able to experiment quickly with potentially valuable data sets and being able to react faster to developments. But it also brings a number of risks to the organization. Even more easily than before, users can produce nonsense or end up reinventing the wheel. And multiple versions of the truth can arise simultaneously. Finally, BI self-service can lead to a patchwork of Business Intelligence software tools being created, which makes the manageable items more complex.
Broadly speaking, what it comes down to is that BI self-service has a lot of added value for the (experienced) analyst and seasoned end user. And it requires the BI manager to manage even more than before and to take more responsibility for the overall Business Intelligence process. If he or she is successful at it, BI self-service may be an excellent addition to the existing range of BI services.