Blog by Anand Chitale, technology evangelist.
But someone choosing to put salmon and ice cream on the same plate is unlikely to enjoy it. Given their experience, will they ever willingly eat either again? And perhaps more tellingly, will they ever visit a self-service buffet again and try another choice?
In other words: one bad choice could have long-lasting effects. Would the same thing follow for self-service analytics? I think it might. A bad experience early on might well prevent business users from coming back for more. But what can be done to help prevent this, and ensure that business users have good experiences of the world of self-service?
Improving the self-service experience through better support
Going back to the analogy of the all-you-can-eat buffet, there may well have been a number of people who could have helped avoid the diner’s problem:
- It is possible that the person in question did not know what they were selecting. Maybe it would have been helpful if the kitchen staff had labelled the dishes more clearly, perhaps in several languages, to avoid confusion.
- Perhaps the person did not realize that they could take two plates. An alert member of the waiting staff might have prevented the problem by suggesting that two plates was a better option.
- The person in question may have been experimenting. Those around them could have applauded the experiment, and then suggested trying again for a better combination when it was unsuccessful.
- The person may genuinely not have known that the combination would not taste good. The waiting staff could have advised on a more suitable combination, based on the person’s likes and dislikes.
The same applies to self-service analytics. The analytics team has a role in making sure that the data available for self-service is prepared well, and is kept up-to-date, just as the kitchen staff are responsible for the quality of the food in the buffet. But their role does not end there. They play a key part in making sure that business users can obtain support to help them make the right decisions about what data to choose, and then what to do with it, especially when business users are still quite new to self-service and are not sure what they are doing.
The designers of the software have a key role to play in making sure that it supports the right decisions. Users need to be helped to ask the right questions, and visual analytics is a very good way to do this.
The emphasis is on the self but not alone
This is not the same as the analytics team coming in and doing the analysis in response to a question from a business user. There is a place for a service like SAS Results. You might compare that to a silver-service restaurant: carefully-chosen menu, a defined set of dishes, but no chance for the user to say ‘Can I have the sauce flavored with x, instead of y?’.
With an all-you-can-eat buffet, the user can select what they want on their plate. They can try different options, have second helpings, throw something away and start again. They can take just one thing to see if they like it, or add a few others for flavoring. If one combination doesn’t work, they can add something else to see if that helps, possibly with the advice of a member of the kitchen staff. That is the key advantage of self-service.
Visual analytics is particularly good for self-service. They say a picture tells a thousand words, which means that a visual representation may be significantly easier to understand and interpret, not to mention guiding the user through the decision-making process. This should make it much easier for business users to understand what they are doing. The latest SAS visual analytics release helps users see the basic data profile more easily.
Self-service requires partnership
Ultimately, self-service analytics is a partnership between data scientists and business users. Both sides need each other, and the quality of the relationship will determine the quality of the analytics.