Context: 

For those involved in operational monitoring and process efficiency, you may be aware that UX Process Mining has received more press coverage in recent months. Before we dive into the details, it is probably worth explaining what ‘Process Mining’ means in general and how this discipline can help companies large and small to reduce operational costs and improve business outcomes.

Process Mining? 

The theory of mining a process has been around since the late 1990s. It relates to the idea of capturing and then analysing the bread-crumb trail of activity that is recorded on the server-side of an application. This trail could be in the form of database logs, transaction audit files etc and once aggregated, integrated and formatted can help show the passage of transactions within and across system boundaries. Sometimes, this may also include information about the human participants – but quite often this detail is obscured by intermediate systems or middleware. Given that data may be recorded continuously for many years, the operations team would then need to visualise or ‘mine’ this data to make sense of what they are looking at. The classic example of process mining is to confirm the heritage of pharmaceutical products or the production of foods and beverages.

When visualisation is implemented well, it should allow the user to navigate the data across different dimensions (e.g. time, system, person etc) and identify outliers to standard process and monitor KPIs over time. The net result is that operational staff can use this insight to help improve process and support compliance regimes.

Unfortunately, for those who need to improve the operational processes that are performed by the end-users, there is often a disconnect between the action taken by a user and the collection of transactions that may occur as a result. Furthermore, if you are wanting to understand which applications a user is looking at, perhaps as part of an exploratory process, then there is unlikely to be any meaningful trail of activity at the back-end.

UX Process Mining

As the name suggests, this is a variation on traditional process mining – one that focusses on the user’s interaction (or User eXperience) with the client-side applications rather than just the updates that occur at the backend. Imagine if you could know when applications were launched, received focus, lost focus, were moved or resized etc. These events, when analysed properly, can provide critical insight to user activity levels, navigational paths, dwell time and process complexity. Better still, once the actual business outcomes are correlated with the UI metrics then you can begin to understand how to improve operational efficiency and target end-user training where the biggest paybacks are likely to occur.

With the right kind of data capture, there are many queries that you execute to reveal operational insights. A selection of these are shown in the table below:

UX Process Mining

The data that is required to deliver Application, User and Infrastructure queries (the top table) can be generated very simply by looking for UI events that relate to application launch, window focus, mouse movement, keyboard clicks and machine stats. Not all UI integration platforms provide this level of logging – even though this information is obviously available at the operating system level. That said, this is a critical foundational layer of UX mining data that most of the remaining queries use to apply additional post-capture processing and heuristics.

 

For example, the next table, User Journey, illustrates how, when you have built up a body of recent or historical UX data, you can start to look for trends in application usage and identify best-practice and then deviations from there. Those UX mining tools that offer the ability to do this for a subset of applications, users and time periods will help the operational analyst target on specific areas which can have the greatest impact for time spent.

Finally, the bottom table, Business Outcomes, is where most operational and business value can be gleaned. For example, if you have the ability to monitor user journeys AND you know what best practice looks like AND you know which users are delivering the best business outcomes (e.g. highest customer satisfaction, the most sales closed, the quickest handling time etc) then you can baseline these users and compare them with the rest of the user population. This, for the first time, will allow the operations team to prove, with hard evidence, which navigation flows and which applications can help deliver the best, or indeed the worst outcomes.

For those organisations who would like to glean this data from their end-user population, then this possible through tools such as Glue42. While this is primarily an application integration platform it can also be used to ‘instrument’ applications for the purposes of UX Process Mining.

Share This