Data Sources That Can Measure and Improve IT Service Satisfaction During a Crisis

Data Sources That Can Measure and Improve IT Service Satisfaction During a Crisis
  • Homepage
  • Blog
  • Data Sources That Can Measure and Improve IT Service Satisfaction During a Crisis

Delivering Proactive IT Services: Part II

When business relationships are being strained by a global crisis that mandates social distancing, enterprise platforms are put to their most important test. Measuring the true experience of your customers, both internal and external, in this moment is more important than ever.

Unfortunately, the most common source of sentiment data — self-reported surveys — are rarely accurate, comprehensive, or revealing enough on their own. Data is needed to supplement surveys and contextualize results. What are the best sources of such sentiment data? They include:

  • Scraping corporate social intranet chatter using NLP
  • Viewing activities logged on ITSM platforms
  • Customer experience measured in APM systems

Obtaining data from these sources aggregates nearly every touchpoint or pain point a given end user has experienced. This data can be modeled to give a quick snapshot of a user’s pain points — not just what they elect to say on paper. Putting this data into the hands of service desk agents allows for better handling of tickets. They can understand the relationship the user has with IT, and what pain that user might currently be experiencing. 

Data can also be sorted into specific views by group, department, which applications they use, and other variables that ultimately reveal a global picture of organizational ITSM health. In the long run, modeling IT service customer sentiment data allows for complex analyses, including trends, predictive projections, and actionable visualizations. 

Taking a data-based approach to measuring IT service sentiment improves the connections between IT and its customers. With their finger on the pulse of customer needs, IT services can remove common sources of pain. More importantly — especially in a moment when people rely on enterprise platforms to work remotely — having true visibility of customer sentiment allows IT to proactively get to the root of processes or common incidents that drive dissatisfaction. The end goal is to make IT services smarter, whereupon they are no longer just reacting to issues but proactively making strides to improve the IT experience.

Sourcing IT Service Sentiment Data from Outside the Survey Bubble

Customer sentiment surveys, by themselves, may be a woefully inadequate measuring stick. Many surveys have a low response rate, and respondents can have major self-reporting biases. Depending on how they feel when they fill out a survey, they may thoughtlessly enter in all 0/5 or 5/5 scores.

Instead of relying on just CSAT surveys, IT leaders must uncover what employees are saying when they’re not actively in front of the lens of IT workers. To accomplish this, IT organizations can look at data coming from corporate social networks. These include tools like those provided by Slack, Chatter, Chatbox, Yammer (Microsoft’s enterprise social intranet), and others. 

Scraping and analyzing this data using AI/ML-driven NLP techniques can allow IT service organizations to perform a sentiment analysis. These analyses can elicit valuable information such as:

  • How customers experience the IT service request process
  • What service areas tend to generate the most pain
  • Which applications/platforms are the source of the most pain
  • What problems user groups tend to seek advice for before reaching out to the service desk, and what solutions are commonly offered

Data scraped from intranet chatter can allow IT to rationalize the survey responses they have been seeing. IT leaders may, thereby, not feel a need to replace surveys outright, but they can at least give them more context.

Using ITSM System of Record Data for a More Detailed Picture of User Sentiment

Aggregating IT service agent and user activity from across major systems of record allows IT leaders to directly measure customer experience based on actions, not just words. Data imported from ITSM tickets can be analyzed to immediately highlight common problem areas and top request items. IT leaders can then aggregate this data to create a combined view of customer sentiment, which in turn enables deeper analyses through slices, drill-down, and AI-driven predictive analytics. 

Looking at specific user views, for example, can generate an impression of a specific IT customer’s sentiment from the perspective of past ticket data. Service desk agents can ask themselves, “What was this customer’s experience in the last three-to-12 months?” The answer can be gleaned from activities and IT touchpoints that directly affected the specific user. 

Viewing this data can allow service desk agents to answer specific questions about the user, like:

  • Do they have lots of recent tickets? 
  • Are many of their tickets still open? 
  • Have they been interacting with the service desk at all? 
  • What assets do they have ownership over, and which do they use on a regular basis?
  • What are the recent relevant communications they’ve had, that were not necessarily with IT?

Considering that last question, a customer sentiment modeling analytic can allow IT agents to see whether they’ve brought up an issue recently in social chatter. They may have even solicited solutions from peers, so IT agents can know in advance what solutions have already been tried and what obvious questions the user has already had answered for them. A view such as this can dramatically improve interactions with IT service delivery touchpoints, and at the same time, it can serve to accelerate the time to resolution for any issue called in by the user.

Putting Data to Work to Identify Sources of User Dissatisfaction

Individual user sentiment scores can serve as a scaffold for IT service leaders to build better relationships on a person-by-person basis. Creating an expressive KPI affords an instant view of customer sentiment — are they lighting up as a red customer or a green one?

Collectively, this sentiment data holds even more power. IT leaders can use it to trend data. Are certain areas driving increasing levels of dissatisfaction? Are low-level IT problems going to escalate into an incident that will impact business productivity? What types of employee groups tend to have the best experience for the given timeframe? And which ones have the worst?

Using predictive analytics can even allow IT leaders to consider how certain scenarios would play out. A common barrier to proactive action is that change failures tend to cause business disruption. With this perspective, IT can understand which changes can have a big impact.

User views can be aggregated to different roles, including specific groups, departments, etc. This allows organizations to solve challenges like high employee churn by analyzing the experience of individuals with a specific tenure, such as their first six months within the enterprise. IT leaders can determine whether newly on boarded employees will have a good impression of enterprise technology use and IT service, helping them blunt factors that could otherwise lead to higher new employee turnover.

Views can also be shifted to rapidly identify direct sources of negative sentiment. For example, how do users feel about the use of the enterprise’s chosen cloud productivity suite? How do they feel about their ability to access their desktop files and emails remotely? Answers to questions like these can allow IT to proactively identify and address common sources of pain.

How to Implement a System of Intelligence to Measure — And Proactively Improve — IT Customer Sentiment

To implement sweeping, proactive changes that improve end user sentiment towards IT services and enterprise technology as a whole, IT leaders need access to a wide variety of high-quality data from ITSM platforms, corporate social networks, and APM systems.

IT leaders should blend data to provide cross-function insights, such as past incident logs, software assets used/incident tickets, overall user sentiment, etc. Machine learning and AI can implement dynamic metric calculations to determine a final user sentiment score as a bellwether or snapshot. Providing these scores to IT service agents allows them to, at-a-glance, know what pain points drove a user to seek help, and it can allow them to develop relationships that avoid pitfalls common to an impersonal back-and-forth interaction typical of most IT service delivery.

The ultimate goal in such a program is empowering agents to be proactive in providing customer services. By using a service-centered approach, IT agents and group leaders can identify opportunities to improve processes while simplifying their support approach. They can get clear answers to key questions like, “Would the cost of buying brand new laptops be offset by IT support cost savings?”

Proactive approaches may be needed as business processes are changed forever by COVID-19 disruptions — we may never be able to look back, even when “normal” operations resume. Visualizing IT service customer sentiment data in aggregate is key to driving true organization-wide changes. We’ll cover how to visualize and make the most use of your modelled data in the final part of this series: Visualizing IT Service Data to Proactively Improve Customer Satisfaction.

Want more information on how analytics can empower a proactive approach to IT service delivery? Watch our recent webinar: “How IT Can Proactively Anticipate and Deliver Services That Drive Employee Productivity

Watch the Webinar

Related blog posts