The Most Common Failure: Rich Telemetry, Thin Voice
Almost every enterprise DEX program collects device and application telemetry comprehensively. Agents run on every managed endpoint, around the clock, and report boot times, crashes, CPU and memory pressure, network quality, and application performance. That half of the picture is usually complete.
The other half rarely is. Employee sentiment, the voice of the employee about their own technology experience, is collected occasionally, from a small or self-selected group, and often only from people who just closed a ticket. The program ends up with millions of device data points and a few hundred opinions.
That imbalance matters because DEX is, by definition, about perception. Forrester defines digital employee experience as “the sum of all perceptions that employees have about working with the technology.” Telemetry measures conditions. Only sentiment measures the experience itself.
The short version. Telemetry tells you what happened on the device. Sentiment tells you whether it mattered to the person using it. A program that has only the first can make devices healthier. It cannot prove that work got better.
What incomplete sentiment costs you
- The DEX score becomes a technology score. Most platforms combine technical measures with sentiment. With no sentiment configured, the score is just device health with a friendlier name.
- Friction that telemetry cannot see stays invisible. Confusing workflows, slow access requests, poor training, policies that block work, and applications that are technically healthy but hard to use generate no alerts. Employees can describe them in a sentence.
- Fixes cannot be validated. Telemetry can show that a crash rate dropped. Only employees can confirm that the change made their day better, or that the fix created a new annoyance.
- Priorities skew toward what is measurable. Without sentiment, teams fix what is easy to count rather than what employees care about most.
- The business case stays thin. Leaders outside IT respond to what employees say. A program that cannot show how employees rate their experience, by group and over time, struggles to earn a seat in business planning.
- Maturity stalls. On our maturity model, the move from Stage 2 (Gauges On) to Stage 3 (Flying on Instruments) depends on correlating telemetry with experience outcomes. Without broad, regular sentiment, there is nothing to correlate against.
What Comprehensive Sentiment Looks Like
“We run a survey” is not the same as having sentiment data. The difference shows up in five dimensions.
| Dimension | Partial (common) | Comprehensive (the goal) |
|---|---|---|
| Coverage | Ticket submitters, a pilot group, or headquarters staff | The whole population in scope, across personas, locations, regions, device types, and frontline roles |
| Frequency | Annual or one-off | A recurring pulse to everyone, plus event-triggered questions after migrations, fixes, and onboarding |
| Context | An emailed link to a long external form | Short questions delivered in the workflow, close to the moment being rated |
| Correlation | Stand-alone results in a slide deck | Joined to telemetry by persona, device, location, and application so causes can be found |
| Closed loop | Results stay inside IT | Employees see what changed because of their feedback, and response rates hold up over time |
Why Enterprises Fall Short on Sentiment
The reasons are rarely technical. Every major DEX platform can deliver in-context surveys. The obstacles are organizational, political, and legal.
1. IT lacks the mandate to talk to the whole workforce
In most enterprises, HR and internal communications own the channel to employees and guard it closely. A DEX team that wants to reach every employee on a regular cadence needs permission it does not have, and often lacks the executive sponsor to get it. The result is a compromise: a small pilot group, or survey only the people who contacted support.
2. Competition with the HR engagement survey
HR teams worry that a technology pulse will compete with the annual engagement survey, confuse employees, or depress participation. Without an agreed survey calendar, the DEX program loses that argument by default.
3. Fear of survey fatigue
“Our people are over-surveyed” is the most common objection. The evidence points somewhere else. McKinsey reviewed more than 20 academic articles and found the number one driver of survey fatigue was the perception that the organization would not act on the results, not the number or length of surveys. Fatigue is mostly a symptom of the next problem.
4. No capacity to act on what employees say
Collecting feedback creates an obligation to respond. In a Quantum Workplace study of 1,068 employees, about two in three said their organization fails to act effectively on survey results, with lack of time, resources, and prioritization named as the biggest barriers. When no one owns follow-up, response rates fall, leaders conclude employees do not want to be asked, and the sentiment program quietly ends.
5. Reluctance to hear the answer
Sentiment results can read like a report card for IT, for a service provider, or for a leader who sponsored a rollout. Some teams avoid broad collection because they suspect what it will show. That instinct protects the team in the short term and leaves the program blind.
6. Works councils, unions, and privacy review
Combining employee feedback with device data raises legitimate questions about monitoring. In Germany, works councils have co-determination rights over technical systems capable of monitoring employee behavior or performance (Works Constitution Act, section 87(1) no. 6), and similar consultation duties exist elsewhere in Europe. GDPR adds purpose and minimization requirements. When these conversations start late, whole countries are excluded from sentiment collection, and the program develops regional blind spots that are rarely reported.
7. The platform was bought for telemetry only
The engagement or survey module is not licensed, not deployed, or blocked by notification policies. VDI sessions, shared and kiosk devices, Macs, mobile, and frontline staff are often left out of the rollout entirely, so even a well-run campaign cannot reach them.
8. Poorly designed first campaigns
A long, generic survey sent at a busy time gets a low response. Leaders read the low response as proof that employees will not engage, rather than as a design problem, and the program never tries again.
How to Build a Comprehensive Sentiment Program
- Get joint sponsorship from IT and HR. Position technology sentiment as the missing dimension of the employee experience HR already measures, not a competing survey. Agree on one survey calendar.
- Write a survey governance charter. Set frequency caps per employee, blackout periods, an approval workflow for new campaigns, minimum group sizes for reporting, and who can see what. A charter turns a political negotiation into a policy.
- Involve works councils, unions, and privacy teams early. Document the purpose, the data collected, aggregation and anonymity rules, retention, and access. Where required, put it in a works agreement before launch, not after. This is not legal advice; involve your own counsel.
- Mix three kinds of collection. A short recurring pulse to the whole population for the baseline, event-triggered questions after migrations, fixes, and onboarding milestones, and an always-available feedback channel for the issues people never report. Use your DEX platform’s built-in surveys, or a dedicated IT experience management tool such as HappySignals (see sentiment tools).
- Treat coverage as a metric. Track the share of the population that gave feedback in the last 90 days, response rate by segment, and which groups have no data at all. Report gaps the same way you report device coverage.
- Close the loop visibly. Publish what changed because of employee feedback. A simple “you said, we did” message after each cycle is the most reliable way to keep response rates up.
- Correlate, then act. Join sentiment to telemetry by persona, location, and application. The most valuable findings sit where the two disagree: healthy devices with unhappy users, or poor metrics that nobody minds.
Our Governance page covers the operating rhythm that keeps this going, and our Research page explains why most technology issues never reach the service desk in the first place.
Other Failure Patterns That Block Maturity
Sentiment is the most common gap, but rarely the only one. These are the other patterns we see most often, and the stage on our maturity model where each one usually stops progress.
| Failure pattern | What it looks like | Usually blocks | What fixes it |
|---|---|---|---|
| Tool as the strategy | The platform purchase is the plan. No defined outcomes, personas, or success measures. | Stage 1 to 2 | Define two or three business outcomes and a baseline before expanding the rollout |
| Partial fleet coverage | Windows laptops are covered; Macs, VDI, mobile, shared devices, and contractors are not. | Stage 1 to 2 | Report coverage by device type and population, and close the gaps deliberately |
| No dedicated owner | DEX is a side duty for an endpoint engineer. Forrester’s 2025 survey found only 20% of organizations have a dedicated DEX team. | Stage 2 to 3 | Name a DEX lead with time, budget, and authority to set priorities |
| Dashboards without action | Insights are reviewed but never turned into a remediation backlog with owners and dates. | Stage 2 to 3 | Run a weekly operating rhythm that converts findings into work |
| Incomplete sentiment | Telemetry everywhere, feedback from a small or self-selected group. | Stage 2 to 3 | See the sentiment program steps above |
| Automation without guardrails | Remediation scripts scaled without risk tiers, staging, or rollback, until one breaks something and automation is frozen. | Stage 3 to 4 | Apply the controls in Governing Autonomous Remediation |
| No business translation | Results are reported in device metrics that executives do not use. The budget is questioned at renewal. | Stage 3 to 4 | Express outcomes as time returned, cost avoided, and experience targets such as XLAs |
| An IT-only program | HR, Finance, and business units are informed but not involved. | Stage 4 to 5 | Bring them into governance with shared measures |
| Project, not practice | The team disbands after rollout and the platform drifts into a monitoring tool. | Any stage | Run DEX as an ongoing operating discipline; see DEXOps |
Areas That Are Usually Overlooked
- Onboarding and role changes. The first weeks shape how new hires see the organization, yet the journey is rarely measured end to end.
- Frontline and shared-device workers. Retail, clinical, warehouse, and branch staff often sit outside both the telemetry rollout and the survey audience.
- Non-technical friction. Access requests, approval chains, policy restrictions, and training gaps cause much of the frustration employees describe, and none of it appears in device data.
- Regional and language coverage. Surveys offered in one language, or excluded from countries pending consultation, leave large populations unheard.
- Communicating improvements. Employees rarely hear when IT fixes something on their behalf. Telling them builds the trust that keeps feedback flowing.
- The experience of AI tools. Copilot and other AI assistants are often measured by license counts rather than by whether employees find them useful. That is a sentiment question as much as a telemetry one.