Dr. Serena Huang does not romanticize data. As Global Head of People Analytics and HR Technology at PayPal, she has spent 17 years pulling engagement numbers, retention curves, and manager effectiveness signals into something leaders will act on. Her rule is simple. If an analytics program cannot point to decisions that changed because of the data, it is decoration.
Our conversation with Serena centered on the gap every HR leader knows but rarely talks about. Companies ask employees what they think, then fail to show what changed. She argues this is not a measurement problem. It is a workflow and trust problem, and solving it is the real frontier of people analytics.
The post that follows translates her operating philosophy into the structures People teams can implement without hiring a 20-person analytics org.
Why Most Employee Feedback Programs Stall Before Action
Serena’s first observation was that most feedback programs measure collection, not change. Companies report survey response rates and engagement scores. Very few track whether employees actually see something shift as a result of what they said.
The data backs her up. Gartner’s HR engagement research found that only a third of employees believe their organization will act on feedback, and nearly half wish the company did more in response. The gap is not in asking. It is in closing.
Her fix is structural. Build feedback collection and feedback response as two halves of the same workflow. That is exactly what the AllVoices employee engagement solutions approach treats as non-negotiable: every input has an owner, a review cadence, and a visible response path.
How to Translate Survey Data Into Decisions People Can See
Serena walked through how her team turns quarterly pulse data into action. Three moves define the rhythm.
First, cut the data by manager so accountability lives at the team level, not just the company level. Second, give every manager a short list of two or three behaviors to change and a check-in date. Third, re-measure the same segment and publish the delta, even when it moves the wrong way.
She emphasized that transparency about misses is what earns permission to keep asking. If you only show the wins, employees learn their feedback becomes a marketing asset.
What data do you actually need before the first action
You need engagement or sentiment scores, a minimum-cell-size rule, a view of manager span of control, and turnover data. That is enough to start. Most analytics projects stall because teams chase a complete dataset before anyone acts on a single insight. Serena’s rule is to ship a first decision inside two weeks of data arriving, even if the decision is small.
How do you know whether feedback actually changed something
Run the same targeted question again 30 to 60 days later in the same segment. Compare the score and the qualitative themes. When employee feedback processes are paired with a same-question re-measurement loop, employees see whether the organization heard them. That is the proof point that drives continued participation.
What Actually Works for People Analytics Teams
Principle 1: Serve managers, not dashboards
The consumer of people analytics is the front-line manager. If a manager cannot read the output in 90 seconds and take one action, the report failed. Serena advocates for short manager packets that pair one number, one theme, and one next step. Deloitte’s people analytics research makes the same point. The analytics function exists to translate information into solutions and prompt decisions.
Principle 2: Blend survey data with intake and investigation data
Pulse surveys are necessary but insufficient. They tell you how people feel. ER intake tells you what they are experiencing. When you pair them, you see which team cultures are silently eroding and which are noisy but healthy. This is where tools like the AllVoices data and insights module earn their place in the stack, because one view connects sentiment with incidents and resolution quality.
Principle 3: Close the loop in public
Serena pushes for visible response cycles. Publish what changed, what did not, and why. Employees forgive slow progress faster than invisible progress. The worst outcome is a survey every quarter with no trace of response in between.
Where Employee Relations Fits in a Feedback-Driven Culture
A survey program without a functioning ER backbone will produce disappointment at scale. Employees raise themes, managers cannot resolve structural issues, and the loop breaks. Serena was explicit that analytics teams should be deeply partnered with ER because the casework itself is the most honest signal of workplace health.
When ER cases are tracked in a structured system, themes emerge that no survey will surface. The AllVoices HR case management system gives People teams the ability to connect a spike in discrimination intake to an engagement score drop in the same business unit. That connection transforms engagement data from a lagging indicator into a diagnostic tool.
How should people analytics and ER teams partner
Share a data model. Run joint reviews monthly. Align on thresholds that trigger action. When an ER theme crosses a volume line for a business unit, the analytics team should trigger a targeted survey in that unit rather than wait for the next quarterly cycle. That compression of the cycle is what turns analytics into a responsive system.
Frequently Asked Questions About Acting on Employee Feedback
Why do employees stop trusting feedback programs
Because they do not see the response. Collection without visible action becomes extraction. Once employees conclude their input will not alter anything, response rates drop, qualitative content flattens, and the remaining data loses diagnostic power. Rebuilding takes 12 to 18 months of consistent visible action, and it is more expensive than never breaking trust in the first place.
What is the minimum viable people analytics stack
A pulse survey tool, an HR case management platform, a turnover data source, and a basic analytics layer to join them. Teams often overbuild by pursuing a mature data warehouse before any decisions have been made. Start small, act fast, and add complexity once the action loop is stable.
How does Vera and AI fit into acting on feedback
AI speeds up the analysis layer by clustering themes, generating manager summaries, and drafting response messages. The Vera AI co-pilot for employee relations helps ER teams move from raw intake to thematic summary in minutes, which shrinks the window between feedback and decision. It does not replace judgment. It removes friction.
How do you prevent survey fatigue
Ask fewer questions more often, and respond visibly between cycles. A 10-question survey that produces visible action beats a 60-question survey that produces a deck. Serena’s teams use very short pulse instruments targeted at specific segments to avoid the bloat that makes respondents cynical.
Can people analytics reduce turnover
Yes, when paired with manager action. Analytics alone does not hold people. Managers hold people. A well-run program identifies the manager behaviors most correlated with staying, coaches managers on those behaviors, and measures both the behavior and the retention outcome quarter over quarter.
The Bottom Line for HR Leaders
Serena’s framing was sharp. The question is not whether you have the right survey vendor. The question is whether any employee in your company can name something that changed because they spoke up. If the answer is no, your measurement machine is producing data that erodes trust rather than builds it.
Start by pairing every input source with a named response path. Join sentiment, ER, and retention data so themes are visible. Publish the wins and the misses at the same cadence. And build the analytics function to serve managers rather than executive slideware.
To see how AllVoices threads feedback, ER intake, and analytics into a single workflow that closes the loop, book a personalized demo with our team.
.avif)
.avif)
.avif)
.avif)
.avif)
.avif)