What it is and what it's for#
The pulse is a continuous measure of how the workforce feels. Instead of a long survey every six months, each person gets one random question a day, answers it in a moment and gets on with their work.
The difference from a survey is one of nature, not of size:
- 1
A survey is a snapshot
It is launched, it is closed and it gives a result with a date on it. It is useful for comparing waves.
- 2
The pulse is a series
It is always open. It is no use for a single day's figure: it is useful for seeing where the organisation is moving.
Results are read under Analyze › Pulse Check Dashboard, on the Organisation pulse screen, which aggregates every response and cuts it by question, by category, by location and by area. Every block, with screenshots, is in Pulse Check Dashboard.
Pulse and surveys do not compete. The pulse tells you that something is moving and on what topic; the survey lets you ask in depth once you know where to look.
How it's set up#
Only one thing is configured: the catalogue of questions. Distribution, calculation and the cuts are automatic.
The questions#
Questions are kept in Manage › Surveys › Sentiment Analysis Questions. Each one has a Topic (it groups questions among administrators; people are not shown it), the wording in Question (ES) and Question (EN), the Projects it is asked in (at least one), the dates From and Until, and optionally Only in these locations and Only in these groups. It is switched on with Question on: a new question starts switched off and stays out of the draw until you switch it on. In the list, each question shows who it reaches and how it is going (answers, average and comments), with a button to read its answers. A question that already has answers cannot be deleted, only switched off.
The fields and steps are in Sentiment Analysis Questions.
Always word questions positively and about a single idea. A question that bundles two things together ("do you have the means and the support you need?") produces a number that cannot be interpreted: you don't know which of the two failed.
The Topic is what makes the pulse useful. A question filed under a topic it doesn't belong to breaks nothing visibly: it simply contaminates the series of the wrong topic and nobody notices.
How it's distributed#
Each person gets one random question a day, picked from the switched-on questions that apply to them by project, dates, location and group; if none fits that day, they are asked nothing. There are no reminders and no chasing: if they don't log in that day, they don't answer and nothing happens.
This has a direct consequence for reading the data: because distribution is random, the number of responses to any one question is small and variable. The screen says so explicitly, and that is why it always shows each question's response count next to its result.
How to read the screen#
At the top you choose the period: 30 days, 90 days or 12 months (it opens on 90 days). A line underneath sums up how many answers there are, from how many people and out of how many active staff. Below there are four figures, each compared with the previous period: the Pulse index, which is the average of the responses (1-5 scale) taken to 0-100 so it can be compared between periods; Participation (people responding over active headcount); the Unfavourable share (answers of 1 and 2); and People responding. How each one is calculated is in Metrics.
The index moves very little: almost every organisation lives within a narrow band. The variable that really moves is participation, and a drop in participation usually comes before a drop in the index.
The rest of the screen expands on the figure:
What to look at now
The readings the system infers from the period's data, for example which category is lowest. It is the entry point if you are short of time.
Questions
One row per question with its result. Clicking one expands its evolution: at the top, Answers (with the previous period's), People, Previous index and Unfavourable; below, a table with Month, N, Index and vs period. When there is no previous period to compare against, that is stated instead of inventing a change.
Monthly trend
Index and participation month by month. It is the chart to look at before any other.
Answer distribution
How the responses spread across the scale. A decent average with many responses at the extremes is not the same as a decent average with everyone in the middle.
Index by category
The actionable cut. Here there are enough responses per topic, and it is where you decide what to act on.
By location and By area
The same results split by site and by organisational unit. Only groups above the minimum sample appear: those with fewer than 5 respondents are hidden, and the footer says how many.
Employee voice
The recent comments, with the most critical first. It is what gives the numbers context.
What the employee sees#
- One question, the first time they log in that day. Never more than one.
- They answer it on a short scale and, if they want, add a comment.
- There is no obligation: they can ignore it and carry on using the app as normal.
- They don't see anyone else's result or their own history: the pulse is not a scoreboard.
- They are not identified in the results. Cuts by site or by area only appear when there are enough responses for nobody to be recognisable.
The pulse only works if the workforce perceives it as genuinely anonymous. If management comments on a small team's result in a way that makes it possible to guess who said what, participation collapses and does not recover.
Frequently asked questions#
Why does a question have so few responses?
Because distribution is random and daily: each question only reaches part of the workforce each day. It is normal, and that is why the response count is always shown next to the result.
Can I find out who answered what?
No. The pulse is anonymous by design and small cuts are hidden so that it cannot be worked out.
Why doesn't a site appear in the cut by location?
Because it does not reach the minimum sample: fewer than 5 respondents in the period. It will appear once it accumulates enough responses.
How often should the pulse be reviewed?
Once a month, looking at the series. Reviewing it daily leads to reacting to noise.
What do I do if a category's index falls?
First, check whether participation has fallen too: if it has, the problem is with the programme before it is with the climate. If participation holds and the category falls, then there is a signal, and that is the moment to launch a survey that goes deeper into that topic.
Limits and warnings#
- A single day's figure is worth nothing. The pulse is read as a series. Comparing two isolated weeks produces false conclusions almost every time.
- The result of an individual question is noise. The unit with a large enough sample to decide on is the category, not the question.
- A stable index with participation falling is bad news, even if the headline number does not move: it means the people who are worst off have stopped answering.
- Groups below the minimum sample (5 respondents) are not shown. It is an anonymity protection, not a fault. In small sites you may never see their cut.
- Changing a question's wording breaks its series. If you need to measure something else, switch the question off and create a new one rather than rewriting the existing one.
- A question that is switched off, has no project or has the wrong dates stops contributing without warning. Nobody gets an error: that part of the climate simply stops being measured.