Compare a heatmap across survey periods
How to overlay two surveys on the Dashboard heatmap so every cell shows the change (delta) in score, how improvements and declines are colored, and the caveats that keep the comparison fair.
On this page
A single heatmap tells you how each team, location or manager scored on each driver right now. But engagement work is about movement: did the survey you ran this quarter actually shift the numbers you set out to change last quarter? The Dashboard heatmap can answer that directly. Turn on the comparison and every cell stops showing a score and starts showing a delta - the exact change between your latest survey and an earlier one you choose.
This guide covers how to pick the two periods, how the change is calculated and displayed, how to read improvements against declines, and the caveats that keep the comparison honest.
On the Dashboard heatmap, click Compare Survey, pick an earlier survey to compare against, and each cell will show the difference between your latest survey's score and that earlier survey's score. Green means the score went up, red means it went down, and only rows that exist in both surveys are shown. Close the comparison to return to the normal single-survey heatmap.
Before you start
A few things to know so the feature behaves the way you expect:
- This lives on the Dashboard heatmap, the driver-by-demographic grid under Analyse -> Reports (the main Dashboard). It is the same heatmap covered in Read a heatmap - the comparison simply changes what each cell displays.
- The comparison may be gated for your account. The Compare Survey button appears only when the comparison capability is enabled for your organization. If you do not see it, it is turned off at the account level.
- You need at least two report-ready surveys. The comparison always uses your most recent survey as one side of the equation, so you need one more finished survey to compare it against. If there is nothing to compare, the button is disabled and shows an Insufficient Data tooltip.
- Audience matters. This is an Account Admin view of the organization-wide heatmap. What a Sub Admin or manager can compare is limited to the surveys and people they are allowed to see.
How the two periods are chosen
This is the single most important thing to understand, because it is not a free "pick any two surveys" comparison.
One side is fixed: CultureMonkey always uses your latest active survey (the most recent report-ready survey you have access to) as the baseline. You do not choose it.
The other side is the one you pick. When you open the comparison, you choose an earlier survey from a dropdown of your other report-ready surveys. The heatmap then shows, for every cell, how the latest survey moved relative to the survey you picked.
So the mental model is always:
> Delta = latest survey score − the survey you selected
If you want to see "how did we change since the annual survey a year ago," you select that annual survey, and the grid shows this survey minus that one. You are not comparing two arbitrary historical points against each other - you are always measuring your newest results against an older reference point you choose.
The comparison is only as useful as the survey you pick to compare against. The two most common choices are the immediately preceding cycle (did our last round of actions move the needle?) and the same period last year (what is the year-over-year trend, controlling for seasonality?). Pick the one that matches the question you are actually asking.
Turning on the comparison
- 1Open the Dashboard heatmap. Go to Analyse -> Reports and scroll to the driver heatmap (engagement by driver). By default it shows the scores for your latest survey.
- 2Click Compare Survey. Look for the Compare Survey button (with the exchange arrows icon) at the top-right of the heatmap card. If it is greyed out with an Insufficient Data tooltip, you do not yet have a second survey to compare against.
- 3Pick the survey to compare with. In the Compare survey scores dialog, use the Select Survey dropdown to choose an earlier survey. Each option is labeled with its title and published date so you can tell cycles apart.
- 4Click Compare. The heatmap reloads in delta mode. A banner appears reading Comparing with: [survey name] so it is always clear which reference point you are looking at.
- 5Read the deltas, then close. When you are done, click the small x next to Comparing with to drop back to the normal single-survey heatmap.
How the change is calculated
For every cell (a driver crossed with a demographic group, for example Communication for the Bangalore location), CultureMonkey takes the group's score in your latest survey and subtracts its score in the survey you selected. The result is rounded to one decimal place and shown in the cell.
A worked example, reading a single cell:
- In the latest survey, the Sales team scored 8.2 on Recognition.
- In the survey you selected to compare against, the Sales team scored 7.5 on Recognition.
- The cell shows +0.7 - Recognition for Sales improved by seven tenths of a point since that earlier survey.
The same subtraction is applied across every row and every driver, including the special rows:
| Row | What the delta means |
|---|---|
| A driver (e.g. Recognition, Communication) | Change in that driver's average score for the group |
| Overall Score | Change in the group's overall engagement score |
| Participation Rate | Change in the percentage of the group who responded |
| e-NPS | Change in the group's eNPS, on the −100 to +100 scale |
Note that Participation Count is deliberately left out of the comparison. A raw count of respondents is not comparable between two surveys with different invite lists, so it is dropped from delta mode rather than shown as a misleading number.
Reading improvements versus declines
In delta mode the color of a cell no longer means "high score" or "low score." It means direction and size of change. Green is improvement, red is decline, and amber is roughly flat:
| Cell color | What it signals |
|---|---|
| Dark green | A significant improvement since the compared survey |
| Light green | A moderate improvement |
| Amber / yellow | Little or no meaningful change (stable) |
| Light red | A moderate decline |
| Dark red | A significant decline since the compared survey |
The bands are set on a normalized scale, not on the raw number, so that changes are comparable across metrics that live on very different scales. A driver score can only move within a 0-10 (or 0-5, or percentage) range, participation is a percentage, and eNPS spans −100 to +100. CultureMonkey scales each delta against the largest change that metric could possibly show before choosing a color. That is why a +5 change on eNPS looks calmer than a +5 change on a 10-point driver score - relative to what is possible, the eNPS shift is smaller.
Practically, this means you should:
- Scan for red first. Dark red cells are where a group went backwards the most since your reference survey. These are your regressions and usually deserve the first look.
- Then confirm your wins in green. If you ran action plans on a specific driver, the cells for that driver are exactly where you check whether the work paid off.
- Treat amber as "watch, do not celebrate." A stable cell is not necessarily good - if the score was low last time and barely moved, amber is telling you the problem is still there.
A group can be dark green (big improvement) and still sit at a low absolute score, or dark red (big decline) while still being your strongest team. Delta mode answers "which way did it move," not "is it good." Flip back to the normal heatmap to see the actual scores whenever you need the full picture.
Which rows appear, and why some are missing
Delta mode only shows a group when it exists in both surveys. CultureMonkey intersects the demographic groups from the latest survey with those from the survey you selected, and compares only the ones common to both.
That has a few consequences worth expecting:
- A team, location or attribute value that did not exist last time will not appear. A team created after the earlier survey has nothing to subtract from, so it is left out of the comparison.
- A group that has since been removed or renamed can drop out. If the label no longer matches across the two surveys, there is no pair to compare.
- A cell can be blank even when the row is present. If a group answered a driver in one survey but not the other, that particular cell has no pair and stays empty, even though the rest of its row is filled.
None of these are errors. They are the comparison being careful not to invent a delta where there is nothing valid to subtract.
Caveats: keeping the comparison fair
The math is simple; the interpretation needs care. A delta is only meaningful when the two things you are subtracting are genuinely comparable.
- Compare like with like. The comparison is most trustworthy when both surveys asked the same drivers and questions. If you compare a full annual engagement survey against a short pulse that only covered two drivers, most cells will be empty and the few that remain can mislead. Choose a reference survey with a similar scope.
- Watch for audience changes. A team that reorganized, doubled in size, or had heavy turnover between the two surveys is not really "the same team" answering twice. A shift in who is in the group can move a score as much as any real change in sentiment.
- Mind the participation delta. If a group's participation swung sharply, its score delta is on shakier ground - a different slice of people responded. Read the Participation Rate row alongside the driver rows before drawing conclusions.
- Small groups are noisy. For a group of six people, one person changing their mind can swing the average by more than a point. Deltas on small groups exaggerate easily. Weight your attention toward larger, more stable groups.
- Respect anonymity thresholds. As everywhere in CultureMonkey, a group too small to meet the response threshold will not surface a score in either survey, so it will not produce a delta. This is expected, not a fault.
Treat a red cell as a question ("why did this slip?"), not a conclusion. Pair the comparison with the open-text feedback and the participation numbers behind it before you decide what actually changed.
Frequently asked questions
Can I compare any two surveys against each other?
Not directly. One side is always your latest report-ready survey. You choose the other survey to compare it against. To measure change between two older surveys, you would need the more recent of them to be your latest survey.
Why is the Compare Survey button greyed out?
Because there is nothing to compare against. The button needs at least one report-ready survey in addition to your latest one. Until a second survey has finished and is report-ready, it stays disabled with an Insufficient Data tooltip.
Why did some teams disappear when I turned on the comparison?
Delta mode only shows groups present in both surveys. A team that did not exist (or had no responses) in the survey you selected has nothing to compare against, so it is left out until it appears in both periods.
Does green always mean the group is doing well?
No. Green means the score improved since the compared survey, regardless of the absolute level. A group can improve a lot and still be low, or decline slightly from an excellent starting point. Switch off the comparison to see the actual scores.
Is Participation Count shown in the comparison?
No. Raw respondent counts are not comparable between surveys with different invite lists, so the Participation Count row is dropped in delta mode. Participation Rate (a percentage) is compared instead.
Can managers and Sub Admins use this?
The comparison respects each user's data scope. A Sub Admin or manager only sees surveys and groups they are authorized to view, so their comparison is limited to that scope. This article is written for the Account Admin, organization-wide view.
Where to go next
- Learn the base heatmap first: Read a heatmap
- Get oriented on the Dashboard: Tour of the dashboard
- Understand the eNPS row: What is eNPS and how is it calculated?
- Why a group has no score: Response threshold
Your feedback helps us improve the Help Center.