People science methodology built for scalable, impactful culture change

Foster retention and productivity with AI-first, industry and academia-backed methodology.

4.7 / 5 on G2, Spring 2025

Trusted by 500+ organizations across 30+ countries
Cerence AIBayportAujan Coca-ColaAlumilRobertshawAstra Service PartnersGarrett
Drivers

Fifteen drivers, named and defined

  • Fifteen research-backed drivers ship with every CultureMonkey account. From Autonomy to Social Connection, one standard set, so scores stay comparable cycle to cycle and line up with industry benchmarks.
  • A driver is a theme that a group of rating questions measures together. The Help Center's own definition, with the drivers drawn from workplace-engagement research and refined across thousands of surveys.
  • A focused engagement survey uses eight to ten of the fifteen. Fewer well-chosen drivers with a few questions each beat a sprawl, and every added question costs response rate.
Every driver, defined

Autonomy

The freedom employees have to do their work without micro-management (a focus on what gets done over how).

Show your working

The Help Center's own example: three drivers, 500 rating answers.

Recognition100 answers3.20
Growth100 answers3.00
Wellbeing300 answers4.40
Overallall 500 answers, averaged once3.88
Open text · separate reporteNPS · separate report
Stored 1 to 5Displayed7.8 / 10
Score

How the number is calculated

  • The overall score is a flat average of every rating answer in scope. Deliberately not the average of the per-driver averages, so a driver with more answers pulls the headline number harder.
  • The difference is not academic. The Help Center's worked example: averaging the driver averages gives 3.53 out of 5, while averaging every answer gives 1,940 divided by 500, or 3.88.
  • Stored on 1 to 5, shown on 0 to 10 by default. The same mean can also display out of 5 or as a percentage. Open text and eNPS never feed it.
The full worked example

“I feel recognized for the work I do.”

Scale

One scale, eight ways to label it

  • Every rating question uses a five-point scale. Point 1 is most negative and point 5 most positive: 4 and 5 count favorable, 3 neutral, 1 and 2 unfavorable.
  • Eight preset label sets ship with the scale, plus a smiley display. Agreement, Satisfaction, Quality, Frequency, Performance, Importance, Focus and Happiness. Custom labels and reverse-coded questions are not supported.
  • Questions come from a research-backed library already mapped to drivers. Templates run from five or six questions to around forty, and every preview shows an estimated completion time.
Rating questions and scales

Your score, clearly explained on a live report.

eNPS

One question, its own arithmetic

  • eNPS is a single 0-to-10 question with its own formula. Promoters score 9 or 10, detractors 0 to 6, and eNPS is the percentage of promoters minus the percentage of detractors, from -100 to +100.
  • The Help Center's worked example: 60% promoters, 15% detractors, eNPS +45. Passives, who score 7 or 8, dilute the result without counting against it, and the score reports as a whole number.
  • It never feeds the engagement score. eNPS lives in its own report with its own benchmark, so the two numbers can move independently of each other, and often do.
eNPS in depth

Illustrative loyalty barometer: eNPS rises from plus 18 in Q1 to plus 47 in Q4 against an industry median of plus 21. Each quarter shows the promoter, passive, and detractor split.

+18Positive
42%Promoters
34%Passives
24%Detractors
−3vs industry
Driver set · survey settings

A survey uses one set or the other, never a mix.

2 modes
SummaryReportQuestionsBenchmark
Standard drivers compare against your industry and employee-size band on the Benchmark tab.
Custom

What you can change, and what it costs you

  • You can replace the default drivers with a set of your own. A survey uses either the default set or the custom set, never a mix, and custom drivers are managed on their own settings page.
  • Custom drivers carry no industry benchmark. A driver only you measure has no external reference point, so it will not appear on the Benchmark tab. We state that here rather than on a call.
  • Sub-drivers are a separate, finer taxonomy for follow-up questions. Their scores roll up on their own, apart from driver scores, and most accounts settle at somewhere between five and fifteen.
Custom drivers and sub-drivers
Proof

Checkable beats impressive

15

drivers in the research-backed framework your engagement surveys score against

10M+

employee data points behind the benchmarks and sentiment models

4.7/5

average rating from HR leaders on G2

G2 High Performer, EnterpriseG2 Momentum LeaderG2 Fastest ImplementationISO 27001 certifiedSOC 2 Type IIGDPR compliant

Go deeper on the People Science hub, how the AI analysis works, how benchmarks are built, and the technology that runs the platform.

In their words

Hear it from people leaders

Mollie Mock Kaufman, Director of Talent Development & Training at United Bank
CultureMonkey made it easy to automate our employee engagement surveys while still giving us the flexibility to customize them. Their real-time data analysis, combined with insights from the People Science team, helped us uncover key focus areas, align our strategy with industry benchmarks, and strengthen our leadership development efforts.
Mollie Mock KaufmanDirector of Talent Development & Training · United Bank
View case study
Heather Kane, Change Management & Engagement Lead at Robertshaw
QR codes, multilingual support, and open comment analysis helped our managers understand culture deeply and have more meaningful conversations, something that was not possible before.
Heather KaneChange Management & Engagement Lead · Robertshaw
View case study
Elizabeth Egan, Director of Talent Management & Organizational Development at Cerence AI
CultureMonkey is able to translate that accordingly, not just a translation you might get on Google, but one that is actually utilized in each language, understood as the question is intended to be understood. And that is done by the employee choosing the language they would like to take that survey in, which is a really big difference maker.
Elizabeth EganDirector of Talent Management & Organizational Development · Cerence AI
View case study
FAQ

The methodology, question by question

The drivers

The fifteen themes every engagement survey scores against.

What are the fifteen employee engagement drivers?
The fifteen default drivers are Autonomy, Work Environment, Leadership, Management, Work Life Balance, Involvement, Communication, Rewards, Recognition, Growth & Development, Purpose Alignment, Innovation, Wellness, Meaningful Work and Social Connection. Every account ships with the same set, so scores stay comparable and line up with industry benchmarks, and each driver is measured by rating questions mapped to it.
Where do the drivers come from?
They are drawn from workplace-engagement research and refined across thousands of surveys, which is the Help Center's own description. Each question was written and validated to measure a specific dimension of the employee experience and is permanently attached to one driver. We do not publish a validation study, and we do not claim one here.
How many drivers should an engagement survey cover?
Eight to ten is the common range for a focused engagement survey; a targeted pulse might use just two or three. You do not have to use all fifteen. Fewer drivers with a few questions each keep the survey short, and shorter surveys protect response rates, which matter more to a usable score than extra breadth.

The calculation

The exact arithmetic behind the headline number, as documented.

How is the overall engagement score calculated?
It is a flat average of every rating answer in scope: add up every rating employees gave, then divide by the number of answers. The result is stored on 1 to 5 and shown on your reporting scale, 0 to 10 by default. Open-text comments, eNPS and lifecycle sends are excluded. The worked example is published in the Help Center.
Is it an average of the driver scores?
No, and the difference matters. Averaging the per-driver averages would let a driver with 50 answers sway the headline as much as one with 2,000. CultureMonkey averages the raw answers instead, so in the Help Center's worked example the score is 1,940 divided by 500, or 3.88, not the 3.53 that averaging the averages would give.
Why is the score shown out of 10 if it is stored out of 5?
Every rating answer is stored on 1 to 5, and the reporting scale defaults to 0 to 10, with 0 to 5 and percentage displays available. A stored 3.88 shows as 7.8 out of 10. The same mean is rescaled, so driver, question and overall scores stay comparable. Reading the results matters more: start with the trend.

The scale

The five points every rating question uses, and the labels on them.

What rating scale do the questions use?
A five-point scale: point 1 is the most negative answer, point 5 the most positive. Answers of 4 and 5 count favorable, 3 neutral, 1 and 2 unfavorable, so a question can read 68% favorable, 14% neutral, 18% unfavorable alongside its average. Eight preset label sets and a smiley display cover the wording.
Can we change the scale or the labels?
You can choose among the eight preset label sets, Agreement through Happiness, and the smiley display, but custom label text is not supported, and neither are reverse-coded questions, where agreeing would be the bad answer. Every scale runs negative to positive and the scoring assumes it. We would rather say that plainly than imply flexibility that is not there.

eNPS

One question with its own formula, kept apart from the engagement score.

How is eNPS calculated?
eNPS asks one 0-to-10 question. Scores of 9 and 10 are promoters, 7 and 8 passives, 0 to 6 detractors; the score is the percentage of promoters minus the percentage of detractors, from -100 to +100, reported as a whole number. With 60% promoters and 15% detractors, eNPS is +45.
Does eNPS affect the engagement score?
No. eNPS lives in its own report with its own benchmark and never feeds the engagement score, which is built only from rating answers. The two numbers can move independently, and often do: loyalty can lag a genuine improvement, or hold up for a while as experience slips.

Customizing

Custom drivers and sub-drivers, and the benchmark trade-off they carry.

Can we add our own drivers?
Yes. Custom drivers are managed on their own settings page and score exactly like the defaults: a plain average of the answers mapped to them. A survey uses either the default set or the custom set, never a mix. Sub-drivers are separate, a finer taxonomy for follow-up questions whose scores roll up on their own.
Are custom drivers benchmarked?
No. Industry benchmarks cover CultureMonkey's standard drivers and eNPS. A driver only your organization measures has no external pool to compare against, so it will not appear on the Benchmark tab, and a survey built entirely from custom questions may show little or nothing there. That is expected, not a bug.

The limits

What the method does not do, stated in the same voice as what it does.

Is there a statistical model behind the score?
No. The score is an average, not a model: no weighting, no regression, no significance testing produces the number, and the documentation says so. What that buys you is a figure anyone can recompute by hand. Anonymity thresholds still apply before any cut of these numbers is shown.
Are the benchmarks competitor data?
No. A benchmark is an aggregated reference score drawn from a wider pool of organizations, keyed on your industry and company size, never a named competitor's number. It covers standard drivers and eNPS only, and not every combination has one, so an empty row is expected. A tenth of a point is noise.

See the calculation on your own data

Run one survey and recompute the score by hand if you like. Every number above is documented in the Help Center, and it will match.

ISO 27001·SOC 2 Type II·GDPR·★★★★★ 4.7/5 on G2