How to measure organisational culture: methods and limits
By Deibson Silva · Reviewed on
To measure organisational culture is to describe with data how each area of a company really decides, how it coordinates and what it rewards, and to compare that with what the company says about itself. It is done by combining aggregated behavioural profiles, anonymous surveys, conversations and a reading of real decisions, because no single method is enough.
What are you measuring when you measure culture?
The way things get done when nobody is giving instructions. How a hard decision is taken, what happens to the bearer of bad news, who moves up, how long it takes to say no.
Every company has two cultures. The declared one is on the website and in the welcome presentation. The lived one shows in meetings and in decisions. Measuring serves to describe the second and to see how far it sits from the first.
What you get is a description by group: this team decides with data and takes its time, that one decides fast and corrects later. No single number comes out, and no mark for the company.
Which methods are there, and what does each one show?
Six, which complement each other. The table says what each one shows, what it leaves unseen and when to use it.
| Method | What it shows | What it does not show | When to use it |
|---|---|---|---|
| Behavioural profiles aggregated by area | How each team tends to decide and what moves it | What people think of the company | At the start, to compare areas |
| Engagement survey | What staff think about specific drivers | Why they think so | Once a year |
| Pulse survey | How one driver changes over a short time | The whole picture | After an action |
| Interviews and discussion groups | The reasons and the examples | How many people think the same | After the survey, to understand it |
| Reading of decisions | What the company rewards in practice: promotions, departures, recognition | What those who do not decide feel | Each year, alongside the survey |
| Observing meetings | Who speaks, who decides and how people disagree | What happens outside the room | When two methods contradict each other |
Why is a survey not enough?
Because it gathers opinions, and culture is made of behaviour. A team can score highly on 'I feel listened to' and always decide what one person says.
The answer also depends on the moment. The same question, asked in the week of a difficult close or in the week of good news, gives different results. And each person understands the words in their own way: 'autonomy' means one thing in sales and another in production.
That is why the survey is read next to another source. The map of profiles says how a team tends to decide; the survey, how it feels about that; the conversation, why.
How is anonymity protected?
With a minimum number of responses per group and with care when data is crossed. If a team has three people, any result of theirs points at them. The prudent course is not to show groups of fewer than five and to combine small teams with others.
There is a less obvious trap. If you show a department's total and all of its teams but one, a subtraction is enough to know what the missing one said. To prevent that, when one group is hidden the next smallest is hidden as well.
Written comments need more care than scores, because the way someone writes identifies them. With few responses it is better to summarise the themes.
Anonymous is a word with legal weight. Recital 26 of the GDPR leaves anonymous information outside the regulation, but only when the person is not identifiable, taking into account all the means someone could reasonably use to identify them. A survey with the name hidden and the area, shift and length of service in view may not qualify.
What are the limits of measuring culture?
The first is size. In a company of thirty people there are hardly any groups that can be shown, and direct conversation gives more than any figure.
The second is that asking changes things. A survey creates the expectation that something will happen. If nothing does, the next one will get fewer answers and a worse tone.
The third is use. The result describes groups. It helps to decide how a team is organised or what managers are trained in, and it is not for deciding about one person.
The fourth is pace. Culture changes slowly. Measuring all of it more than once a year tires people and barely shows differences; pulse surveys are there to follow one specific change.
And the fifth is comparison. Knowing that another company scores higher on a driver helps little. Comparing two of your own areas, or the same area at two moments, helps far more.
What do staff and their representatives have to be told?
Before you ask, each person has to know who is handling their data and why. Article 13 of the GDPR requires it. Article 5 adds that data is collected for specified purposes and is not then used in a way incompatible with them: what was collected to understand the culture does not turn up later in another decision.
If the measurement uses personality or behaviour profiles, the Spanish data protection authority's list places them under its first criterion and expects an impact assessment in most cases where two or more criteria are met. Check this before you start.
And in Spain, if the measurement is going to lead to decisions that bring relevant changes to how work is organised, Article 64.5 of the Workers' Statute gives the works council the right to be informed and consulted. Other countries have their own consultation rules.
What do you do with the result?
Choose. Out of everything that comes up, keep one or two differences that affect a pending business decision. If two areas that have to coordinate decide at very different speeds, there is work to do there. If the difference bothers nobody, note it and leave it.
Then comes a short plan with owners and dates, and a brief second measurement to see what changed. Giving the result back to those who answered is part of the plan.
Where do you start?
With one business question and two areas. For example: why do the projects that engineering and sales share run late? Measure those two areas with two methods, one of behaviour and one of opinion, and talk to five or six people in each.
That gives you a description senior management recognises and a first decision. Widening it to the rest of the company makes sense once that first round has been of use.
Sources
Frequently asked questions
- How often should you measure organisational culture?
- All of it, once a year at most. In between, a pulse survey on what is being worked on is enough to see whether anything is changing.
- Can it be measured in a company of fewer than fifty people?
- Yes, with fewer breakdowns. There will be few groups of five or more, so compare two or three large areas and rely more on conversations.
- Are culture and engagement the same thing?
- No. Engagement is how staff feel at a given moment and can change from one month to the next. Culture is how people decide and work, and it changes over years.
- Who should see the results?
- Senior management sees the whole, each manager sees their area and staff receive a summary with the actions decided. Agree this and say so before you ask.
- Can it be used to decide about particular people?
- No. The result describes groups. Using it to assess one person breaks the trust of those who answered and goes beyond the purpose the data was collected for.