Tell me what’s on your mind/ What you feel, what do you believe/What is inside you that makes you scream? / There’s something I need to say before you walk away – Tell Me, Failed Flight

The very first (and sometimes the only thing) most organizations do when they embark on the Employee Engagement journey is to conduct a survey. An in-depth and well designed organization wide survey still is the best way to capture what employees think and establish a baseline for current levels of engagement. (Till neuroscience catches up, but that’s still some way off)

Do you need an outside ‘expert’ to come do your survey? Maybe. But that depends on various factors.

How large and diverse is your organization? If you are a multinational spread across multiple countries with a large multicultural workforce, then maybe you are better off getting someone who truly knows his stuff. Getting an external source also usually gives you the added advantage of providing industry benchmarks.

Do you have resources to pull off a detailed survey? If you are a small or middle-sized organization, chances are your tech team is already stretched to the limit with work. There is always some new feature to roll out, some bug to fix. Telling them to setup the required tech for an employee engagement survey might not go down too well. On the other hand there are enough online resources that let you conduct your survey quite professionally with minimal effort. (Surveymonkey, Google Apps are two good options)

And the third (perhaps the most important question irrespective of size/budget/resources), do you know what to look for? Or rephrase it as do you know how to design a good survey?

To help answer that third question, let me dive a bit deeper into the two aspects of the survey that will determine its efficacy and ultimately determine ROI of years of effort and money that the company will put into Employee Engagement.


The first question that most people try to tackle is how many questions does one ask employees? Are 5 enough, or do we need 50? And then the next question pops up: About what all aspects do we ask questions on?

If these questions are popping into your head – Stop! You are making the mistake of “putting the cart before the horse”. Before thinking of questions, focus on the goal. This means thinking deep into what factors you think drive employee engagement in your organization and then determining what about those drivers are you seeking to establish from the survey.

Start off by capturing the broad and general goals of the survey. For example you might come up with something like “We want to discover how employees feel about: (a) Compensation (b) Work Life Balance (c) Management … and so on.

If you are a startup in growth mode, Work-Life balance may not be a smart thing to ask employee about. (Chances are very few of them have anything remotely close to a balance. It comes with the territory of a startup) You might be better off checking if employees feel they are aligned on the larger goal because that is a critical success factor. But if you are fifteen year old mid-sized family run firm whose top-management insists they still ‘think like a start-up’ then Work-Life balance might be definitely something you want to think about.

This is the first stage in designing your survey and likely will take a few days and multiple iterations to cover all areas that might have an impact on employee engagement in your organization.

And Well-Designed:

Once we establish the goals of the survey, we take a deep breath; get a big cup of our favourite brew (best to stick to coffee and tea at this point); roll up shirtsleeves and dive deep into designing the survey!

The next step is for you to establish the ‘research’ questions. These are not the actual questions, but the questions that you are seeking to answer. If you have established the goals of the study this part should be a snap.

There are essentially two types of questions (statistically speaking)

  • Testable Questions (or Close ended Questions): These are the questions, which can be eventually boiled down to a point that can be validated/answered by a statistical test. For example a check to establish any significant relationship between the age of an employee and engagement is a testable question.
  • Non-Testable Questions (or Open ended Questions): These on the other hand are, you guessed it, questions that cannot be answered by performing a statistical test on the responses. They are typically used to collect information for which there could be a multitude of responses – for example, ‘What are things employees would like changed to improve their productivity?’

A good survey will usually have a mix of both close-ended and open-ended questions. Too many close-ended questions risk rendering your survey presumptive, since you will have to come up with all options and won’t have a chance to capture ideas and opinions. Too many open-ended questions might make your analysis phase a nightmare since dealing with free-text is anything but easy when the survey has hundreds or thousands of respondents.

“… the null hypothesis is never proved or established, but is possibly disproved, in the course of experimentation. Every experiment may be said to exist only to give the facts a chance of disproving the null hypothesis.” R. A. Fisher

In the final stage before you start thinking of questions, you need to convert each of your ‘research’ questions into a null hypothesis.

A research question ‘Is there a (significant) relationship between active mentoring by seniors and employee engagement?’

becomes a research hypothesis just by flipping the order of the first two words, ‘There is a (significant) relationship between active mentoring by seniors and employee engagement.”

One cannot test a hypothesis in statistics, so we convert the research-hypothesis into a null hypothesis: “There is NO (significant) relationship between active mentoring by seniors and employee engagement”

The corresponding question in your survey might be “Do you feel more engaged with your job role when you are mentored by a senior resource?” and have a response option on a Likert Scale with Five or Seven levels. A typical five-level Likert item is as follows:

  • Strongly Disagree
  • Disagree
  • Neither Agree nor Disagree
  • Agree
  • Strongly Agree

You might choose to remove the “neutral” option (c), and create an even-point scale to use a “forced choice” method.

Ironing out the creases:

When you are done designing your survey, there is one least thing to be done before you unleash it on the organization – get rid of or correct the confusing statements in there. When you have been immersed in creating the survey for days together, you might put statements in the survey that might seem obvious to you but might be confusing to an employee. And in any decent sized organization there will be no chance to correct it once the survey is live.

There are essentially two ways you can pre-empt this (in the best possible way).

  • Use Statistics (again): Send out your survey to a sample of 30 or more people and see their responses (and feedback). Make changes to those areas of the survey that seem confusing to the respondents.
  • Take a hands-on approach: Select a few respondents who are representative of the various segments in your organization (Age/Gender/Role) and be present when they take the survey. This will enable them to ask you questions when they are stuck or unsure of something in the survey. Any question a respondent asks is a defect – in the final version there will be no opportunity for respondents to ask questions. You can then fix those issues right away with a tight feedback loop.

My, my, just look at the time:

One last point about your survey – keep note of the time people take on an average to complete your survey. 2-3 minutes seems to be the sweet spot. Anything more than that and chances are your respondents will get fatigued and just click through without really paying attention. Less is better – as long as your respondents don’t feel underwhelmed and start to wonder what just happened when the Thank-You page pops-up. (The perception of a shallow survey can do more damage to your employee engagement efforts than you assume)

These broad guidelines should help you design a thorough Employee Engagement (or any other) survey for your organization. In the next (and concluding) part on DIY Employee Engagement Surveys, I will talk about the post-survey analysis.

3 thoughts on “DIY Employee Engagement Surveys: Bootstrapping your way to insights. (Part 1)

  1. Nice post. I was checking continuously this weblog and I’m impressed! Extremely helpful info particularly the last phase :) I take care of such info a lot. I was looking for this certain info for a very long time. Thanks and best of luck.


We love discussions! Please do leave a comment on what you think about this post.

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.