Get more from your staff survey

Staff surveys can be a self-serving exercise. So how can you make sure you're asking the right questions?

by Rhymer Rigby
Last Updated: 09 Oct 2013

There's more than a whiff of The Office about staff questionnaires. Having been endlessly exhorted by magazines like this one to listen to their staff, senior management commission - at great expense - a survey full of bar-charts, analysis and impressive-sounding jargon in order to seem to be doing so. But, often, little attention is paid to asking the right questions - and do they even want to know the answers?

If the results show a motivated, engaged and happy workforce, then, great, the bosses are heroes. If the findings aren't quite as spiffy as that, well, statistics can always be massaged to make them look better. And if staff morale is so bad that no amount of processing can make a silk purse out of a sow's ear, there's always the option of burying the evidence. Maybe in a year's time the exercise will be repeated, with questions suitably modified to encourage more acceptable answers, until staff tell the bosses what they want to hear.

'A lot of surveys start from the wrong point,' says Peter Hutton, author of What Are Your Staff Trying to Tell You?, a former Mori director and founder of Brand Energy Research. 'You start from saying "We've got to do a survey", rather than "We should systematically listen to staff",' he says. 'So you go and find someone who does surveys and you say to them: "Tell us what we should be asking", and you wind up asking a standardised list of questions that owes more to consistency than your own needs.'

This is not a terribly productive way of going about things. Management expends considerable time and money producing what amounts to self-serving propaganda, an expensive doorstop of a document from which they will learn nothing. Meanwhile, employees who may not need much encouragement to suspect their bosses' motives end up feeling ignored, frustrated and demotivated. People are generally adept at spotting bullshit, and bosses rarely get away with attempts at spin and dissimulation.

Even if the survey is conducted in a spirit of open inquiry, there are still pitfalls to avoid. The classic one is a failure to act on unpalatable results. Changing the decor in the toilets is one thing, responding to criticism of the firm's culture or strategy quite another. Too often, management seems unprepared to follow through on the feedback to any but the most trivial of questions.

'Employee surveys can often be an embodiment of the phoniness of the modern workplace,' says Stephen Overell, an associate director at the Work Foundation. 'On the one hand, you have people saying they're very interested in how staff feel, and on the other surveys become a kind of end in themselves. You acknowledge people's feelings and then you give up. Surveys are common, but solid evidence of them leading to action is rather rarer.'

A shame, because they can be enlightening and useful exercises for all involved. Thankfully, there are businesses that conduct exemplary staff surveys: they ask the right questions, communicate the results well and act on criticism or suggestions for improvement. But for an organisation wanting to take the pulse of its employees, the business of doing so can seem baffling. What should you be asking? How should you ask? How often should you do it? How should you respond?

Where do the problems lie? For starters, as Hutton suggests, employee surveys tend to be similar from one firm to the next. Says Virginia Merritt, co-founder of organisational development consultancy Stanton Marris: 'Because surveys often ask very general questions, people will simply fill them in depending on how they feel that day. If you've had a bad day, you tend to rant.'

So don't let your survey provider lead you by the nose. You understand your firm better than they do, and a bit of thought about what you actually want to know from your staff will pay dividends. 'What we see is great enthusiasm to try to measure the impact of a lot of initiatives, especially in the area of staff engagement,' explains Merritt. 'Surveys seem to focus on whether people are engaged or not. But what they need to ask is: what are they engaged in?'

One of the reasons surveys tend to ask generic questions, she adds, is that both the companies that use them and those that carry them out love to benchmark. It's a nice idea, but it means in practice that you're asking staff at a firm that builds ships the same questions as employees of an accountancy firm, regardless of the suitability of those questions and whether there's any meaningful value in comparing boatbuilding with bean-counting.

Benchmarking also calls into question what surveys are for. Are they an attempt to disover what your employees really think, or an excuse to indulge your competitive instincts by finding out how 'well' you're doing against your main rivals? Says Hutton: 'You often have a culture where what gets measured gets managed - and people just love to compare themselves to others, even if it's completely pointless.'

His pet hate is that old survey favourite, 'a long list of questions with an agree/disagree scale against them'. Despite its pervasiveness, the agree/disagree scale is often not the best tool for the job. 'It's widely used, but in no other area of market research do all major consultancies default to the A/D scale. I think it's because you can shove any statement in and it looks like an intelligent question.'

Enough of the bad, what about the good? For good surveying, says Hutton, you need to start with company strategy. 'There are three things you should be trying to do: first, to understand what creates and promotes job satisfaction; second, to understand what needs fixing in the organisation - and getting this from the people on the ground. And third, to capture feedback on what works well and what doesn't.'

So which firms do make a decent fist of their employee surveys?

Remarkably, Royal Mail (RM), for all its grim record of strikes, work-to-rules and general employee discord, is held up as an outfit that has done a great job - at least when it comes to restructuring the way it conducts staff surveys. Says its HR director Debbie Moore: 'We're certainly in a much better place than we were 14 months ago. We used to do an annual survey called "Have your say", which was locally based and didn't ask on a regular basis for views of what was going on across the organisation. We were concerned that we didn't know how people felt and what drove employee engagement.'

Of course, the more jaundiced Royal Mail customers among you might think its employee survey is the least of that organisation's problems, and that trying to fix it is like trying to bail out a sinking ship with a teaspoon. But if RM is facing some of the most serious employee engagement and motivation issues of any firm, surely its chances of doing something about them are enhanced by obtaining reliable information on where the problems really lie?

Top priority for RM bosses was to identify exactly what questions to ask. With 185,000 staff - comprising everyone from people in the financial centre and the management team, to agents, sub-postmasters and, of course, the posties themselves - this was a task that sounded simple but in reality called for a great deal of research, says Moore. All that effort paid dividends, though - RM was able to boil down the essence of employee engagement into a commendably concise eight core questions. 'And after that we have different questions for different groups,' she says.

And what of our old foe the agree/disagree scale? It's certainly there, but leavened by plenty of broader, thought-provoking queries with free-form answers - 'things like: "If you were the MD of the Post Office, what would you do?".'

As well as the big annual survey, RM does three smaller, more frequent 'pulse' surveys that sample 30% of the workforce. The first pulse in August 2008 registered an 'awful' level of engagement of 34%, from a response rate of 50%. That's now gone up to a more respectable 52% off a 54% response, although clearly there is more to be done.

The UK's largest building society, Nationwide, is another organisation known for genuinely listening to staff. 'We do an annual survey every December which goes to all employees,' says Jo Hussey, senior manager for employee brand. 'It's done online, so it's quick and confidential.'

Speed and ease of completion are vital for a good response rate, but often overlooked. Hussey explains that much has been done to reduce the survey's bulk. 'It used to run to 104 questions, but we've got it down to 59.'

These questions cover areas such as working conditions, resources available and reaction to the leadership team, as well as more general issues such as how staff feel about working for Nationwide and what they think of the competition. The response rate is an impressive 75%-85%.

Careful thought is also needed when deciding which of the key findings are the most important to act upon. 'Historically, we'd address the areas with the biggest drops, but we know through statistical analysis that just because something has dropped doesn't mean it's the most important,' says Hussey. 'So we look at the drivers of employee commitment through regression analysis. We then use this to give guidelines to individual managers. We say: "These are the top three drivers to focus on."'

A good employee survey isn't easy to do, but it's worth the effort. 'It helps you focus on the important things,' Hussey concludes - and knowing the difference between what's really important and what can wait is one of the key skills of management. The lesson is that, in business as in life, you only get out of your employee survey what you are prepared to put into it.


HOW TO AVOID THE POLL PITFALLS

Generic questions. One size does not fit all. Think carefully about what you really need to ask. The survey is there to tell you what your staff think, not to enable you to measure yourself against very different companies - or provide the survey company with a nice sideline in data to sell.

Too many questions. Surveys that are rambling and long-winded are less likely to be completed and may frustrate - if you start to notice more negative responses after question 75, there's probably an obvious reason why.

Agree/disagree scales. OK in moderation, but not everything has an answer that can be graded 0 to 9. They provide you with easy-to-tabulate data, but not much more.

No follow-through. You need to tell people what they've told you, and then act on what they've told you. If you're not going to, explain why.

Infrequent surveys. If you're going to figure out what's getting better and what's getting worse in time to do much about it, you need to survey regularly. Good practice is a large annual survey and smaller interim polls.

Imperfect anonymity. If you're to get honest responses, people must be confident that they won't be identified. So be careful when giving team-by-team analyses to line managers - if their team is small, it will be easy for them to figure out who said what.

Confusing dramatic with important. The score that rises or falls the most is not necessarily the most critical finding. Lots of people may be upset at changes in the canteen lunch menu, but it's unlikely to foment discontent in the way that a pay freeze or benefits cut does.

Spinning the results. Don't bother. Take unfavourable comments on the chin and people will respect you. Try to spin or use statistical sleight-of-hand and you'll be found out - and probably pilloried all over someone's blog or Twitter account.

Find this article useful?

Get more great articles like this in your inbox every lunchtime