In an ideal world, all organizations would have a vested interest in keeping their employees happy. They’d take those annual employee satisfaction survey results seriously and translate needs into expectations people could depend on. But employee satisfaction — which sometimes just means not hating your job — is not the same as employee engagement. Engagement speaks to something much larger and more impactful on business goals and outcomes.

But there’s a problem with how we define and assess employee engagement today: clinging to outmoded assumptions and methods makes it hard to get to the root of problems. And, even when we can, we don’t always know how to deal.

An A** Out of You and Me

As awesome as it would be to use metrics like pay rates and perks to create a perfect workplace for everybody, it’s utterly unrealistic. Back in the days when metropolitan office jobs were something major to aspire to — and usually reserved for the already-prosperous — that might’ve been enough. But it’s high time we stopped assuming that today’s worker is still only in it for the paycheck.

According to the pros over at Custom Insight, “Employee engagement is the extent to which employees feel passionate about their jobs, are committed to the organization, and put discretionary effort into their work.” That’s a nice, qualitative way of saying it. Jargon aside, employee engagement is evident when getting to do the work means more to someone than the reward given in exchange for it. It means innovation, excitement, and self-motivation. Said marketing exec Billy Cripe in a recent social business Twitter thread, “The Hallmark of engaged employees: work is what they want to do, not just a place they go.”

Measuring Up

Most companies have surveys in place that are designed to determine both employee satisfaction and engagement. Data is great, especially in terms of historical comparisons and tracking — but it’s not everything. First of all, not everyone takes these surveys seriously, because they often don’t see what good it’ll do. Credibility can be an issue, too, if an employee is unhappy. Finally, while every employee’s input is valuable, it isn’t equally so. In fact, Forbes’ Cy Wakeman encourages playing favorites. “Listen to what your top performers tell you,” she writes. “They’ve proved their value and earned their credibility.”

When the data is affected by such a high level of subjectivity, it’s important to gather even more of it. Customer Insight notes that while short surveys are great for tracking if employees are engaged, “they have a hard time explaining why…because they lack detail.” To get a truly accurate look at overall engagement, surveys should cover a vast array of topics and include between 50 and 80 questions. Even then, the translation of results into action should be tempered with flexible, somewhat abstract considerations. For example, are the more negative responses coming mostly from one department? If so, has that department recently undergone a reorg? Does it typically have high turnover? Or, are the positive responses coming from higher-paid employees? How about brand new folks who might not really know how they feel yet? Even something like timing — right after a Christmas party, when people feel appreciated, or after a disappointing quarter, when they don’t — can have a major impact on your data.

Be careful to take circumstances into account, or you could easily end up with statistics that are basically useless.

After the Facts

Obviously, employee engagement doesn’t end with assessment. Perpetuating and improving it is beneficial to the overall ambitions of any organization, and is tangibly reflected in boosted productivity, profits, retention, and perhaps most importantly, a company’s ability to attract talented people. It even creates more loyal customers, since employees are wilfully accountable and strive to deliver awesome services and results.

Companies that use data wisely to support and shape work experiences will prove to their employees that they’re engaged with them, too — and that’s much more valuable than simple statistics.