What Do You Want From Your Data? Five Must-Have's

In a recent post, I wrote about the pitfalls of taking your data at face value and last week, I blogged about not knowing the true impact of critical organizational attributes without diving more deeply into the data (read: do more than explore it visually or look at up/down trends).

My business partner sent me an excellent article this morning by Stacy Harris that asks the question: Are You Considering Firing Your Employee Engagement Partner? I’ve been thinking a lot lately about data and how to make sense of it and I recommend this article for anyone else on a data journey. I think Stacy’s research applies to any type of data collection method. I’m summarizing what I took away from the article:

It’s not enough to have a survey; you need a strategy. A stand-alone survey gives you just that: one view whereas designing a survey that supports a business strategy provides data that can be aggregated with other data sets for a more accurate and complete picture.

If you ask generic questions, you get smiley or frown-y answers but no insight. Whether you have 12 questions or 40, if they don’t speak to your culture, your workforce or your targeted customers, I’m hard pressed to see how you will get any “aha’s” that are worth the time and expense of a survey.

What specifically do you want from a survey or any “listening post” that you design? “Because we’ve always done one” or “Because everyone else does it” are not specific objectives. What business issue do you hope to resolve? What’s keeping you or your boss up at night?

How are you analyzing the data? Scores alone may not tell you what you need to know. Trends without relationship to some rationale are just up or down points on a chart. Desktop tools exist to assist. Help is available (call me).

If you’re hung up on benchmarks, you may never get to the “why” of your own project. I know that some organizations swear by a benchmark study and who can argue, as long as the benchmarks map precisely to your own situation and as long as it’s not your only measure. Like generic questions, without the appropriate construct, a benchmark exercise can leave you with no specific roadmap for your success.

Collecting data is a critical component of every function these days. It's a project like any other, with objectives, outcomes and measures. Is your data giving you what you need?

What's Wrong With Taking Survey Data at Face Value?

I’ve been designing survey questionnaires and analyzing the data for so long that I often forget that some people may not be doing the deep dive and asking the hard questions of the data they’ve collected and for which our clients hire us (thank you, you know who you are). Maybe a little exploratory analysis, a tad of correlation, a glance at the verbatim comments and we’re done until the next time. Did we do a survey? Check. Did we do anything with it? Sure, sort of. Do we have a deep understanding of what the data means? Well….

What’s worse than not gathering intelligence from customers and employees?

Coming to the wrong conclusions!

I’m reminded of this fact by two articles I read last week: In This Case, Let’s Examine Dissatisfaction in the February issue of Survey magazine and Guest Perceptions on Factors Influencing Customer Loyalty in the current issue of the International Journal of Contemporary Hospitality Management.

In the case of customer dissatisfaction, the article suggests several calls to action:

Understand whether you have a category problem rather than a brand problem. In other words, your competitive space may allow easy switching with or without loyalty programs so make sure you know what your issue is before investing in programs that will not alleviate it.

Your market strategy will drive a customer’s perception of satisfaction. If you are a low cost provider, you have accepted that a lower level of quality and service is part of the equation. The danger zone you could find yourself in is in trying to be low cost while also attracting a customer who looks for a different level of product and service.

Benchmarking. I’ve never been a fan but lots of companies do it and the swirling vortex that you get sucked into is that you compare your performance to companies who target different customer segments.

Dissatisfaction may not arise from what you do but rather what other, similar companies do that you don’t do. Customers constantly evaluate decisions based on alternatives; some amount of dissatisfaction arises with your product and service even if you are executing your strategy perfectly.

My suggestions for arriving at the best conclusions possible from your data analysis:

• Keep your strategy upper most in mind when designing the project and return to it often when analyzing data. This means knowing who your competition is; who the ideal customer is and what your competitive advantages are.

• Design survey questions to be particular rather than general. The more generic the question, the less likely it is that you have actionable data and the more likely you potentially are arriving at the wrong conclusions.

• Don’t confuse happy with satisfied. If you want to meet a customer’s needs, you are aiming for satisfaction. If you want happy, that’s a whole different level of expectations.

• Perform data analysis from several different perspectives. Not all survey questions should be treated equally in reaching conclusions.

I’ll write more about the last topic next week.

Transform Your Metrics From "So What?" Into "Who Knew?"

There are a lot of 3D movies out; have you noticed? I don’t seek them out but I appreciate the fact that people may enjoy a film more when it is multi- dimensional and they can feel immersed in the action.

I think we get too fond of our metrics; we have them because we’ve always had them. We track metrics and manage them and present their variances against performance goals. The problem? They often are one-dimensional and not very meaningful outside of our own function. And, if they aren’t tied to a real business outcome, it’s hard to make a case for the programs we want to implement. We may not monetize metrics, which is the language of our bosses; so there’s a sense of “so what?” when we present.

So, how do you make a common metric like turnover (employee or customer) more 3 dimensional and get people immersed in your action?

Embed Metrics With Data: Not just the obvious data of people in/people out. Drill down; explore data. There’s an “aha” in there I promise and since you have the business context, there is no one better positioned to see it and explain it.

Use Data Sets from Other Departments: Make your metric multi-dimensional by bringing in data from HR, Sales, Marketing, Operations, Process, Call Center: whatever data set you have, add to it in a smart way by collaborating with other departments who also have valuable data that isn’t yet insight. We have to dismantle data fiefdoms and share. Where does turnover impact the business? How does it impact the business?

Try Simple Statistical Tests: This is the point at which people click off because they think it’s not in their skill set. If you have Excel on your PC, you have a statistical toolkit. Invest in a great little e-book that provides a huge amount of good information and it is well presented (Using Excel to Solve Business Problems by Curtis Seare). Try out various assumptions to see which are more powerful. Who is leaving? What is driving turnover? How does it affect customers? How does it impact employees? Where does it affect business goals? Experiment with results and keep testing.

Provide a Business Context: Sometimes people get hung up with statistics, even simple ones and forget that the most important point is taking what statistics can tell you and mapping that to what you know about the business.

Tell Me Something I Don’t Know: aka Monetize the Results. When you know what turnover really costs the company and what it costs to improve the situation, you will have the attention of people who haven’t seen your metrics/data/ideas presented in a way that they understand.

Then, your metrics are multi-dimensional and provide real intelligence for the organization.

How are you helping your decision-makers get immersed in your metrics? Are they in 3D?