Leaving value on the table….year in, year out. Virtually every organization is doing a regular survey of employees’ thoughts and opinions every year or two, with some firms doing more frequent ‘pulse’ surveys. However rich this data is, and as powerful as some of the resulting analyses may be, I still see a blindingly significant amount value being left on the table in the world of survey data due to the lack of integration with other data sources.

In most cases, survey data is only integrated along a few key data points such as demographics, tenure, and supervisor – information that is most often provided by the survey participant. Where the survey is ‘confidential’ – allowing the survey firm to capture unique employee identifier in the process – other data points may be incorporated such as performance ratings, or promotion history. This integration is done as an offering of the survey firm, leveraging a point-in-time extract of HR data to drive various cuts of the data in reporting, and sometimes deeper analytics.

Even under the most robust of these scenarios, there is value left on the table. While the survey firms may have the data to continue answering subsequent questions, the companies themselves are essentially ‘locked out’ of delivering basic analytics or answering the follow-up questions (without incurring extra expense with their vendor). Additionally, the difficulty and cost of providing the data extract means that the original data extract will likely be used even months after it was delivered. The result: many questions go unanswered, and eventually, they just go unasked.

Integrating the survey data into a data asset that the company can access is the way to break these limitations and release the value of the survey data all year long.

When is the last time you reviewed your survey report? Chances are that you, and most other leaders in your organization know where the report is, but are not using it. It is a printout, or maybe a link to a website buried in an archived email. In the best of situations, an action plan has been created based on the results and it is being checked regularly for progress. What would be different if survey data was available more regularly throughout the year? What if the survey data could be leveraged throughout the year as a source for answering questions as they come up (without the need to engage a survey vendor or workforce metrics vendor to get it done)? What if survey results could be easily embedded into key metrics reporting, providing ongoing visibility into this valuable data?

Maintaining Confidentiality while Integrating. The main challenge is in preserving the confidentiality of the survey data while integrating it at the employee level. In many ways, survey data is treated with greater protection than compensation and even some health data – which is accessible by some of the employees within the organization for system support and also for analysis. Keeping survey data out of reach of all employees hands will most likely require an out-sourced data asset with sufficient security that prevents employees from accessing the information at two levels: 1) no access to individual responses, and 2) metrics only accessible for groups of ‘x or more’ respondent (typically 5 or 7). There are ways this can be done, with the data being transferred from survey vendor directly to the data warehouse/metrics vendor – keeping the detail data at arms-length from the employees.

Integration releases the full potential of survey data. When survey data is integrated at the employee level with the broader set of data from HR systems (performance, potential, promotions, terminations, recruitment source, etc.), it provides an organization with the ability to easily answer questions like the following:

• What is the engagement level of our highest performing employees?
• What is the predictive accuracy of the ‘intent to leave’ question?
• How satisfied are the employees that are in ‘critical roles’?
• Does ‘manager satisfaction’ translate to higher retention in our organization?
• Are recently promoted employees who have been promoted recently more/less engaged?

It is not that these questions are new. They are not. What’s new is the ability for a company to answer these questions without incremental cost – driving down the cost of information to drive the business is a key objective of ‘business intelligence’ activities. In especially rich data sets, the survey data can be linked to financial, productivity, customer, and quality metrics – all of it with greater flexibility than ever before.

Survey data in a changing organization. A month or six months later, the organization has changed – some employees are gone, new ones are in, and the organizational structure may have shifted. Survey data that is locked in time becomes less actionable as the year progresses. When integrated with the employee and organizational structure data, the survey results can be ‘re-cast’ each month to reflect the latest organizational structure – keeping the survey data more ‘fresh’ through the year. This is similar to tracking high performer rates – even though the results are based on a point-in-time event (performance review), we continue to look at the latest data of what percentage of the organization (or business unit) are high performers.

Confidential v. Anonymous. Surveys that can be tied to a specific employee are considered ‘confidential’. The respondent’s results are not shared with the employer, but are available for the survey firm to combine with other data from the employer to provide richer reporting and analysis. ‘Anonymous’ surveys are not attributable to a specific employee, and are most often tied to a supervisor or business unit based on the respondent’s selection from a list. More companies are moving toward ‘confidential’ surveys to improve the analytical value and also the accuracy of their surveys.

Quality of Data under Confidential and Anonymous Surveys: There is the frequent question of how honest the employees will be in a ‘confidential’ survey where the employees know that somewhere the results can be attributable to them, and whether the response rates will suffer. Even with the best communication and assurances of confidentiality, leaders fear employees will hold back or opt-out. However, the accuracy challenges of the ‘anonymous’ model need to be acknowledged as well – as there will always be some individuals that take the survey multiple times whether intentionally or not (e.g., a leader of 5 people gets 7 responses on the survey). In an ‘anonymous’ survey, the respondent is the provider of all information: demographics, tenure, supervisor/unit, and sometimes performance rating. And a less likely, but very real risk to survey data quality, the ‘anonymous’ survey may provide cover for what would be called ‘moral hazard’ in the world of risk management – namely, some employees may be providing intentionally misleading information attributed to their own or another leader.

Going ‘confidential’ usually has some impact on response rates – I have heard numbers of about 3-5 percentage points being the norm. It should not be assumed that the drop is due to employees’ resistance to the new model, as one of the key drivers is the systematic prevention of duplicate responses. Provided there is effective communication for the change, ‘confidential’ surveys should more than offset these response rate ‘costs’ with the increased analytical & reporting value. The survey firm Kenexa has an interesting paper on this – where they use the terminology ‘Attributed’ and ‘Non-Attributed’ to cover these different survey options. Here is a link: http://www.kenexa.com/ResourceCenter/ThoughtLeadership/Attributed-Versus-Non-Attributed-Surveys

QUESTIONS you should consider asking within your organization:

1. Is your survey ‘confidential’ or ‘anonymous’? Has the value of a ‘confidential’ survey been considered? Are there cultural obstacles/concerns to ‘going confidential’? What are the accuracy concerns that are currently known/accepted in the ‘anonymous’ model?
2. What analyses and reporting needs have gone unmet? Chances are that some great work is not being done either because they couldn’t be done based on the survey data, or because of the cost to have them completed by the survey vendor. There may also be some influential leaders that have questions they want answered, and are not aware of the potential to fully integrate the survey data.
3. What is the engagement level of our top performers and those in critical roles? The value is in the integration of the data to answer more impactful questions from which strategic decisions can be made. It is critical to not lose track of the fact that this is about improving the business. If that is not clear, it will just be tossed aside as another interesting idea.

Nicholas Garbis, Sr. Consultant
Infohrm – global leader in workforce planning, reporting, and analytics
www.infohrm.com

This entry was posted on Friday, October 16th, 2009 at 5:15 pm and is filed under HCM. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.