Making sure you can analyse your survey data as you mature
A few weeks ago Redthread Research published their latest People Analytics technology report. One of the many interesting findings relates to how People Analytics use cases change as they become more sophisticated. They identify 3 use cases:
Phase 1: PAPs use vendor tools for understanding a specific HR area (e.g., engagement), integrating data from other HR data sources (e.g., HRIS), and presenting it in dashboards; senior leaders begin to leverage dashboards.
Phase 2: PAPs use vendor tools to integrate a broader set of people-related data and some operational data, and provide a continuous stream of data; other leaders increasingly use these more robust dashboards and insights.
Phase 3: PAPs use vendor tools to export the integrated data, and add it to a data lake or to run additional analysis on the tools of their choice, such as Tableau and Power BI; leaders broadly adopt the dashboards and other capabilities to answer business questions.
This is very similar to the way we’ve observed the market for some years. The caveat that I’d make is that there is a dynamic where the types of use cases that typically required a ‘Phase 3’ a few years ago move down the maturity levels. It is certainly worth assuming that this dynamic will continue, with many of today’s advanced use-cases becoming widely available in the future.
What is difficult to envisage is a time when all use cases will be available within tools. Each business will always have some organization-specific needs, often linked to the strategy of the organization. If this was not the case we’d expect to see the increasing sophistication of tools decreasing the needs for firms to hire data analysts, whereas in reality we see analyst demand rising at the same time.
Our role in this is relatively straightforward: All of our clients are exporting data (text + linked data) from tools where they’ve been collected - mostly employee engagement / experience / voice tools. The text is, at the point of export, unstructured data. We then analyse this data effectively creating new structured data which is needed for further analyses and to use in BI tools. About half of our clients are predominately aiming for us to help them interpret this data, identifying rich patterns and helping them focus on which areas to prioritise for action. You could think of these as Phase 1 & 2 clients. We’re often using BI tools like Tableau to help report this data.
The other half are additionally wanting us to return structured data that enables them to incorporate text feedback into their existing dashboards and analyses. For these ‘Phase 3’ clients we’re providing a necessary data-transformation role between the tool which captured the data & their BI-based reports. We will discuss these requirements up-front so that when we return data it will be in the right format for their tools.
From our experience, it would be wise for all organizations, whatever their current position, to assume that they will increase in maturity over time. In the Redthread model they will therefore move closer to ‘Phase 3’ and will be wanting to export their information out of the survey tools so that they can perform more business-specific or advanced modelling or reporting.
Preparing for maturity
There is one ‘hidden’ barrier that we see time & time again hindering the progress of People Analytics teams as they move towards maturity - the difficulty that many system providers make getting well structured data out of their tools. On numerous occasions over the last few months we’ve had to ‘reverse-engineer’ data to build acceptable amounts of information on who provided the answers. This is time-consuming work but it can be done.
Some of the employee engagement tool vendors go a step further and tell clients that legally they can’t export data for further analysis. This is not likely to be true, it’s fundamentally down to two reasons:
First, there is a commitment that you will make to your employees when you collect the data. You can change this commitment. A large section of large enterprises are using survey tools where a defined number of users in the organization can see all data. All it requires is that the commitment you make defines this.
Second, we believe some of the vendors perceive that controlling the data increases the ‘lock-in’ that they have. Many of the large survey providers state that the cost of a survey tool is falling towards zero. It is the non-survey functionality where the value is created (benchmarking, analysis, reporting etc). Controlling access to data forces you to be reliant on their tools for your analysis. It is one mechanism to stop you switching to a lower-cost tool.
How not to fall into this trap
The best time to get what you want & we predict you will likely need is at the procurement stage. You need to ensure that data access is written into your contract.
Make sure your RFPs define what you need in terms of data access. You need individual-level data. This can be anonymised (eg using a hashed employeeID) but it should have some defined demographic data attached.
You should work closely with your legal team to ensure that your commitments to employees when you collect the data do not prevent you from analysing the data outside the tool at a future date.
If a provider can’t provide you with reasonable access you should probably choose another. There are top-level providers who do give access, and ones with a reputation of making this difficult. Choose the former.
Summary
It is becoming apparent that companies should plan for how they will add survey-data to non-survey employee data as they become more mature.
There are technical challenges but primarily legal challenges that might prevent you doing this. The most problematic are likely to be legal ones. Engaging with your legal team up-front is the easiest way of capturing value.
Finally, to incorporate qualitative data such as text into your analyses and reports you’ll need to transform it into structured data. To understand the sizeable benefits that are achievable here please get in contact.