Making Data-Driven People Decisions

Having great data about our people allowed me to make better prioritization decisions and be in lockstep with our team's needs and desires. This was especially true as I scaled from dozens of employees to hundreds and from one office to half a dozen over the course of 18 months.

One of the first things I’ve done at every organization I’ve joined is start a monthly employee survey. The topics repeat every quarter, but important measures like employee net promoter score are asked monthly. Surveys should take 5-10 minutes, no more.  

When you start collecting data from this early in your company’s life, it allows you to have in-depth longitudinal information about our employee base, their sentiment, and their performance as you scale to hundreds of employees. 

Most companies wait until there is an enlightened HR lead to start building programs like this. The truth is if you wait until you have a big-time VP, you're likely putting them at a disadvantage when they join as they’re going to have to build the same information through their first few months on the job. I’ve seen other companies start to measure their employee experience when things seem “off.”The problem with this approach is that you have no baseline to compare against. 

Start collecting data early 

Having monthly data makes it possible to see trends as they develop. You’ll then be able to make micro-adjustments as opposed to major changes as you scale your company and culture. 

Here’s how to do it

Automate data collection early.

Most Human Resource Information Systems (HRIS) have built-in employee survey functionality. Some, like Gusto, have pre-made questions -- if you’re on a system like that, you should turn on the automatic surveys. If your HR system does not have a survey functionality I would suggest you choose one of two approaches: 1) start by creating your own survey via google docs or similar. This will require some management but will get you started. 2) Sign up for a service like Lattice to automate the collection and reporting. 

Ask many of the same questions

Done well, you should rotate through a few sets of questions on a quarterly basis, but ask a few of the same questions each month. Rotational themes might be Strategy, Vision, Management, Culture, Comms, Work Practices. Questions that get asked each month might be Employee Net Promoter Score, Happiness, Engagement, Workload, and open feedback. 

It’s important that the survey results are mostly anonymous. There should be some meta information to help diagnose and fix issues, but personally, identifiable information should be precluded. When your small team and location will likely suffice, when you’re larger you’ll want subteam, manager, and ideally the ability to view things like how long the employee has been with the company. 

Review the data

Review the data on a monthly basis. I’ve found that it’s best to write a summary of it that can then be shared with other managers (and potentially the company). The summary shouldn’t sugarcoat or obfuscate. When you get larger you can import the survey results into your Business Intelligence tools. 

Act of the data 

Acting on the data is the most important aspect. It’s important that you tell the company what steps you’re taking to rectify anything that comes up. This doesn’t mean that you need to fix every anonymous gripe, but you do need to focus on alleviating worrisome trends and nipping issues at the bud. It’s important to be able to explicitly defer action as well -- “explicitly” being the keyword. In my experience, reviewing the survey results and intended actions at Townhall and smaller departmental meetings is a great way to explicitly highlight your actions. 

Closing 

Having this information allowed me to make better prioritization decisions and to be in lockstep with the needs and desires of our team. This was especially true as I scaled from dozens of employees to hundreds and from one office to half a dozen over the course of 18 months -- while avoiding many of the common pitfalls of hypergrowth. By having historical survey data I was better able to get in front of things that were slowly breaking.