How to Build a Hypothesis for your People Analytics Projects
It is crucial to devise a good hypothesis for your people analytics project in order to connect people data to real business problems. A good hypothesis also helps to ensure that you select the right analytical approach for your project and that the project isn’t biased by pre-conceived notions.
The expectations for a HR Business Partner (HRBP) have shifted since the origin of the HRBP over 20 years ago. Now, HRBPs need to adopt a data-driven mindset and learn to craft excellent hypotheses to be an effective bridge between people analytics and business issues – this is how to ensure that you add real business value with your people analytics projects.
HRBPs are regularly having conversations with the business and must be able to get to the heart of the problem that the business is trying to solve. Then, they must be able to translate that to the people analytics team - or to the work they themselves are doing - to drive the right insights and take them back to the business.
But don’t just take our word for it! Here’s what some of our guests on the Digital HR Leaders podcast, hosted by David Green, have to say about building hypotheses:
Series 5: Episode 5: The Role of HR in Driving Successful Org Design with Julie Digby, Global Integration and Transformation Leader at Mars.
David Green: What does digital transformation mean for HR professionals in terms of some of the skills that they need to develop, particularly when working in close cooperation with the business?
Julie Digby: I think we need to get more data and data driven, more comfortable with it and when I say that I don't necessarily mean that we need the data manipulation in the hands of the strategic business partners. But we've got to have the tools and we've got to have the specialists who are by the side of the strategic business partner that can access and show them the data and test their hypothesis. So, they need to have the ability to constantly be saying, okay, what's happening to our organisation? Does that still fit the strategy?
Series 5: Episode 2: The Evolution of People Analytics at Microsoft with Dawn Klinghoffer, Head of People Analytics at Microsoft.
David Green: You talked about this concept of data minimisation and I think it'd be really good to share that cause that's not something I hear from a lot of practitioners.
Dawn Klinghoffer: So, we have a required training every single year for HR employees to take, data minimisation is exactly what we try to teach people in terms of, only pull the data that you know that you're going to need for the analysis. You can always add later, but there needs to be a business reason whenever you're adding data elements to the analysis.
I think that most people just want everything and the kitchen sink from the beginning because they're worried that if they don't pull in everything, they're going to miss something. And again, all the work that I try to get my team to do is hypothesis driven. And so, if you are, if you were being really true to a hypothesis driven analytics methodology, then what you're doing is you should know what data elements that you should pull in.
Series 5: Episode 4: How McKesson used ONA to Drive Sales Performance with RJ Milnor, VP, Talent Management Operations at McKesson.
RJ Milnor: So that consulting piece is important. What we're increasingly finding is to appropriately scope the work. That consulting skill set is so important as well because we need that understanding of the business and why it's important to them. And to make sure we have it, we have it set up and the research setup appropriately.
David Green: And develop the hypotheses and get down to the right level so it's the right questions rather than maybe the questions that could be at the top level, really drilling down to the...
RJ Milnor: It's a fantastic point and I think we've seen examples of that recently and I'm very proud of the team for identifying that where we might be given a question or hypothesis, and we realise, you know, that's a surface level problem. It's a real problem. And our clients and the business are feeling real pain.
Although hypothesis building is a scientific practice, it doesn’t have to be overcomplicated. In this bitesize learning video, Jonathan Ferrar and Ian Bailie explain how to build a clear and concise hypothesis for any analytics projects you are embarking on.
This bitesize learning video is taken from our online training course Framing Business Questions and Developing Hypotheses for People Analytics, in which Jonathan and Ian go into more detail, breaking down the first two critical steps in any analysis project: framing business questions and developing a hypothesis. They dive into 5 key topics within these steps to help you ensure you’re adding business value with your analytics topic, looking at how to frame the business question, whether the business question is relevant through to how to actually define hypotheses.
What is a Hypothesis?
According to the Wikipedia definition a hypothesis is:
‘A supposition or proposed explanation made on the basis of limited evidence as a starting point for further investigation’
A hypothesis is a statement that introduces a research question and proposes an expected result. Quite simply when we talk about hypotheses what we mean is that we're creating a really simple statement that we can use to determine whether something is true, or false.
Interested in learning more about running a successful People Analytics project? Take a look at our online People Analytics certifications on myHRfuture
An example of a hypothesis might be:
“Attrition in salespeople is higher this year than it was last year.”
Is that true or is it false? And how might we be able to test that?
It is important to mention when we're talking about hypotheses that often the concept of creating one can seem quite daunting, making us feel like it’s leading into a huge analytics or statistical project, however, it’s worth noting that this doesn’t have to be the case. Ultimately when creating hypotheses what's really important is that you come up with a simple statement that can be tested.
Three techniques to help you build a Hypothesis
When building a compelling hypothesis there are a variety of different techniques that you can employ to help you ensure yours is simple and compelling.
It’s a statement not a question
The first technique to note is very straightforward, ensure that you create your hypothesis as a statement, not as a question. It is very simple - when you've written your hypothesis, if it ends with a question mark then it is not a hypothesis that can be tested. So, be sure to write it as a statement with a full stop at the end. Remember a hypothesis outlines a theory that can be tested; therefore, it must be positioned as a statement.
The ‘If – Then’ Technique
The second technique is the ‘if – then’ technique. This technique is very straight forward and does exactly what it says on the box. When building a hypothesis using this technique you will end up with a statement that outlines
“‘if’ you do this… ‘then’ this will happen”
This technique from a purely scientific, analytical method allows you to look at the data and analyse the consequence. If a specific action is taken, then a certain outcome is expected. An example might be:
“if we pay people more then they will stay longer”
or
“if we provide more online training courses then people's performance will improve”
The “Yes, No” Technique
Often hypotheses are more commonly phrased in terms of correlations or effects, where you directly state the predicted relationship between variables. This final technique is the “yes, no” technique. The yes, no technique is where you fashion your hypothesis as a statement that makes a prediction that only has a yes or no answer. An example might be:
“Attrition is increasing in the factory”
In that factory, attrition is either increasing or it's not increasing. It’s important to note that with this form of hypothesis crafting you may end up with supplementary questions, which requires you to create multiple hypotheses. Multiple hypotheses are used when we wish to consider multiple hypotheses statements simultaneously.
So, as we’ve covered when building your hypotheses for your people analytics projects remember that it should always be a statement not a question. To help you build a solid hypotheses statement, you can leverage either the “if… then” or the “yes, no” technique.
Get Certified in the HR Skills of the Future and Press PLAY on your Career!
myHRfuture is a learning experience platform supporting HR professionals like yourself, invest in your career. Our online HR certifications comprise of a blend of HR training courses, videos, interviews, podcasts, case studies and articles all expertly curated to help you build tactical skills while providing real life context to what you’ve learnt.
All of our online HR certifications reflect the core HR skills you need to develop for the future, giving you a professional edge while providing evidence of the skills you’ve mastered, helping you get ahead and press PLAY on your career all for £25 per month. What are you waiting for? Sign-up today!
ABOUT THE AUTHOR
Manpreet Randhawa is the Head of Digital Content for myHRfuture.com. In her previous role as the Change Management Lead for People Planning, Design & Analytics at Cisco Systems, she was responsible for defining and executing on the change management strategy to successfully implement and sustain the digital and cultural transformation across the enterprise. Manpreet is very passionate about change management and technology and how to use both to transform the employee experience and prepare companies for the Future of Work.