Six steps to double the business value of HR data and metrics: part 2

wpid-data.jpg

In the second of our two-part series exploring HR data and metrics, Nick Kemsley from the Centre for HR Excellence at Henley Business School looks at the final three of six practical steps any HR function can take to double the business value of the information they are providing to their organisation.

As CEOs and boards push to squeeze maximum value out of their organisations, the need for insight to guide their decision making and to manage both short- and long-term risk has never been greater. HR functions need to focus on providing insight, not just information, to align metrics to business goals and to examine outcomes at least as much as process metrics. The first part of this article looked at the first three of six practical steps HR needs to take in order to overhaul the business value of their approach to HR data and metrics:




  • Identify the insight needed to underpin strategy delivery and manage organisational risk.
  • Understand the decision-making data your board needs to support shorter-term performance.
  • Determine the critical gaps between this and what you currently provide.

This article looks in detail at the final three steps towards doubling the business value of HR data and metrics, which are:



  • Find the most pragmatic way to plug the gaps.
  • Optimise the format in which the data is presented.
  • Develop the right skills to analyse and talk about the data.

Find the most pragmatic way to plug the gaps

The question to ask: “What is the easiest way to generate the missing insight required, starting with what we already have?”

In step three, we said that we should understand the nature of the gap. Is it about the data itself or what we do with the data? If it’s about the data itself, the first resort should always be to look at what data capability you already have. Do you have the data but are simply not presenting it? For example, data on leavers by grade. Can you create the data from other data you have? For example, time spent on training per head. Or do you have to get the data through new measurement or from another source, for example, by benchmarking with other organisations?

If it is about what you do with the data, then the first resort should always be to look at how insight can be developed simply, as opposed to through effort-intensive data analysis. This might involve finding trends with existing data to create insights and project forwards, correlating one set of data against another or engaging with data earlier – for example the difference between strategic workforce planning and simple resource planning. Only when we have exhausted all the easy stuff, should we look to system change to plug the gaps.

For example, a multi-national organisation has identified that a major shift from volume transaction processing to higher margin service provision is vital to the future of the business. Twelve months earlier, a number of “strategic projects” were identified. One year later, however, there seems to be little progress on the strategic shift which the business was trying to drive. This is being discussed in an HR meeting and there is a view that resources may not be being properly allocated to these projects in reality. A member of the HR information system team then proposes that, since HR has data relating to roles and project time booking, she can write a quick algorithm that can, in a rough way, categorise individuals into three “buckets” (core business, business support and growth projects), dependent upon where they spend most of their time. This does not require any change to existing data. The algorithm is run later that day across a two-year period, and the day after that the HR business partners for each business area quickly check the results are accurate and amend where necessary. The results are then plotted on a graph.

The HR director takes just this single slide to the next board meeting. What it clearly shows is that, despite there being an enormous amount of talk about priorities, at grass-roots level people are not being properly allocated to support growth projects. In fact, it shows that the number of people working in business support functions like finance has actually grown, while those working on business-critical endeavours has decreased. This is a fantastic insight for the CEO. He asks that HR lead a piece of work with his board members on re-allocating staff by prioritising work and improving efficiency, and that they report on progress as the number-one agenda item on each board meeting going forward. In an all-employee telephone conference later that month, the CEO publicly thanks HR for the insight they provided.

Optimise the format in which the data is presented

The question to ask: “Who is the customer for the data insight and what is the best way of giving it to them?”

The first resort should be to put yourselves in the shoes of the data’s customers and consider some of the following:




  • What data is presented and what is not? The temptation is to present lots of data so that we are seen to have “all this” at our disposal. Better to filter what we present to suit the business need. Give yourself the challenge “if it doesn’t align to a strategic risk or an important business performance issue, it doesn’t go in”.
  • What are the priority insights? What point are we trying to make with the data? The way that it is formatted should make the point clearly. Choose graphs or tables that serve to make your point evident.
  • How is insight developed? Does insight come from the report or from the discussion around it? As a general rule, you should create the insight before it is placed in front of your customers, not ask them to make the connections. A good approach is to have an HR meeting prior to any “pack” being submitted to ensure that it is aligned to priorities, and to get into some of the texture and detail behind the data and draw out any insights. Some degree of “narrative” is often appreciated that takes a stab at explaining why something may be happening. Any pack should stand alone as a source of insight, not be dependent upon a presenter for the insight to be gained.
  • What is the appropriate level of detail? How is the data “layered”? Key insights should be pulled to the front of any packs or reports, and clear linkages made to the business issues they relate to. Supporting graphs or data should be next. Other “data of interest” can then be included at the rear. Give yourself the challenge “if they only read the front page, what do we want them to take away”?
  • What claims will the data support? If you are making claims using data, think hard about what you can and cannot justifiably say with it. Don’t be afraid of using phrases like “indicates” or “suggests”. Be aware of how sensitive the findings are to margins of error in data accuracy.
  • What is the right format and regularity? Is it a report or pack? What is the right balance of graphics and words? If the insight is to be gained by comparing one set of data with a second, plot them on the same graph. Make pictures do most of the work. What regularity suits which data? What can be reported on a quarterly basis, what monthly, what is in every pack and what are one-offs? If there is a major risk uncovered by the data, don’t be ashamed of taking more space in the discussion/pack to go into depth on it at the expense of other data items.

For example, a UK-based services business was recently told by its CEO to review the metrics pack that it presented to the board each month. In a team meeting arranged to discuss the data pack, the HR team was asked to tick off the elements of the data pack that related directly to topics in last month’s board meeting minutes. There was nothing that related directly. Instead, there was a lot of data on diversity. The question was then asked: “How much has gender balance in the business changed in the past 12 months?” The answer was by less than 1%. “So why do we present it monthly?” was the next question.

Another example is a global business embarking on a push with respect to diversity that included the following finding from a company-wide employee survey in their people pack: “The number of gay, lesbian, bisexual and transgender employees in the organisation has doubled in the last three years.” Has it? Given that this is a discretionary disclosure, the only thing this actually shows is that the number of people who declared this preference in the survey has increased.

a third example is of an HR function that made a big play about the organisation getting tougher on performance when in fact they had recently changed the performance rating structure, which clouded any conclusions which could be made at that stage.

Develop the right skills to analyse and talk about the data

The question to ask: “Do we have the skills to analyse and talk about data in the right way, and if so, how can they best be harnessed?”

There are two considerations here:

1. Skills to gain insight from data.

2. Skills in presenting insight from data.

Typically, to gain insight from data, you need a combination of micro- and macro-data skills. What I mean by this is both the ability to delve into the detail of data and understand its strengths and weaknesses (micro-data skills), and the ability to see patterns in data, work with imprecise data and scenarios and see connections and correlations (macro-data skills).

These are often not found in the same person, since what may make you good at one may make you bad at the other. The right question to ask therefore is: “Where do these skills exist in my organisation and how do I best bring them to bear?”

For example, an organisation was struggling to get strategic workforce planning up and running. They were finding that the individuals responsible for recruitment could not engage the business in the right kind of conversation, and could not think beyond a spreadsheet full of numbers as an outcome. As a result, discussions about changing workforce needs were not being had early enough to do something about them. They realised that, in addition to skills in planning operational recruitment activity, they needed capability in scenario planning and judgement in interpreting ambiguous data. They lacked these macro-data skills in the HR function, but realised that the wider business had them in many areas. They therefore gave responsibility for strategic workforce planning to a mixed, cross-business group of different disciplines and people, from resourcing, talent, finance, and business planning. This allowed them to pool the different and often conflicting skillsets in service of the creation of a strategic workforce plan.

When it comes to presenting data, HR can often shoot itself in the foot by struggling to provide context for the findings, comment plausibly on the statistical relevance of the data or talk about it in a way that is appropriate for the argument being made. Considerations here include a grounding in basic statistics, an understanding of how data is collated and analysed and a solid view of what the data will and will not support in terms of an argument.

Summary

What I have tried to do here is to map out a checklist for examining and pragmatically overhauling your approach to data and metrics. These are not so much solutions, because one size does not fit all, but a sensible set of questions and considerations which, if applied, will develop insight and present opportunities for an HR function to add real business insight through data and therefore build its own credibility.

Nick Kemsley is co-director of the Henley Centre for HR Excellence








Using data and statistics more effectively


This article is the latest in a series on using data and statistics more effectively in HR:


Comments are closed.