For most ‘business facing’ roles inside big companies, data and the understanding and manipulating of it is becoming more important. It seems to be everywhere (data is the new oil) and everyone is using it to demonstrate their point. Often though you don’t need to dive too deep. I tend to like an evidence base (tell me how much, how many, how often etc.) but increasingly the process of getting to that evidence seems fraught. When clients ask their teams ‘how do we know that?’ or ‘can you compare what happened last year?’ they tend to be met with lots of blank looks and sucking of teeth (a bit like taking your car to be repaired – “not before Christmas mate” and “this is going to be an expensive job Mister”).
When discussing these issues around data with my clients the stories they tell are the same. ‘We find it hard to get our data’. ‘It’s always broken, inaccurate or incomplete’. I recently ran a series of workshops for a global business to explore what the demands for information were across the group. A telling case was in the first workshop (held predominantly with a global IT group). The general consensus was their organisation data was in pretty good shape, especially people data and especially that managed by IT (which often tends to be Active Directory – the de facto standard for enterprises to manage user’s IT profiles). However as we went on to run over 20 further workshops across the global business, people from heads of business units to PAs all said the same: ‘our data is rubbish’. ‘It’s incomplete, we don’t even know who our people are’.
The changing face of data
We’ve probably all known the pain of bringing data together into one coherent, reliable store (there’s a whole discipline around it – Enterprise Data Management). In HR this issue is often well recognised – but postponed. To be addressed with the shiny new HR Management System (HRMS) that’s going to be implemented (one day). Often these large HRMS programmes do lead to a measurable improvement; I am seeing a lot of workday implementations currently and the promise a big systems transformation brings is welcome.
And this is a big but – for every successful programme I see there are many challenging, messy and abandoned paths on the journey. An HRMS programme of work is extensive and touches everyone. Source systems have to be identified, understood, retired or committed to. A new ‘data schema’ is needed and it takes teams of people with deep technical skills to glue it all together. And then there always seems to be the parts that don’t work quite right. It is a recurrent theme I have witnessed over 20+ years.
There is a new, emerging approach though. You can sense it in a few of the things happening around you. They can be loosely grouped around having an impact on the way users (you & I) access, handle and get value from data. Until recently the only widespread route to doing anything you wanted to with data was in Excel, and that’s a pretty messy place for people data. Over the past few years though we’ve seen the emergence of data visualisation tools – they make data a little easier and a little more fun (I love Tableau) and we’ve seen a huge shift to more user friendly (yet still sophisticated) tools through the rise of the iPhone, iPad and supporting applications.
These tools put data into the hands of casual interest business users, helping you to make sense of it yourself and present insight back to others to make a persuasive case. I call this ‘human scale data’ as opposed to the big data trend in machine data. I know data isn’t everyone’s bag and it is especially tricky to handle HR data given it is often so broken and hard to source (especially in comparison with Finance data). I’d maybe even go so far as to say one of the reasons HR struggles for a seat at the board next to Finance is the challenge of data. Perhaps these new approaches may offer a useful and far less effortful route to insight.
So with all this in mind, what is a data schema and how does being ‘schema less’ make such a vast difference – a whole world of difference?
Here’s Wikipedia’s definition of a schema (warning: if you are not a database professional, may cause dizziness). My simplistic view from a business management background, is they are a representation of how data is stored. They are detailed; every piece of data requires defining for a database to work, is this piece of data a number (what type), a date (what format), an amount, a title etc.. It defines how data is related to each other; for instance this surname links to this payroll ID, links to this banking detail. Schemas are not trivial and large organisations will have dozens or hundreds of people who need to understand them. [As an aside around HR data – one of the specific requirements of HR is the ability for a database to handle dates in the past, and in the future when another piece of data becomes valid. For instance this person is joining on date xx, this person was promoted on date xx. This demand alone places additional complexity in HR databases.]
It is almost certainly true that every database you interact with in your professional life has a schema defined somewhere in order to work as it should. This gives you and the systems people that develop them the confidence that the number that comes out at the end is the right one. It also stops you (or anyone else) being silly with the inputs. The downside is – it’s complicated. Not only is it complicated, it is inflexible, you can’t easily connect to it and even when you can it’s hard to understand. There’s technical and data engineering magic woven into every piece and the more you tend to do with your data the more complex it becomes.
This means you can’t use it.
Here’s a quick test for a global HRD:
- How many people work for you?
- How many people work for the whole organisation?
- How many are female?
- How many are short term contractors (and what determines if they are?)
- Of those that are female and permanent how many are within 2 levels of a senior management position?
- What is the average salary of them vs. their male counterparts?
- What’s their performance record over time?
- How do they think their manager is performing?
- How quickly and with what resources could you answer the questions above?
You probably have systems all around your organisation that hold the data that would let you answer those questions – but every one of those systems operates within a fixed data schema. That’s wonderful to know (you can rely on what they tell you), but it’s also a nightmare when you want to bring them all together to answer a question being posed about the gender distribution of your senior management and how you are planning for it. Because the schema is fixed you have to understand it for every system in order to access the data. That makes it a massive technical job and yes, the big new shiny HRMS will give you it, but what happens when the FD or the CEO (or a compliance head) asks – ‘what would it cost to relocate, re-structure and promote?’ and you need to include the data from finance?
Is it possible for you to simplify this landscape dramatically?
The Importance of being Schema-less*
Can you get the answers now, without waiting for a long lag-time, can you and your HR team ask the questions yourselves, get hold of the data and answer the questions immediately?
Of course you know what I’m going to say. The answer is yes – if you use a schema-less database.
OrgVue is a schema-less database.
What does that mean?
It doesn’t have a fixed representation of the data it holds.
It means you can cut and paste into it. If you have the email address or the payroll ID or any other way to uniquely identify an employee, start there. Cut and paste them in. 2 people or 100,000 employees – doesn’t matter. Then add in their reporting line. If they exist in the data and their manager (reporting line) exists in the data, OrgVue will find them and you’ve now got an 10,000 person org chart – in 2 mins. (You’ll also have information about reporting line depth and span of control).
Now paste over the top (literally, copy from a spreadsheet, paste into OrgVue) any other data you can get – performance rating, tenure, recruitment sources, costs, management info etc. etc. Anything you have or anything you can source: that old employee engagement survey; the new proof-of-concept Performance Management tool data; the leadership project. Literally any data you have, even messy, broken, incomplete data. If there’s an ID to join it up with OrgVue will let you.
You’ve not needed IT to get involved and yet you can now see your organisation. The first thing you will see is broken data. Because OrgVue is schema-less it doesn’t reject broken or missing data, just shows you where it is broken and allows you to fix it easily (1,000 records with a misspelt location or wrong business unit code? – fix it in seconds). You can even export out the fixed data.
Your whole organisation is now navigable. 10,000 people on a single screen (that’ll show you something you didn’t know), or zoom in to every individual, see their performance, their photo, their costs – their data, any data.
The difficult business environment we all now live in is constrained on the one hand by underlying systems and on the other facing the ever increasing demand to know more, faster. In the middle of the large slow-moving systems transformation programmes to fix it (if you are lucky), it is a breath of fresh air to be able to just get on with it.
You’ll find out things you never knew, or more importantly your board didn’t know. In the process you’ll be fixing your data and helping to simplify IT, a desire of everyone I meet.
For more on OrgVue, see Andrew Marritt’s commentary.
*(with apologies to Oscar Wilde)