Sunday, 18 May 2014

Can HR Systems ever do Big Data? Workday?

'Big Data' is one of those IT buzzwords these days, with everyone jumping on the bandwagon.

Last week I was researching Custom Objects for our implementation, to see if there's anything new in the Workday world on this topic.  Custom Objects are Workday's 'customer-defined' fields, I previously looked into the Workday functionality last year.

On the 'Worker' data object, we're already up to 12 fields used out of 20, prior to starting our Europe roll-out.  Workday 'Big Data Analytics' customers, however, get 100 Custom Objects on the Worker!  This alone would be a reason to license the product, and I think that's why WD keeps the rest of us in check at 20.

What is 'Workday Big Data Analytics' (BDA)?

 BDA is a newer, separately licensed product that enables a customer to analyse Workday and non-Workday data together in one place.  You put non-Workday data into the 'Big Data Store' and use 'Big Data Explorer' as your interface to upkeep (i.e. load) your data and perform analysis.  (In some ways it reminds me of the promise of ERP in the late 1980's...your Financials and HR data, all in one system!)

Advantages?

As I'm curious, I've been looking at the WD presentations on this topic and reading through the forum posts by other customers. 
  • It sounds like it offers pre-defined templates and reports to help you to analyse large volumes of data more easily.
  • You can incorporate this data into your manager dashboards in WD, so a 'one stop shop' for a manager.


Can an HR System really deliver on 'Big Data' benefits?

When it comes to data in the HRIS world, I find a few main hurdles:
  1. Getting data
  2. Maintaining data
  3. Working with data so that it becomes a source of better decision making

I'm not sure how much Workday (or any other HRIS or HCM system) is really able to capitalize on Big Data, for a few reasons.
  1. Getting data and linking it across systems can often be a challenge.  Not all systems have a unique key structure that matches other systems.
  2. Other systems roll up data and define data in a different manner--it becomes an 'apples to oranges' exercise to compare it.
  3. Even when you are able to match data, it's often a one time exercise.
  4. HR data is not always maintained.  Someone may have a great idea to keep company property or driver licenses for company vehicle drivers in the HR system, but on-going maintenance can be an effort if systems and processes are not robust or meeting local needs.
  5. Even if you have the best data in the world, it can be a struggle to define a common set of reports or to figure out how to configure manager dashboards automatically at a highly detailed level so that you are providing targeted information at a manager's fingertips, on demand, rather than a 'general manager dashboard' which is the same for all managers.

When I think Big Data, this is how I think of it

Here in the UK, we have a large supermarket chain, Tesco.  They do some 24 hour supermarkets, some mega-stores that carry clothes and electronics, as well as having little stores tucked into train and bus stations and in neighborhoods.  Tesco appeals to a variety of customers as they offer foods from value through to upscale.

Most people (including myself) have a Tesco clubcard.  You can scan it when you pay and get discounts, rewards, and every other month, they send targetted coupons based on your spending habits.  Yes, it's very big brother, but no one is forcing you to take or use a card.

Tesco knows that I bought bananas at 6 AM on a Sunday, wine on Friday, and that I buy cat food from them.  Tesco has a super big database somewhere, and I suspect their analytics team is gigantic.

Big Data can change behaviour

A few rounds back, they sent me a cracking good coupon for 1.10 off cat food, so I used it.  In the last round of coupons, they sent me 75 p off of cat food.  I didn't use the coupon but bought my cat food elsewhere.  In this round of coupons, they've now raised the ante, and are offering me 95 p off.  Will it change my behaviour in cat food buying?  It might, as they're now making it more valuable to me. 

Big Data knows what I need before I know that I need it.

The bottom coupon absolutely tickled me!  I am a US person living in the UK.  Those very smart data crunching computers (or the smart IT wonks programming the logic) have analyzed the data and are thinking that they can woo me into their USA product range.  I can only assume that they put together the data facts, that I don't buy tea and milk, but instead coffee and a variety of other tidbits that lead them to this conclusion.

So why can't HCM, HRIS, HR Systems, etc. be this smart too?

I see a lot of hype from the HR data solutions that they're harnessing the power of Big Data, but I'm not sure that they're there yet. 
  1. I don't think that they have the data volumes necessary to fuel Big Data, even if you can combine data from other systems such as T&E or Sales.
  2. You'll never be in the 'daily, operational, transactional' level of data needed to get the deep data insights.  For example:
    1. An average HRIS may give you a report of terminations in the last fiscal year and an exit reason cited by the employee (where existing)
    2. Big data HR Systems may bring in data from other systems--what was the sales volume of the emps leaving?  Did you lose a top performer or underachiever?
    3. A Business Intelligence or Management Information team may use either of the above two sources to spot data trends...and then provide a further level of insight.
      1. An increase in terms at location X coincides with the HR policy changes to not match 401k and to reduce company contributions to benefits.
      2. Recommendation:  adjust the policy to influence the future term numbers.
      3. Check after six months, any impact?  Rinse/repeat.
However, we'll never get to that Big Data level of detail...without having the full data set, for example, knowing the compensation and benefits received at the new company, we're just making a one-sided, best guess, based on the data that we have.

Tesco doesn't know what I'm buying at the competing supermarkets...but it has an idea based on what I am buying and not buying at Tesco.  It knows that I have a cat based on all the cat related things that I buy, but I don't buy cat litter there (Her Excellecy prefers another brand that I buy elsewhere).  Those data gurus at Tesco periodically shoot me a fabulous deal on cat litter, knowing that it's a *missing* item on my list. 

We'll never get to that level of data in the HR systems world, regardless of how good the product may be at providing templates, reports and analysis tools and that is why we'll never do Big Data properly in an HRIS setting.

Of course, very happy to be proven incorrect on this point, be it on Workday on another HCM.  :)

2 comments:

  1. That is right, unless you cross HR data owned by the company with personal data collected on the internet.
    Most companies already do that through the recruitment process where they have the feeling they "know" the candidate (skills, personality, interests, motivations) but fail to keep these "people performance indicators" up to date after they have been recruited :)
    For example, an information like "how often you've been on linkedin last month compared to the past 6 months" might be an effective indicator to anticipate a departure.
    Although collecting such data is probably illegal, the scope of HR data owned by a company can be enriched just listering and keeping information from employees themselves during performance reviews or internal surveys. Something also we can't imagine now is the impact of Y generation sharing so much about their mood, interests, life and quite often work (twitter, facebook) and how a company could use this information to learn more about them and propose what motivate them best. Not only while recruiting them.

    However considering where we are today, what people related data a company have, and how clean it is, I agree there are not so much to do on Big Data for HR at the moment.

    ReplyDelete