Human Capital Management Analytics

12Oct/110

Today if your hear the name Apple, Tata Motors, 5 Guys or Pixar your first thought is “I just don’t like what they do. I Love what they do.” You; laughed during Cars, were awe inspired by the iPod. drove 15 miles to get that burger, and were fascinated by a profitable $2000 mass-produce car. You also didn’t just experience them once… Cars 2, iPad, second third fourth burger, Range Rover. They have brought and defined Innovation for our generation and possible the next several.

Does this mean you have to be innovative to be like Apple or 5 Guys or to compete against them? Absolutely. Just ask Palm, GM, In n Out Burgers or Disney Studios. In fact, based on a survey from the Conference Board, innovation is the #4 priority for CEOs in 2011. Innovation impacts your business in more ways than you can possibly imagine. Bill Gates once said “Microsoft’s top 10% is impactful, but if I lose my 20 most innovative software engineers Microsoft will go out of business within 12 months.”

You can’t quantify innovation by looking at a person or giving them a test. Innovation sometimes only happens when the right-mix of people get together in a team setting. So how do you identify your “innovators”? Does your company have enough of them? Does the innovation increase profitability?

You can determine who the innovators are by determining the following;

How much does a person or team produce? - Does this team or person deliver on what ever they are given or is it hit and miss?
Did they produce a new thing? – Is it a completely never before seen product, drug or process? Was it the next generation of something? Was it something that was a copy of another product?
What impact did it have on the market? – Did what they produce change the market place? Did it displace a competitor? Did it define a new category of products? Did it make the company better? Do analysts, customers or the press say what was produced was innovative?
What was the profit contribution? – How much profit did it bring the company? Did it raise the stock price? Innovation is important to success. But it shouldn’t cost so much that you never recoup the costs.
Do they repeat? – Anybody can get lucky once. But delivering innovative things 2, 3 or 4 times in a timely manner means the producers are Innovators.

With this you can then identify and put policies in place to further foster your innovators or have a business case to hire them.

A lack of innovators within a company may not bankrupt it. But as we have seen time and time again a company with the right amount of innovators will make more profit, have a more engaged workforce and take the customers of those that don't.

Share on Facebook

Filed under: Analytic Programs, Business Case, HCM Analytics Example, HCM Analytics Roadmap, Workforce Engagement No Comments

22Sep/110

The Future of HCM Analytics-iHRIM Article

Posted by Brian Gaspar

Had an article published in this issue of iHRIM October-November magazine on the Workforce Predictions and it will be handed out at HRTech in two weeks.

iHRIM article - The Future of HCM Analaytics

The world of analytics has shifted emphasis over the past 2 decades - from looking at the past to looking at the future. Organizations who have made this shift, realized the basic fact that the billions and billions of pieces of information being produced every single day can be used to answer, fairly accurately, what will probably occur at some given point in the future.

Although predicting the future sounds mystical and straight out of science fiction (like in movies like The Minority Report and Paycheck where predicting the future for all facets of life is the norm), the reality is that we’re already doing this today and you are being impacted by these predictions in almost every aspect of your own day to day existence – it’s called predictive analytics.

Examples from everyday life:

Supermarkets know more about people who use their frequent shopper cards than do their friends on Facebook.

http://www.buseco.monash.edu.au/centres/acrs/research/whitepapers/hidden-side-of-loyalty.pdf

Credit rating companies such as FICO or Experian gather every person’s financial history in order to predict how credit worthy that person is – every single day. This single predicted score (e.g. 800 = very low risk or 600 = very high risk) is provided to lenders in order to determine how much money to lend, to employers to determine how responsible a person is, or to the owner of the credit score who will be provided with recommendations on how to improve their score. http://en.wikipedia.org/wiki/Credit_score
Major League Baseball captures every move made by every player during every game. It then crunches this information to determine a player’s and team’s weaknesses, strengths and performance in given play-situations and based on this ‘predicted performance’, coaches determine who their starter should be against a particular team and even how much players should be paid. http://www.bloomberg.com/news/2011-03-31/baseball-is-set-for-deluge-in-data-as-monitoring-of-players-goes-hi-tech.html
In 6 hours the NSA collects enough digital information on people, which if printed, can fill the Library of Congress, and these data mining and predictive capabilities were instrumental in finding and tracking Osama Bin Laden. http://www.popsci.com/technology/article/2011-05/every-six-hours-nsa-gathers-much-data-stored-entire-library-congress
Companies are starting to mine millions of tweets and stock transactions a day in order to predict if their stocks will perform better or worse depending upon the general public mood.

http://www.hpcwire.com/hpcwire/2011-05-11/a_global_mood_ring_for_financial_markets.html

There are a few common threads across these examples. First, all of them are using historical information (1 second past is historical) to make predictions. Second, the predictions being made are about people’s behaviors. More specifically predictions on how they will perform, what they will buy, how they could be improved, who they are connected to and what is their mood. Third, they are looking at information that comes from a few to several thousands of sources. Fourth, decision makers can determine the best course of action to minimize challenges or exploit opportunities and fifth, organizations are spending billions of dollars annually on making predictions about people and it has proven effective. With HR being all about…well people…this would seem to be a natural fit.

HR is currently delivering information to managers and executives on items that they already know about like, what their headcount is, what it is forecasted to be by quarter-end or year-end, what a worker’s past performance rating was, when they will be going on vacation and even if they are taking too much vacation. While these items are important to document, it does not tell the business anything new and is primarily geared towards delivering on the next phase of Human Resource “back office” automation.

In order for HR to truly deliver on the promise of being “business partners”, it needs to tell the business scenarios’ that are predicted to happen in the future, across all levels of the organization, which they do not already know about and to provide recommendations on how to fix or exploit it. Analytics, like headcount reports, need to shift from a mere ‘what they already know’ to one which will not only a) tell an organization that 5 additional people need to be hired because 5 specific individuals are expected to join the competitors but also b) the reasons they will leave e.g., the competitors have several positions that match our workers skills, that they are paying more since inflation is on the rise, that a pay increase has been given in the past 24 months, or that the competitors are beginning to out innovative our organization.

With Human Resource organizations having difficulty delivering fully robust and accurate headcount reports, this may seem like something that should be explored at some distant future and that’s the wrong approach. HR struggles to get basic funding for important analytics initiatives and a 2009 Forrester Report showed that companies spend less than 1% of their business intelligence budget on HR. With HR making up, on average, 31% of a company’s total operating costs, the assumption is that this should be one of the top 3 funding areas. The reality though is that all other functional areas have been able to quantify a business case by showing that their analytics initiatives lead to actions the business must take and/or didn’t know about before, such as increasing sales, competitors stealing customers, reducing spending, improving productivity, or even to meet federal laws. The current model for today’s HCM Analytics just does not provide that “business driver” to spend more on it and this “business driver”, as discovered by other functional areas, would be to tell the organization what it does not already know, why it is happening and how to fix or exploit it. All of this is predictive analytics and without it HR will continue to struggle to obtain the funding and support for their must-have analytics initiatives.

HR can do predictive analytics today by partnering with organizations (e.g. sales, marketing and engineering) that have predictive analytics already in place to make quantifiable and actionable predictions about the workforce in areas that the business does not know about. Several companies have done this and Oracle for example, was able to build a business case using this approach. In 3 weeks Oracle was able to predict which top performers were predicted to leave the organization and why - this information is now driving global policy changes in how to retain key performers and has provided the approved business case to expand the scope to predicting worker performance.

HR is at a cross roads where analytics is concerned. The current approach simply does not work which is proven by the fact that companies spend less than 1% on a problem that costs them 31%. In order for HR to be successful in analytics it must embrace the shift towards predictive analytics. Without making the shift, HR will not be able to deliver on the promise of being a “Business Partner”.

Share on Facebook

Filed under: Business Case, HCM Analytics Example, Workforce Predictions No Comments

18Jul/11Off

Today’s “In Memory” HCM Apps; Functionality of 10 Years Past and Gone (again) in 5 years

Posted by Brian Gaspar

I was recently at a company presenting on HCM Analytics. Also in the room was a less-than-independent consultant who was working with the company helping them with their HCM Application Roadmap. The consultant brought up two points, actually their only comment in the meeting, which I found rather interesting and got me thinking about how old school “In Memory” HCM Apps are. The comments were:

Relational Databases have challenges
“In Memory” is a better architecture than a Relational Database

I was stunned by the first comment and snickered about the second (quietly in my head that is) - let me explain why.

Relational Databases have Challenges

Why I was stunned by this comment: The consultant just alienated themselves and lost all credibility with their client, who incidentally is the founding fathers of the Relational Database (something the consultant obviously did not know about them as they said it a few times). The rule #1 for any consultant is that you never ever tell a client, repeatedly, that what they have or are building is terrible, unless they are paying you too. If you are not paid to do that, you don’t say anything about it. The consultant lost even more credibility, if that was even possible, when they were not able to tell us what any of these “challenges” were.

“In Memory” is a Better Architecture than Relational Databases

Honestly I was a bit confused by this statement since an “In Memory” architecture is not exactly the same thing as a Relational Database. I am just going to assume the consultant meant that the storing of data for thousands of workers in 3 skinny tables of simple name value pairs completely “in memory” is somehow better than storing it in a Relational Database on Hard Disk Drives.

Since there is a lot going on with that seemingly simple statement I am going to break this down into its three main components:

1. Skinny Tables are Better Than Relational Databases:

Relational Databases are designed and made primarily for 7 things; storing of data, defining relationships, flexibility, scalability, speed, triggering of events and reporting (aka Analytics). Skinny Denormalized Name Value Pair Tables were designed for only 1 thing, speed. I snickered at this comment by the consultant, because, Relational Databases were built to and did replace this type of design almost 30 years ago, by the client we were both talking to. Let’s talk about why it replaced it:

Speed: In this case, speed is how fast you can retrieve a piece of data from a table. If you do not optimize a Relational Database for speed (e.g. indexes, cache important data “in memory”, pre-calculated values, optimize your queries, etc.), out of the box, a Skinny Table will be faster. If you do optimize the relational database (which every single vendor and company does), it’ll both be on par for the retrieval of data and significantly faster for things like Reporting and Analytics.

The Rest: To be blunt, companies will make absolutely no forward progress in delivering any new business capabilities using this architecture and will only be able to deliver on what was available 10 years ago. It simply is not architected to support what the business expects from today’s HCM Applications. For example, you can and should forget about ever having:

a fully expanded organizational chart (does not support b trees)
ad hoc reports or ad hoc anything (missing relational & flexibility)
multi-dimensional reports like for example, showing a trend of headcount and headcount costs for the past 3 years for 5 departments in 3 different locations graphically and in a table. (missing relational)
visual drag & drop organizational modeling (missing relational, flexibility & support of b trees)
Predictive Analytics (missing relational & statistical algorithms)
dashboards and embedded analytics usable and configurable by end users (missing relational & flexibility)
dynamic business processes such as OnBoarding or international transfers (missing relational & triggering of events natively)

This means you get a whole bunch of data entry pages blended with dozens of static reports, which you can’t personalize or create new ones when you want to. If a vendor does have any of the above, they are either doing it in a Relational Database or have built highly proprietary tools. If they are doing it with highly proprietary tools, then I wish you luck. Because you are definitely going to need it in when you have to explain why your TCO jumped through the roof, learn it, find skill sets for, integrate with other applications and for the ongoing management of it.

2. Storing Data “In Memory” is Better then Storing in a RDMBS on a Hard Disk Drives

The consultant implied that this “In Memory” architecture was a brand new concept. The reality is that all companies and vendors have been “caching” important data on a web server for more than a decade. While the name “In Memory” is cooler than the name “cache” of 10 years back, the architectural principle is exactly the same i.e. it eliminates the step of retrieving data from the Hard Drive (spinning up the disk, seeking the data and providing the return). Important data could be in the form of a document (e.g. Word, pdf etc) or data from a database.

However, instead of caching some data, these vendors have put all of their data “In Memory”. This eliminates some of the problems of caching some pieces of data and not catching other pieces of data, such as data integrity problems associated with the requirement of having two copies of the data (1 in cache & 1 on the disk drive). However, this created a whole set of completely new problems and limitations. The most notable ones are that; you’ll have lots of application down time and can forget about adding a new field or creating a report on the fly. If you want to add a brand new field (e.g. new field to required for a new legislation report) in a form or create an ad hoc report using Skinny Tables stored “In Memory”, you have to first bring down the environment, then rebuild the skinny tables (most likely from a relational database) and then redeploy them. This means you will have weekly multi hour down times. You don’t have any of these problems with Relational Databases on Hard Disk Drives.

3. “In Memory” is faster than Hard Disk Drives?

Hard Disk Drives are going the way of the 3.5 inch floppy drive. Solid State Storage is the storage medium for most consumer oriented vendors (e.g. Apple’s Air, iPad and iPhone). Which means absolutely everything is stored in memory (e.g. operating system, applications, and even a relational database, etc). Solid State Storage is faster, smaller and consumes less energy than Hard Drives. Think about how much slimmer an Apple Air is versus your corporate Thinkpad. Then think about much faster your iPad brings up pages versus how fast your PC brings up Word. Several vendors are already selling Servers and Mass Storage appliances built with Solid State Storage, not Hard Disk Drives. As memory prices continue to fall, servers with Solid State Storage will be as common as they are in the consumer world.

HCM Application Vendors using “In Memory” architectures are geared for speed by skipping the step of going against the Hard Disk Drive. That is pretty much the only benefit. The reality is that this is a rebranding of an already well established and used by everybody design. The storing of all of your employee records “in memory” (aka caching) will never be able to do what a Relational Database can do and when the speed difference disappears in 3 - 5 years with the replacement of Hard Disk Drives with Solid State Storage, so will this design. Why then would you invest time and money on something that was replaced once by Relational Databases, can only deliver on functionality which existed 10 years ago and will be replaced (yet again) by something faster in 3 to 5 years?

Share on Facebook

Filed under: Business Case, HCM Architecture, Workforce Optimization Comments Off

17Jun/110

Future of HCM Analytics – Workforce Predictions (with poll)

Posted by Brian Gaspar

I gave a webinar on hr.com this week discussing the Future of HCM Analytics...with a bit of Oracle Fusion HCM Workforce Prediction demo (slides here). I was excited about this webinar, because, I got to talk about future state topics, showed a demo, tried out some new material and was the most excited about being able to give an online poll. The poll's questions focused in on what companies were spending on HCM Analytics.

Webinar polls are usually notoriously hard to do. You are not in a room with the attendees. This means you can't easily guilt or bribe attendees into giving answers to your polls. Plus the minor fact that everybody is on mute!!! So they can't verbally respond even if they wanted to. However with GoToMeeting by Citrix (very impressed with their "Poll" feature) and a few electronic Starbucks cards I was able to achieve a 75% response rate for the 350 attendees.

Poll Question #1: Does your company have 1 or more workers dedicated to HCM Analytics?

Yes: 49% Respondents have 1 or more workers dedicated to HCM Analytics.
No: 51% Respondents have less than 1 worker dedicated to HCM Analytics.

Poll Question #2: Does your company spend more than 1% of its Business Intelligence budget on HCM Analytics?

Yes: 1% of respondents say their company spends more than 1% of the BI budget on HCM Analytics.
No: 38% of respondents say their company spends less than 1% of the BI budget on HCM Analytics.
Don't Know: 51% of respondents say they do not how much their company spends on HCM Analytics.

The numbers are certainly a bit bleak for Human Resources in terms of the % of Analytics budget it actually gets. But I wasn't surprised. I see adoption numbers of HCM Analytics all the time and this is very much aligned with those. However the main take away from this should not be on the lack of budget or people to support Human Resource Analytics initiatives. The take away should be that a company's leadership does not get what it needs from Human Resources with their traditional HCM Analytics offerings. Leaders spend money on what it needs and what has a Return on Investment. Traditional HCM Analytics simply does not provide an ROI. The future of HCM Analytics is Workforce Prediction and it's already proven to have an ROI.

Share on Facebook

Filed under: Analytic Programs, Business Case, Conference, Events, Workforce Optimization, Workforce Predictions No Comments

9Jun/110

Workforce Planning – Overcoming the Barrier to Adoption

Posted by Brian Gaspar

I am dating myself here, but for the past 18 years I have been a user, builder, practitioner, consultant and teacher of Workforce Planning. Over these almost 2 decades I have learned that the primary barrier to adoption has been that Human Resources attempts to replace an already existing, core and engrained Workforce Planning solution with their own. It’s not that Human Resources does this intentionally or think they can offer a better solution and overlooks the current one. They just do not have deep enough domain in the various Lines of Business to recognize that there are supporting structures and processes in place that the Lines of Business use to plan work for their employees. Examples of these already existing Workforce Planning processes are:

Sales has territory management, territory optimization, etc.
Engineering has several processes to choose from such as Waterfall, Agile, etc.
Manufacturing, just like Engineering, also has a few to choose from like “Just In Time” Manufacturing.
Call Centers have commonly adopted Workforce Optimization as their solution.
Corporations can also adopt companywide processes, such as Six Sigma or Total Quality Management.

Overcoming the Adoption Barrier
It’s natural that the Lines of Business would and will continue to reject HR’s own solution to Workforce Planning. They will never believe nor could Human Resources show that they have a greater domain than they do in their own business. However they do need HR’s help to make it dramatically more effective than it is today. So how does HR overcome this barrier?

Understand HR’s Role: Human Resources should absolutely play a role in Workforce Planning. But that role should not be the lead role. The Lines of Business know their business best, is already doing workforce planning and Human Resources should work with them to determine how they can support an already existing process. Not telling them how to do Workforce Planning for their business.
New Solutions not Needed: There are lots of Workforce Planning vendors, consultants and methodology out there. The first place to look for providing your Lines of Businesses with a solution for Workforce Planning is not here. For the most part, the Lines of Businesses already have some solution they are using and need to figure out how to embed already existing Human Resources solutions into their processes. For example, yearly Performance reviews do not work effectively for customer facing Professional Services teams. Performance Reviews should be done either at the end of every engagement or for extremely long engagement, at the end of every milestone. Either way, Professional Service teams want to fit in Performance Reviews in order to determine who should role onto the next project and who should get some additional training.

If Human Resources follow the above, not lead and not replace, they will be able to successfully partner with the business and will have a much easier time getting subsequent initiatives adopted. Best of luck….

Share on Facebook