Using Big Data to improve learning, even with the Small Data we have today

Although the possibilities with Big Data for learning seem enormous, the probability of the everyday learning person achieving the Big Data dream seems small. But the concept nonetheless pushes us to rethink our traditional approach to learning.

Big Data is big and it’s mostly about your online data

Given the adoption of social-media tools around the world, the amount of data about people, their preferences and online behaviour never ceases to increase. IBM’s Chief Executive says “there will be 5,200 gigabytes of data for every human on the planet by 2020.” What we publish, like and share online helps marketeers, social scientists and others to personalise their offers (and sell you more stuff).

DAN STEER BigData Image1

Following the same principle, we could better tailor learning content to our people

Many ideas around Big Data for learning are centred on capturing data points concerning how formal learning happens, in order to improve that process. At its most banal level, one imagines capturing massive amounts of “happy-sheet” data to see what learners prefer and tailor formal learning to increase that satisfaction. (Example: How long people spend on modules). A better use of Big Data might seek out information on how learners actually perform in formal training and e-learning (example: test results) in order to improve the learning activities therein. Better still would be using all sorts of data captured across the organisation to personalise and adapt learning content to the preferences and interest points of the learner.

DAN STEER BigData Image2

But it wouldn’t be easy to achieve

Although some organisations have started looking into Big Data, most do not have anything in place to capture the potential amount of data available. The worst learning organisations have no data, the average has hearsay and some Excel spreadsheets, and some others have an LMS with (only) completion rates, test scores and some other pointless numbers. Are we really able to capture Big Data today?

And Big Data for learning is not really about learning anyway!?

Let’s face it: Only learning people care about (formal) learning content + processes anyway. Some people might care about developing competence, which is already better. But what we really want is performance improvement. The focus on Big Data in the learning world must not be on getting preference-based data out of formal learning activities, but putting more useful performance and behaviour data from anywhere into learning and performance strategy.

When I read about how we will be able to see what happens during learning, I feel like we missed the Big Data point. What is interesting is using data about behaviour and performance across the whole organisation to see what is going on, what is going wrong and (possibly) what needs to be learnt. It is not about using data about what happened in learning initiatives to make more learning initiatives. Roughly put: Big Data should be focused on Kirkpatrick’s 3rd and 4th levels of behaviour + performance results, not learning satisfaction + acquisition.

DAN STEER BigData Image3

Imagine an example where we capture data on customer-handling behaviour and quality in a sales call-centre in order to provide prescriptions for better repetition of keywords, objection handling or other desired behaviours. Or the example of knowing that employees who use particular search terms in Google acquired more relevant knowledge and are more likely to find solutions for their clients. Or that projects that encounter problems X, Y or Z in a given phase are less likely to finish on time. This kind of data could provide valuable input into performance objective setting, desired behaviour, learning design and learning deployment. Or all sorts of things that are not even related to learning needs.

Suppose we have the right motivation and the data. Are we ready to use it?

When the motivation, budget and systems are in place to capture large amounts of potentially more interesting behaviour and performance data, we will be faced with the new problem of analysing it all. When the machines do a good job of presenting the data and showing trends and exceptions, we will need a massively improved skill-set in statistical analysis and forming hypotheses about what it all means. If the data is the ingredients of the “better-learning processes cake” one must not imagine that we will simply put it all in the oven and wait for the oven to cook up a solution. The (scientific) ability of the learning professional will have to improve in order to analyse all the new data, suggest reasons for existence of certain observables, question and form hypotheses and propose learning initiatives to test, measure and constantly further refine learning initiatives. If indeed learning is the issue.

And that work can already be started with the small data packets we have.

Many learning people are still in what Donald H Taylor calls “the training ghetto”, removed from the real action and lacking the business acumen to consult with and support the organisation. But there is plenty of behaviour and performance data out there that learning professionals can already analyse and use to support the organisation in the change of processes and environment, as well as the proposal of effective learning initiatives. It’s time to drop measuring training programme satisfaction and inventing formal learning cures for problems that are badly defined or simply don’t exist. It’s time for scientific learning consultancy: Help your sales people to form and test better hypotheses on what gets actually gets more revenue. Help project managers to share project-evaluation data on which processes, people and risks cause failure and focus on supporting that. Find out why quality is low in production and focus on supporting that. You will see that they have plenty to say and need lots of help, even with the data they have today.

The dream of Big Data analysis is theoretically nice and if done well could really help to create a more scientific approach to performance improvement. But there is enough small behaviour and performance data out there already to keep the professional learning busy. Let’s worry about capturing other data later.

More?

68721-250X350

Donald H. Taylor is een van de keynote sprekers op het internationaal congres Next Learning op 22 april 2015. Hij spreekt er over “Stop your Learning & Development Department right now. Start performing!”

Ook prof. dr. Viktor Mayer-Schönberger is keynote en deelt zijn expertise over “Big Data: next in Learning!?” 

Dan Steer, Kluwer-trainer, ‘learning geek’ en auteur van dit artikel zal aanwezig zijn op Next Learning 2015 en zijn ervaringen delen op deze blog en op twitter #NLE15

Kluwer Opleidingen is partner van Next Learning 2015. Daardoor geniet u als lezer van LearningLive een korting van 200 euro op de inschrijvingsprijs en kunt u gratis mee op de bus met de Kluwer-VOV-delegatie vanuit Antwerpen naar Den Bosch. Meer info vindt u hier.

 

 

 

 

 

 

Auteur

Dan Steer is freelance trainer bij Kluwer en learning & development consultant. Dan Steer is een Infinite Learning© kampioen, en gedreven door alles wat te maken heeft met SoMe, SoLearn en Enterprise 2.0.

Lees ook

Nieuws per domein

Meest gelezen

Let's connect