Friday, 26 September 2014

Erring on the human side and 7 steps to avoid it

I've notched up another year on the age scale today so forgive me if I'm in a fairly reflective mood.  As usual my blog is less about a planned and insightful look into the world of learning technologies and more random stuff as a result of a trigger that occurred through my work and/or life.  Today's inspiration comes from a stuff up that I made yesterday, but it contains a lesson that I thought was particularly relevant in technologies.

Anyway yesterday I made the not-so-subtle stuff up of including someone on the distribution list for a fairly sensitive email who the email was about rather than to. This type of error I fortunately don't make too often, but the outcomes can be really damaging to relationships and when you work in a relatively small market area (yes, we classify New Zealand that way).  This made me think a little on our communications and the way we do things as humans.  I knew who the email was about and wanted to include the relationship owner in the email.  Somehow I managed to not only include the relationship manager but also the individual... sigh.

The issue here isn't that uncommon and I'm sure at some point in our lives we've made the mistake of putting the wrong people down on communications and sometimes that can have massive consequences in the business world.  Why does it go wrong?  Well, largely it's because of the way our minds work - now I'm no psychologist, but we associate things and make patterns in our head that seem reasonable and we sort of go on auto-pilot - we do things without thinking in other words.  This strikes me as one of the most important reasons to spend time early in a technologies implementation to minimise the human processes.  The old expression goes that to err is human; effectively making mistakes is part of our make up (some of us even make it an art form), the argument here is that if we want to make less stuff ups we need to take 'us' out of the equation as much as possible and in particular the parts where we slide onto auto-pilot.

The simple rule is that whenever we can take humans out of the simple automation parts we do, because we are the ones more likely to make the mistakes.  It doesn't mean that humans don't make decisions because that's still a likely part of your learning and a great way to involve people in decisions that have the big impact, it just means that we take the monotonous tasks (where errors are most likely to creep in) out of the equation.  So if you're looking to reduce human error here's a list of my favourite ways of doing just that:
  1. Bring in your data from a single source of truth. Rubbish in = rubbish out.  If the information in the system is wrong you're starting in a precarious place.  So rather than manually put this information in, we look for ways to sync this information via a single source of truth (like an HR system or HRIS or Payroll system etc)
  2. Remove multiple sign-ins to systems. If possible rather than risk the wrong information when logging on we utilise single-sign-on (SSO) so that the user is automatically and seamlessly transferred from one system to another without getting the chance to get it wrong and sign-in as the wrong person or have to turn to support or helpdesk.
  3. Where decisions are automatic without need for subjectivity then automate them.  When we're setting up someone's learning environment I'm always on the lookout for things that we can set to happen on a rule based or logic type assignment.  For example in Totara when I'm looking at setting up audiences I like to look at what rule sets can be used to dynamically set up people so that they receive the right training, end up on the right courses and even see the right information in the system.  The reasoning here is simple, that I can still make mistakes, but by testing and running through those rules that set my audience up, I end up with an automated system that will logically and consistently apply the rules to make sure the right people are in the right places.
  4. Use dynamic rules.  Rules based on simple logic are great - even better if they are dynamic; that means that the rules get reapplied continually to change the way things are setup as the data changes.  This is the best way to have up to the second data in your reports and the right people in the right learning.
  5. Use dynamic links If we're linking and providing other information we look to set those links up in a dynamic way that change with the content rather than breaking.  There's almost no experience worse in the modern technology based world than a broken-link and you'll lose your target audience quicker than something that goes really quickly (sorry, slight loss of a decent metaphor).
  6. Invest in thorough testing.  Don't forget to never underestimate the ingenuity of fools when claiming something to be fool-proof. We sometimes work with clients who don't want to invest in thorough UAT (user acceptance testing), but this is a really important part before launching anything.  A 'Titanic' type failure on your launch is something you can't recover from so investing a decent amount of time for your own testing is really a must.
  7. KISS.  A great modern acronym for keeping it simple (there are a variety of ones for the final S if you find Stupid offensive).  Keep your systems as simple as possible, Boolean algebra is like any other kind of algebra in that it can often be simplified down.  It can be really hard to follow if you've not tried to do this or used a convoluted way around the problem.  Again working with someone else and explaining what you've done is a good check.

As a final note, I managed to right the email issue yesterday by immediately spotting what I'd done and owning the error.  Immediately I sent out an email to those on the list saying 'sorry' - I also spoke to the person who shouldn't have been on the list and did the same.  The most important thing when erring is to remember that it's what we do as humans; how we deal with the after-effects are what defines us as people.  When you get it wrong (there's really no 'if' about it) own the error.  I've said so many times before but in learning technologies as in life, it's all about ownership.