When things go wrong people often scramble for someone or something to blame. Following the choas caused by the UK Governments handling of exam results, the thing became the algorithm.
Politicians scrambled to wriggle their way through arguments to find the perfect fomula to help deflect calls for resignations. With many not really understanding what it was, unable to defend itself, ministers no doubt thought they’d found the perfect algorithm for the problem and a great scapegoat.
What is an algorithm
The basic definition of an algorithm is ‘a step-by-step set of instructions for solving a problem or performing a task.’
It is not some mysterious thing which boffins working with some dart art. And most important thing to not is an algorithm is only as good as the instructions behind it.
Run before you trust
Gavin Williams MP, minister for Education, after being forced into a humiliating U-Turn ‘we immediately took the action yesterday”. After days before defending the algorythm as robust he suddenly discovered it wasn’t.
Weird. With computer algorithms it is possible to test the impact of an algorithm. Companies do this all the time, setting up replicants of a live environment to run the algorithm in a safe space. The outcomes can then be analysed and algorithm tweaked.
Exam results were known of OFQUAL in advance, the algorithm could have been run and the extent of its impact examined before disrupting thousands of peoples lives.
By failing to make sure the algorithm worked as intenteded it is the governments failure not the passive algorithm which was just doing what it was told
Danger of blaming algorythms
It is dangerous for all of us for government ministers to promote distrusts in algorythms. In so many parts of our lives algorithms are of huge benefit from speeding up the download speed of a video to giving you more relevant content in your social stream.
Blaming algorithm could kill trust in something that in so many aspects of our lives brings benefits. Algorithm help to save lives, widely used in healthcare they help doctors take a number of factors to predict outcomes and prescibe treatment.
Transparancey is key
An algorithm is only as good as the rules that guide it. This is why it is important that those creating algorithms need to be transparent about these rules, together with known strengths and weaknesses of the algorithm.
Algorithm’s and marketing
Sometimes this is not possible. Google, for instance, would not want search marketers to know their algorithm to be able to skew the system based on the alogrithm and not the quality of the result.
A lesson that all marketers can take from the governments handling of the exam results crisis, when using algorithms, make sure you understand the rules and what impact they will have.