Today’s WITI had confirmed my long-time suspicion on the relations of numbers, expectations, and behaviour, to be true. It talks about Goodhart’s Law, named after a British economist, which posits that, “When a measure becomes a target, it ceases to be a good measure.” In other words, according to WITI, as soon as you set some metrics into its sole goal, its usefulness will soon decrease. One way in which this can occur is individuals trying to anticipate the effect of a policy, and then taking actions that alter its outcome.
It reminded me of the time when I was venting out my frustrations to my friend A about sending multiple bespoke, customised cover letters and CVs almost every day in the midst of my job hunting, only to no avail. She mentioned she once read an article about companies using AI, instead of employing real human beings, to go through CVs so they could filter the most eligible candidates by recognising a series of keywords placed in the CVs. As someone who had been keeping myself informed in the areas of technology, this did not shock me, but for some reasons I refused to pander to the idea and idealistically, or somewhat naively, believed that every single cover letter and CV I sent would be vetted through by an actual human being who would find that my skills and qualifications would befit their organisations.
It was until I engaged a professional CV consultant for guidance (yes I have gone to that extent) that I was again reminded how prevalent this practice is — which is called ATS (stands for applicant tracking system). It works exactly like A mentioned — the system scans CVs for specific keywords to determine if the job application should be passed along to the recruiter. In doing so, it weeds out the unqualified candidates (ones who don’t have all the right keywords in the system) instead of identifying ones who are the best fit for the job. Interestingly, she did not have much comments for my existing CV — no comments about structure, grammatical errors, etc. — but more towards changing the keywords e.g. I wrote ‘orchestrated’ for one of my project management points, but she made me change it to… ‘managed’ because this was apparently the word that would get picked up by the ATS. In our session, she mentioned that sometimes, to defeat and pass through this system, applications would insert these keywords into the empty spaces (e.g the footnotes) of their CVs, put the text colour in white so they were invisible to the naked eyes but still would get picked up by the system. This was the Goodhart’s Law, and I think this is the same when it comes to our exam-oriented education system as well. When the goal is attached to a preconditioned measurement, you can bet some people will bend the rules and even cheat to game the system.
I thought about this a lot too when I was working on some social media strategy work for some brands before. What I found was they were so initially preoccupied with the numbers and SEO optimisation — which is good — but to the point that by placing SEO-friendly keywords randomly across their content, they lost and forgot to own their own tone and voice and organically grow their presence through the values that had supposedly defined them from day one. In the end, their initiatives drowned in the midst of thousands of other brands who, relying on the same SEO-friendly keywords out there, sounded and had the same brand demeanour as they are. Only then they were advised to grow organically by hiring professional content writers with more, for the lack of a better word, soul — and in many cases, it was a slower and more thoughtful process — did they begin to reclaim their brand voice and tone again.
Reading in my tabs:
- This is a really heartbreaking news: China passes controversial security law for Hong Kong that would give Beijing broad powers to crack down on a variety of political crimes.
- Also, burn the whole planet down.
- Defund the police, but don’t replace it with surveillance tech.
- App bans as a weapon of international diplomacy, as India banned TikTok and other 58 Chinese apps.
- “The consequences of this uprising are too many to count. The case that the police bring danger, escalation, and expense to many situations rather than peace and resolution is now far more widely accepted. Now it must be defended and implemented, and protected from both backlash and that dreary dragging resistance to change that makes the river sluggish.” Rebecca Solnit explains how social change is also something that accrues over time, organisational prepare, people change, moods, and habits evolve.
- “Museums must practice empathy and close the gap between themselves and their communities; they must provide space for conversations on the issues that matter to the lives of their audiences, neighbours, and employees. Museums must be sites of advocacy, not just for the artistic and art-historical traditions that they hold so dear, but for basic rights to life, safety, shelter, well-being, and economic, and intellectual sustenance. Museums must dismantle regimes of power even when that power emanates from within.” An excellent article on rethinking the model of museums (and many other organisations) as the seismic societal shift wavered between presumed neutrality and tentative solidarity.
- The trauma of being a Black social-media manager.
- A free online exhibition about digital identity and privacy.
- I eat sambal with almost everything, so very glad to see that Lara Lee had written an ode to the glorious condiment in Vittles.
- “This place is an ark now. Behave as you would on an ark.”