Selecting for Success: why hiring fails, and how to fix it
The views expressed in this article are solely those of the author.?
Summary
Leveraging human capital within business is an under-utilised method of gaining competitive advantage within the service economy. Candidate selection that is more predictive of future productivity is one of the two primary methods service economy businesses have to materially increase their human capital. This is because the variation in productivity between candidates is always greater than the variation in productivity within a candidate or colleague over time.
The predictiveness of candidate selections is, however, generally very low, subject as it is to human fallibilities of bias and noise. Furthermore, hiring managers have means and motive to reduce time and effort spent on selections, whilst candidates have means and motive to misrepresent themselves by commission or omission –? most of the data used for selections is necessarily provided by the candidate.?
Typical selection processes in service economy businesses do not, however, mitigate these challenges. Many businesses have no defined process, selecting a workforce of no greater productivity than average. Where businesses have created a selection process to better predict candidate productivity, their implementations require considerable indirect costs and add substantially to hiring manager time and effort per hire. The best outcome of such situations are longer role vacancies with commensurately larger opportunity costs, whilst the worst and quite probable outcome is non-compliance with the process which is effectively the same as there being no process at all.?
Science and technology readily available today can be combined to both increase the predictiveness of candidate future productivity, whilst controlling implementation costs and keeping time and effort on part of hiring managers constant or even reduced. This can be achieved through the adaptation of decision-making protocols available within the field of cognitive science, designed to improve predictiveness of hires, delivered through the technology of an intuitive end-to-end process hosted in the cloud, and using generative AI to create first drafts of the required materials at speed for very low cost. By boosting the predictiveness of their candidate selections to increase the productivity of their workforce, and by reducing time and effort on part of hiring managers to maximise compliance whilst minimising implementation time and costs, businesses will gain a competitive advantage.
People vary in productivity?
Candidates and colleagues vary considerably in their knowledge, skills and experience. They also vary in personality, intelligence, cognitive styles, and other psychological attributes including motivation, beliefs, goal orientation and resilience, as well as their more observable behavioural habits and communication styles. Because these highly variable attributes interact with highly variable working contexts to produce variable outcomes, we should not be surprised by the extremely wide distribution between people in their actual or potential productivity within a specific job. This distribution is even wider in the service economy because the fruits of office labour are less tangible and formulaic and are potentially more scalable than those of manual labour, resulting in a much wider range in productivity.
Whilst people can and some do become more productive in the same tasks over time, this potential future variation is on average not nearly as wide as the actual variation between individuals at the present time. This is why choosing who joins the business (Selection) and who is retained and promoted by a business (Performance Management) has greater leverage over a business’ human capital than other available methods – such as structure, skills development and RACI’s.
Selection = Performance Management
Selection and Performance Management are the primary methods through which businesses can increase their human capital. And their essential task is one and the same: evaluating people in terms of their productivity. The difference is that Selection is a predictive judgement of candidate potential future productivity, whereas Performance Management is an evaluative judgement of colleague actual past productivity. But because the task is fundamentally one and the same between them, they suffer the same challenges, which I shall elaborate on, focusing on Selection.?
Predicting productivity is hard and error-prone
Predicting who the most productive candidate will be is extremely hard, the main reasons for which will now be discussed.?
Nobody has a crystal ball?
Predicting who the best candidate will be is subject to a fundamental ‘crystal ball’ constraint: as long as (or at least our experience of) time remains unidirectional, the future is inherently unpredictable. An accidental death of a loved one cannot be predicted. What if this loss sends the candidate into a downward spiral? And their coping involves alcohol, to such an extent that this degrades their focus, memory, and reliability? This is just one example of how the lottery of life can profoundly impact future productivity, whilst being essentially impossible to predict.?
However, we are all equally subject to the crystal ball constraint. It therefore necessarily delimits the scope of our predictions, but within the scope of what is foreseeable, our predictions remain woefully compromised.?
Selections are biased and noisy, and experts are no better?
Candidate selection is typically a human judgement. In every domain studied including selection, human judgements are shockingly bad at predicting the future, subject as they are to considerable bias and noise.?
Bias has received heavy-enough press to reach public awareness, spawning a mini-industry of ostensibly bias-busting training. Even if some of the more exotic ones do not pass social science’s current emphasis on repeatability, it is safe to conclude that there are likely hundreds of biases affecting our decisions. Some of the culprits more likely to affect selections include:
Cognitive biases, all with different nuances, at heart share the common operation of substituting a complex and difficult question (i.e. which candidate will be the most productive?), with a much easier and likely more satisfying one (e.g. which candidate do I like the most?). As the answer to the complex but relevant question will differ substantially from the answer to the easy but irrelevant question, the quality of our answers and resulting decisions are degraded under bias.?
Whilst receiving far less press, noise -? that is, persistent and mostly unpredictable inconsistencies in judgements that lead to poor decisions - is believed by experts to contribute to far more error in business decisions than bias, including selections. Consider that different evaluators can have wildly different views of the same candidate. Even more troublingly perhaps, the same evaluator can have wildly different views of the same candidate at different points in time. Despite needing to feel consistent, news, business and life events, even blood-sugar, can materially influence our judgements, including selection.?
A seeming solution to these issues is to entrust predictive judgements, such as selections, to experts specifically trained for and experienced in the task. Unfortunately, the scientific literature finds experts whose jobs involve making predictions, are on average no more accurate than the general population in their predictions (see the work of Philip Tetlock for more details).?
The cost-benefit asymmetry of selections?
In large corporations, there is a fundamental asymmetry in the cost-benefit between candidates and hiring managers that encourages selection of sub-optimal candidates.?
Candidates have a greater and more immediate benefit to being selected than does the hiring manager in making a selection: compare the certain and more or less immediate and undivided income for the candidate, with the uncertain, relatively delayed and diffuse productivity gains only some of which redound to the hiring manager’s benefit.??
Given hiring managers will interview multiple candidates for one position, they are likely to spend a multiple of time on the process than any one particular candidate will. This creates a relatively stronger incentive for hiring managers to decrease the amount of time and effort put into hiring, especially relative to candidates.?
Finally, the reward for a sub-optimal candidate being selected is considerably greater for the candidate than the cost of onboarding them is to the hiring manager who selects them, because the reward to the former is undivided, whereas the costs to the latter are divided across the team and the wider company.?
领英推荐
Means, motive, and misrepresentation??
Selection involves evaluating factors which are invisible. Whilst these invisible factors should leave evidence in their wake, information used to base selection decisions on is typically collected and presented entirely by the candidate alone. Because candidates also have a strong incentive to be selected, they have both the means and the motive to various forms of misrepresentation by omission or deliberate commission throughout the selection process.
We are typically cooperative with strangers, giving them the benefit of the doubt, especially if they match our expectations of what a professional looks like (the stereotyping bias in action). Because of the evolutionary arms race that occurs between lying and detection, the most dangerous liars are necessarily the most convincing. And so the more convincing the liar, the more critical it is that they are detected in the selection process, but the less likely it is that they will in fact be detected.
Absence of validated and comparable data?
Due to the increasingly atomised nature of society – driven by the relatively high levels of population density and mobility characteristic of most cities in which many or most service economy jobs are available – it is very likely that there is no reliable method of collecting information on the candidate relevant to their candidacy, independently of the candidate. Even referees are selected by the candidate, and where they are selected independently, their reliability is nearly impossible to quantify.
Furthermore, even where information relevant to someone’s candidacy is theoretically accessible, it is practically impossible to access it. The protection of personally identifiable information has placed a lot of information on candidates beyond the reach of the hirer. Additionally, past work of service economy candidates is invisible because it is mostly intangible. And even if it were tangible, or at least objectifiable as the property of their past employer, it is unobtainable.?
Where there is no defined selection process (which is quite typical), data collected between candidates for the same position is inconsistent. This arises, for example, when questions asked vary between candidates for the same position. Whilst a natural consequence of a good conversation, as well as the desire to ask seemingly clever questions on information presented, it results in highly varied datasets between candidates. One reason why this weakens the average quality of hiring decisions, is that ranking operations have been found to be much easier for us to make, and therefore of higher predictive validity, than categorising operations. Ranking operations however, require standardised data and scales.?
Selection in many businesses is compromised
Many, possibly even most, businesses, are running sub-optimal selection processes. Here sub-optimal means that these processes do not enable the company to identify and select, more than chance would allow, those candidates who are more productive than average.
Some companies have no defined selection process at all. Such companies leave the process entirely up to the whims of the hiring manager. As a result, the ultimate quality of staff selections will likely be no better than chance would allow: the distribution of productivity between recruits will be no different than the distribution of productivity in the wider population of workers in their industry. If, however, there is an increasing preponderance of competitors who have implemented a defined selection process which is increasing the ‘hit-rate’ of their selections, this could over time mean that the average productivity within the remaining pool of candidates may decline, leading the company with no defined process to weaker and weaker recruits.?
Where companies have a defined process, there are two prevalent issues which limit the potential quality of selections:?
The costs of poor selections to business?
In the service economy, staff is one of if not the largest single cost. Whilst these costs are variable between industries, functional domains, and levels of seniority, the variability of cost between a pool of candidates for the same role in the same industry is many times smaller than the variability in productivity between them.?
Because of this wide distribution in candidate productivity – and all the difficulties and constraints already described in identifying which candidates will be the most productive – it is quite likely that the productivity of candidates selected by business simply reflects the distribution within the wider pool of candidates. In other words, average.
What these two propositions imply is that there is a huge and untapped competitive advantage to service economy business in hiring candidates more productive than average. For those businesses not taking this competitive advantage there is, therefore, a massive opportunity cost to their selections.?
There are also substantial direct costs to selecting candidates whose productivity is lower than expected at time of selection. The worst and quite likely outcome is eight months of zero productivity gains from the role, and a sunk-cost of eight months salary. Aside from these direct costs, hiring takes up a lot of time. If we see ten candidates per hire, with two hours of interviewing per candidate, that’s twenty hours interviewing per hire. There will also likely need to be preparation for interviews of around one hour per candidate, so thirty hours per hire – nearly one week of work. This multiplies by the number of people on a hiring panel.?
Selections can be improved
Hiring decisions are essentially predictive judgements about future performance of a set of candidates. And we have already seen how sub-optimal such judgements typically are. To improve the quality of hiring decisions, therefore, we should look to both science and technology. The fundamental purpose of this should be to increase the predictiveness of candidate future productivity, whilst keeping time and effort required on part of hiring managers constant or, ideally, reduced.?
Improved selection predictiveness can be achieved through the adaptation of decision-making protocols available within the field of cognitive science. Whilst the replication crisis might call into question some specific findings of the field, the existence of bias and noise in general is not disputed, and decision-making protocols from the field are designed to specifically control these.?
However, whilst businesses should implement such decision-making hygiene, they should balance this against the cost, especially in terms of time and effort required from hiring managers to comply. This is because stacking more time and cognitive load onto their already stacked schedules, will very likely result in non-compliance some of the time. In hiring, as with hand-washing, one lapse can prove fatal.?
To implement a defined selection process that qualitatively improves the predictiveness of hiring decisions whilst keeping time and effort of compliance constant or even reduced, business must leverage available technology,? including intuitive user-interfaces, preferably built on platforms already ubiquitous, cloud-hosting, and generative Artificial Intelligence in order to produce first drafts of supporting materials.?
By implementing decision-making protocols from the field of cognitive science through intuitive cloud-based and AI-enabled technology, service economy businesses can practicably and reliably improve the quality of their hiring decisions, thereby increasing the productivity of their workforce.?
More specific advice
Below I provide more specific advice on the creation of your own hiring protocol:
Thank you for reading to the end of this article. If you liked this article and would like to see more content on this and related subjects, consider following me.?
Bibliography