Onboarding, learning and development

Hana opens the sixth session with a number that worries her: one in four of Lakeview's newly graduated nurses leaves within the first year. Exit interviews tell the same story each time. The new nurse was given a two-day orientation full of policies and passwords, assigned to a busy unit with a preceptor who was also carrying a full patient load, and left to sink or swim on night shifts within a month. Those who stayed say they learned the job in spite of the system. Dr. Rahimi says the story contains most of this lesson's themes: the first months shape whether people stay, most learning happens on the job whether or not it is designed, and development is an investment that the organization either manages or loses.

Onboarding

Onboarding, sometimes called organizational socialization, is the process by which new employees acquire the knowledge, skills and relationships to become effective members of the organization. It is different from orientation, which is the event on the first day or two. Good onboarding lasts months and has a structure. Talya Bauer summarized it in four levels. Compliance covers the rules and paperwork. Clarification makes sure new hires understand their job and what is expected of them. Culture conveys the organization's norms and how things are really done. Connection builds the relationships and networks that new hires need to get work done and to feel they belong. Most organizations do the first well and the last two poorly.

Research on socialization points to a few outcomes that predict whether new hires stay and perform: role clarity, confidence in their ability to do the job, called self-efficacy, and social acceptance. Practices that build them include a clear plan for the first ninety days, early and frequent feedback, a buddy or mentor who is not the new hire's manager, structured introductions to the people the new hire will depend on, and early tasks that allow quick wins. The manager's behaviour in the first weeks matters more than any program: whether they meet the new hire on day one, explain priorities, check in regularly and introduce them to the team.

The cost of poor onboarding shows up as early turnover and slow ramp-up. Time to proficiency is the time a new hire takes to reach the productivity of an experienced employee. During ramp-up, the organization pays full wages for partial output, and colleagues spend time helping. Dr. Rahimi shows that the cost of ramp-up is the area between the productivity curve of an experienced worker and the curve of a new hire, and that onboarding's job is to make that area smaller.

# Cost of ramp-up for a new nurse at Lakeview, with standard vs structured onboarding
pay_per_month = 7_900          # pay and benefits per month

def productivity(month, speed):
    # share of full productivity, rising toward 1.0
    return min(1.0, 0.35 + speed * month)

def ramp_cost(speed, months=12):
    return sum((1 - productivity(m, speed)) * pay_per_month for m in range(1, months + 1))

standard, structured = ramp_cost(0.045), ramp_cost(0.085)
program = 6_500                # extra preceptor time, residency sessions and mentor pay per nurse
print(f"ramp-up cost, standard onboarding:   {standard:9,.0f}")
print(f"ramp-up cost, structured onboarding: {structured:9,.0f}")
print(f"saving per nurse before turnover effects: {standard - structured - program:,.0f}")

# Add the turnover effect: first-year leaving falls from 25% to 12%; each leaver costs 60,000
hires = 80
avoided = hires * (0.25 - 0.12)
print(f"leavers avoided per year: {avoided:.0f}, worth about {avoided * 60_000:,.0f}")

ramp-up cost, standard onboarding:      33,891
ramp-up cost, structured onboarding:    17,143
saving per nurse before turnover effects: 10,248
leavers avoided per year: 10, worth about 624,000

The calculation persuades Hana that a structured residency program for new graduates would pay for itself even before its effect on turnover, which is worth far more: ten fewer leavers a year at a replacement cost of about 60,000 dollars each is about 620,000 dollars. Owen notes that his figure for replacing a control room operator, who takes two years to qualify fully, would be several times higher. Dr. Rahimi adds that residency and preceptor programs for new nurses are well established in the nursing literature, and that their main ingredient is protected time: preceptors whose patient loads are reduced so that teaching is part of the job rather than an extra.

Learning on the job and the 70-20-10 idea

Much of what people learn at work they learn by doing it. A widely quoted rule of thumb, which came from interviews with executives at the Center for Creative Leadership in the 1980s, holds that about 70 per cent of development comes from challenging assignments, 20 per cent from other people such as bosses, mentors and peers, and 10 per cent from formal courses. Dr. Rahimi warns that the numbers are not research findings and should not be used as a budget formula. The useful point is that formal training is only one part of development, and that assignments and relationships should be designed as deliberately as courses.

Development assignments include stretch roles, rotations, projects across functions, starting something new, fixing something broken and leading people for the first time. They work when they are chosen for what the person needs to learn, when there is support and feedback during the assignment, and when the person has time to reflect afterward. Mentoring pairs a less experienced person with a more experienced one for guidance on career and skills; sponsorship goes further, with a senior person using their influence to advocate for the protégé's advancement. Research has found that women and members of under-represented groups are often mentored but less often sponsored.

Designing training

Formal training is most effective when it starts from a clear need. A training needs analysis looks at three levels. Organizational analysis asks where training fits the strategy and whether the organization will support the use of new skills. Task analysis identifies what must be learned to perform the job, drawing on job analysis. Person analysis identifies who needs the training, because performance problems are not always skill problems: an employee who knows how to do the job but is not doing it may face poor tools, unclear expectations or misaligned incentives, none of which training will fix. A common design model, ADDIE, moves through analysis, design, development, implementation and evaluation.

Research on learning gives practical rules. Spaced practice, spread over time, beats massed practice crammed into one session. Retrieval practice, in which learners recall and apply what they have learned, beats rereading. Practice should resemble the conditions of the job, including realistic errors. Feedback should be prompt and specific. Simulation is valuable for rare, high-stakes situations, which is why airlines train pilots and nurses train for cardiac arrests in simulators. Online learning scales cheaply and works well for knowledge, especially in short modules; skills that involve judgment, interpersonal behaviour or physical technique need practice with feedback.

Transfer of training

Training is wasted if people do not use what they learned back on the job. Timothy Baldwin and Kevin Ford's 1988 review of transfer identified three groups of factors: the trainee's ability and motivation, the design of the training and the work environment. The work environment is often decisive. Trainees who return to managers who ignore the training, colleagues who mock it or a workload that leaves no time to practise it soon revert to old habits. Practical steps to improve transfer include involving managers before and after training, setting goals for applying the skills, giving early opportunities to use them and following up weeks later.

Daniel says Northfield learned this with a problem-solving course for production leads. The course itself was rated highly, but six months later few leads were using the method, because their supervisors still demanded quick fixes and nobody asked for root-cause reports. When the plant manager began opening each weekly meeting by reviewing two problem-solving reports, use of the method rose sharply within a quarter.

Evaluating training

Donald Kirkpatrick proposed in 1959 a four-level model for evaluating training that remains the most used framework. Level 1, reaction, asks whether participants liked the training. Level 2, learning, measures whether they gained knowledge or skill, through tests or demonstrations. Level 3, behaviour, asks whether they changed what they do on the job. Level 4, results, asks whether the organization's outcomes improved, such as quality, sales, safety or turnover. Jack Phillips added a fifth level, return on investment, which compares the money value of results with the cost of the training.

Most organizations measure level 1 and stop. Dr. Rahimi explains why: reactions are easy to collect, while results are affected by many things besides training, so attributing them to training requires comparisons. Where possible, compare trained and untrained groups, or measure before and after with a comparison group, as the cohort learned to do with experiments in MBA-512. Where that is impossible, at least measure behaviour on the job, because training that does not change behaviour cannot change results.

Careers, reskilling and paying for development

Career development has changed from a ladder to a lattice. Fewer organizations promise long careers with predictable promotions, and more employees expect to move across functions and between employers. The organization's role is to make internal opportunities visible, through internal job postings and talent marketplaces that match employees to projects, and to support development that builds skills employees and the organization both value. Many employers fear that training makes employees more attractive to rivals. Research generally finds the opposite on balance: employers that invest in development tend to have lower turnover, especially when development is paired with opportunities to use new skills internally.

Governments in Canada support some employer training. Several provinces offer grants that share the cost of training for employees, and apprenticeship programs come with federal grants and tax credits. Quebec requires employers with a payroll above two million dollars to invest at least one per cent of payroll in training or pay the difference into a provincial fund. Employer-paid tuition for courses related to the job is generally not a taxable benefit to the employee.

Artificial intelligence is changing both what people need to learn and how they learn it. Some tasks in many jobs, especially in writing, analysis, coding and customer service, can now be done faster with AI tools, which shifts the skills needed toward judgment, checking and working with clients. Leo says Kestrel's support engineers now use AI to draft responses and diagnose common problems, so new hires reach basic proficiency faster but need more training in recognizing when the tool is wrong. Dr. Rahimi asks the cohort to treat reskilling as part of workforce planning: identify roles whose tasks are changing, estimate the new skill needs and start development before the change arrives.