An IT director at a mid-sized company faces a dilemma: the team needs Kubernetes experts, but current administrators only know classic virtualization. Three options on the table – train current people (upskilling), reskill developers to DevOps (reskilling), or hire ready-made specialists from the market (hiring). Each of these paths makes sense in a different context, each costs differently, and delivers different results at different times.
In 2026, IT decision makers face unprecedented pressure: technologies are evolving faster than ever, the talent market is exceptionally competitive, and budgets for team development – while high – are not unlimited. According to Gartner research from early 2026, 64% of IT organizations identify competency gaps as the main barrier to implementing digital strategy. At the same time, McKinsey indicates that the cost of maintaining a vacancy for a senior IT position can exceed 150,000 PLN per year – counting not only recruitment costs, but above all lost business opportunities.
This is precisely why conscious choices between upskilling, reskilling, and hiring have ceased to be merely an HR or budgetary matter – they are strategic decisions that affect the competitiveness of the entire organization. In this article, we will analyze all three approaches in terms of costs, time, effectiveness, and risk, so you can make a decision based on hard data, not just intuition.
How does upskilling differ from reskilling?
Let’s start with the fundamentals, because these terms – although often used interchangeably – represent completely different developmental strategies.
Upskilling is developing competencies within an employee’s current career path. A Linux administrator learns advanced automation techniques in Ansible, a data analyst discovers new capabilities of Power BI, a Java developer delves into Spring Boot 3.x. It’s evolution in the same role – the employee does what they did, just better, faster, and using newer tools. Upskilling is incremental in nature: we build on existing foundations, adding successive layers of knowledge and skills.
Reskilling is something completely different – it’s requalification to a new role or new technological domain. A manual tester becomes a test automation engineer, a backend developer learns frontend technologies and transitions to full-stack, a network engineer retrains as a cloud architect. This isn’t evolution – it’s transformation. Reskilling means changing career paths, often abandoning part of old knowledge in favor of building a new set of competencies from scratch.
The difference is fundamental in the context of planning. Upskilling is usually a matter of weeks or months, reskilling – months or even a year. Upskilling can be conducted in parallel with operational work (2-3 days of training, then practice in projects), reskilling requires much greater time commitment and often means temporary decreased employee productivity.
Example from real life: a senior Java developer with 8 years of experience needs a week of intensive training plus a month of practice to work efficiently in Spring Boot 3 instead of an older framework version – that’s upskilling. The same developer, deciding to transition to a DevOps engineer role, needs 4-6 months of learning (Kubernetes, Terraform, CI/CD, cloud platforms, monitoring) plus another 3-6 months of practice to reach senior level in the new role – that’s reskilling.
When is hiring the only way out?
Before we get into detailed comparisons, it’s worth identifying situations where hiring is not so much optimal as necessary.
Urgent business needs with time deficit. When a company must launch a new product within 3 months and needs a team it doesn’t yet have, there’s no time to train people from scratch. Hiring ready-made specialists is the only realistic option.
Niche or highly specialized competencies. Some areas – like security research, machine learning engineering, or blockchain development – require such specific, advanced competencies that retraining employees from other IT areas is simply unprofitable. Reskilling a junior Java developer into an ML engineer is an 18-24 month investment. In that time, you can hire a ready expert.
Lack of internal candidates for reskilling. If a company needs frontend developers, but the team consists exclusively of system administrators without any programming experience, attempting reskilling may be ineffective. Some career transitions require predispositions and interests that cannot be created through training.
Strategic building of new organizational competencies. When a company enters a completely new technological area – for example, transitioning from traditional IT to cloud-native development – it needs to “seed” the team with experts who will be mentors for the rest. Hiring 2-3 seniors from the market, who will then train internal people, is a classic pattern.
Shortages in the labor market in a given location. If a company operates in a region where specialists of a given profile are simply not available, upskilling and reskilling become necessary – but this is a requirement imposed by the market, not a strategic choice.
Hiring makes sense as a complement to a development strategy, but rarely as the only strategy. Companies relying solely on external recruitment pay a premium for talent, face high turnover (new people leave more easily than loyal, internally developed ones), and lose the cultural effect that comes from developing their own teams.
Cost comparison: upskilling vs reskilling vs hiring
Let’s move on to hard numbers. The following calculations are based on data from the Polish IT market in 2026, for mid/senior level positions in medium-sized companies.
| Cost category | Upskilling | Reskilling | Hiring |
|---|---|---|---|
| Direct costs | |||
| Recruitment | 0 PLN | 0 PLN | 15,000 - 30,000 PLN (agencies, job boards, recruiter time) |
| Training | 5,000 - 15,000 PLN | 20,000 - 50,000 PLN | 5,000 - 10,000 PLN (onboarding) |
| Certifications | 2,000 - 5,000 PLN | 5,000 - 15,000 PLN | 0 PLN (usually) |
| Indirect costs | |||
| Lost productivity | 10-20% for 2-3 months | 40-60% for 6-12 months | 50-70% for 3-6 months (onboarding) |
| Mentoring/supervision | 5-10h manager | 20-40h manager + mentor | 30-50h manager + buddy |
| Risk of failure | 15% (non-completion) | 30% (abandonment/lack of aptitude) | 25% (wrong hire, departure within a year) |
| Total cost (12 months) | 30,000 - 60,000 PLN | 80,000 - 150,000 PLN | 100,000 - 200,000 PLN |
Key insight: upskilling is 2-3x cheaper than reskilling and 3-4x cheaper than hiring. But that’s only part of the picture.
What makes up these numbers?
Upskilling: The main cost is training (2-5 days of intensive workshops) plus time for practice. The employee operates at slightly reduced productivity (10-20%) for 2-3 months, because they dedicate part of their time to learning and experiments. In the case of certifications (AWS, Azure, Kubernetes), the cost of the exam and preparatory materials is added.
Reskilling: Here costs rise dramatically. The employee works at limited productivity (40-60% of normal) for 6-12 months, because they’re learning a new role. They need intensive mentoring – a senior from the team must dedicate 20-40 hours to code reviews, consultations, support. Training is longer and more expensive (bootcamps, certification paths). The risk of failure is higher – about 30% of reskilling attempts end in abandonment (lack of aptitude, change of plans, leaving the company).
Hiring: Recruitment is 15-30k PLN (if we use agencies or headhunters). Onboarding takes 3-6 months, during which time the new employee operates at 30-50% productivity. The risk of wrong hire or departure within a year is 25% – which means every fourth new employee either doesn’t work out or leaves, and you have to start over. There’s also a less measurable but real cost: new people often expect higher salaries than company “alumni.”
Time to productivity comparison
Costs are one thing, but time is the second key variable – especially in projects with deadlines.
Time to productivity – the time it takes for an employee to reach full productivity in a new role or with new competencies.
Upskilling: 1-3 months. A Linux administrator learning Ansible, after a week of training and a month of practice, is already creating first playbooks in production. After 3 months, they operate at full capacity. Key advantage: domain knowledge and company context already exist, the employee knows systems, processes, people. They only learn a new tool or technique.
Reskilling: 6-18 months. A Java developer retraining to become a DevOps engineer needs 6 months of intensive learning (Kubernetes, Terraform, cloud, CI/CD, monitoring) to start working independently at junior/mid level. Full productivity at senior level – that’s another 6-12 months of practice. But similar to upskilling: the employee knows the company, its architecture, processes. This speeds up onboarding in the new role.
Hiring: 3-9 months. A new senior DevOps engineer with an excellent CV needs 3 months to learn the company’s infrastructure, architectures, processes, team. Mid-level needs 6 months, junior – even 9. In the first months, every new employee, regardless of seniority, operates at limited productivity because they’re learning the context.
What does this mean in practice? If you have 6 months to deliver a project:
- Upskilling – realistic option, team will be productive after 2-3 months
- Reskilling – risky, team may not reach full productivity before deadline
- Hiring – possible, but requires aggressive recruitment and good onboarding processes
When to choose upskilling?
Upskilling is the first choice strategy in most scenarios. When specifically?
1. Technology evolves, but the domain remains the same. The team works in Java 11 and you need to move to Java 21 with new features? Upskilling. Administrators know VMware, and the company is moving to Proxmox? Upskilling. Analysts use Power BI Desktop, and you want to introduce Power BI Service and advanced DAX? Upskilling.
2. You have time (2-6 months) and budget for development. Upskilling requires investment in training and time for practice, but if you have that luxury – return on investment is the highest of all three strategies.
3. You want to increase retention and engagement. Employees who see that the company invests in their development stay longer. According to LinkedIn Learning research from 2025, 94% of employees would stay at a company longer if the company invested in their development. Upskilling is the most effective employer branding tool.
4. You need to build institutional knowledge. Employees developed internally know context, history of architectural decisions, industry specifics. This is knowledge that cannot be bought with a new employee.
5. New technology is incremental relative to current competencies. If the difference between current and target skillset is 20-30% new knowledge, upskilling is an obvious choice. Example: a sysadmin knowing Bash and Python learns Terraform – this is natural evolution, not revolution.
Example: A manufacturing company with a team of 5 Windows Server administrators wants to implement Infrastructure as Code in Terraform. Administrators have solid foundations (networks, storage, security), they just lack IaC and cloud knowledge. Cost: 5-day Terraform + AWS training (15k PLN), a month of practice under consultant supervision (20k PLN), Terraform Associate certification for the team (10k PLN). Total cost: 45k PLN. After 3 months, the team independently manages infrastructure in code. Alternative: hire 2 DevOps engineers – recruitment cost 30k PLN, salaries 40-50k PLN/month, onboarding 6 months. First year cost: 500-600k PLN.
When to choose reskilling?
Reskilling is a strategy for the future – more expensive, longer, riskier than upskilling, but often the only sensible alternative to mass hiring.
1. Company undergoes technological transformation. From monolith to microservices, from on-premise to cloud, from waterfall to agile/DevOps. These aren’t incremental changes – they’re fundamental changes in working methods. Reskilling part of the team (20-30%) is a way to seed new culture and competencies.
2. There’s excess competencies in one area, shortage in another. The company has 10 manual testers but needs test automation engineers. It has 8 backend developers in Java, but is developing frontend and lacks React developers. Instead of laying off some and recruiting others (which is costly and culturally dangerous), you can reskill part of the team.
3. The labor market in a given area is exceptionally expensive/difficult. In 2026, finding a good DevOps engineer, Data Engineer, or Cloud Architect is a challenge – competition for these profiles is brutal, salaries are rising. Reskilling your people may be cheaper and faster than fighting in the market.
4. You have motivated people with appropriate predispositions. Reskilling works when employees want it. Forcing someone to change roles rarely ends successfully. But if you have a manual tester on the team who learns Python on weekends and dreams of automation – that’s an ideal candidate for reskilling.
5. Long-term talent retention strategy. Reskilling is a powerful benefit. You show employees that their career isn’t blocked, that they can develop in new directions without changing employers. This increases loyalty and decreases turnover.
Example: A software house with 15 senior Java developers has fewer and fewer backend projects, and more and more requests for full-stack and frontend. Instead of layoffs and mass hiring, the company offers 5 Java developers a reskilling path: 3-month React/TypeScript bootcamp (30k PLN), mentoring by external expert (50k PLN), six months of work in full-stack projects with reduced productivity (loss ~200k PLN revenue). Total cost: ~280k PLN. After a year, the company has 5 competent full-stack developers who know the backend part of projects (huge value) and can work on both sides of the stack. Alternative: lay off 5 Java devs (severance 150k PLN), hire 5 React devs (recruitment 75k PLN, onboarding 6 months, risk of departures). Cost + risk + cultural cost: significantly higher.
Hybrid strategy: combine all three approaches
In practice, the best organizations don’t choose one strategy – they build a talent portfolio, combining upskilling, reskilling, and hiring in a thoughtful way.
70-20-10 Model:
- 70% competency development through upskilling – this is the core strategy. The team evolves with technology, people become better at what they do.
- 20% through reskilling – part of the team (especially younger, mid-level) gets a chance to change paths, new competencies emerge in the company.
- 10% through hiring – targeted recruitment of seniors, experts in niche areas, leaders who will be mentors for internal talents.
Example of a working hybrid strategy:
An IT company with a 50-person team plans a developmental year:
-
Upskilling (35 people): All programmers undergo training in new framework versions, security best practices, performance optimization. Cost: 150k PLN (group training). ROI: higher code quality, fewer bugs, faster delivery.
-
Reskilling (10 people): 5 manual testers learn test automation (Selenium, Cypress, API testing), 5 backend devs transition to full-stack (React, TypeScript). Cost: 200k PLN (bootcamps, mentoring, reduced productivity). ROI: new internal competencies, reduced market dependence, increased team mobility.
-
Hiring (5 people): Recruitment of 2 senior DevOps, 2 senior frontend, 1 lead architect. Cost: 100k PLN (recruitment) + ~1.5M PLN (annual salaries). ROI: seeding with expertise, mentoring for juniors, quick filling of gaps in key areas.
Total cost: ~1.95M PLN. Effect: team grows from 50 to 55 people, but competencies grow disproportionately – instead of 50 people doing the same as a year ago, you have 55 people with new skills, new roles, fresh expertise from the market.
How EITT supports each of these strategies
EITT specializes in all three scenarios – upskilling, reskilling, and supporting hiring processes through onboarding new employees.
Upskilling: We offer over 2,500 technical training courses (AWS, Azure, Kubernetes, programming, security, data) in formats tailored to needs: in-person workshops, online live, self-paced e-learning. Working with Java 11 and need to move to 21? We have a dedicated workshop. Team uses Terraform at basic level and you want advanced practices? We conduct training for every level. Our average training rating is 4.8/5 – participants leave with practical skills, not just theory.
Reskilling: We build requalification paths lasting 3-12 months, tailored to employee profile and transformation goal. Examples:
- Manual Tester → Test Automation Engineer: 6-month path (Python basics, Selenium/Cypress, API testing, CI/CD integration, design patterns for testing)
- Backend Developer → DevOps Engineer: 8-month path (Linux fundamentals, Docker/Kubernetes, Terraform, cloud platforms, monitoring/observability)
- Sysadmin → Cloud Engineer: 6-month path (AWS/Azure fundamentals, networking in cloud, IaC, security, cost optimization)
Each path includes: training + mentoring + practical projects + certifications. We work with the company from start to finish – we help identify candidates, design the program, monitor progress.
Hiring support: When you recruit new people, we shorten their onboarding time through dedicated technical training. Hired 3 cloud engineers with different backgrounds? We can conduct a 5-day immersive workshop that levels knowledge and prepares them to work in your environment. This is an investment of 15-30k PLN that shortens onboarding from 6 to 3 months – savings of hundreds of thousands of złoty in lost productivity.
Strategic consulting: We help companies design competency development strategies. We conduct skills gap analysis – we identify what competencies are missing, who on the team has potential for upskilling/reskilling, where it’s worth hiring externally. We build annual development plans with specific steps, metrics, budgets.
Cooperation example: An e-commerce company with a 30-person IT team wanted to implement Kubernetes and move applications to cloud. They had a team of sysadmins who only knew classic virtualization and monolithic applications. Together we designed a strategy:
- Upskilling (15 people): Docker, Kubernetes basics, cloud fundamentals training. Cost: 80k PLN. Time: 3 months.
- Reskilling (5 people): Best sysadmins underwent a 6-month DevOps path (Kubernetes advanced, Terraform, CI/CD, monitoring). Cost: 100k PLN.
- Hiring (2 seniors): Recruitment of 2 senior Kubernetes engineers from market as transformation leaders. Cost: 1M PLN/year.
- Mentoring: EITT-hired consultant supported team in practice for six months (code reviews, troubleshooting, best practices). Cost: 120k PLN.
After a year, the company had a working Kubernetes cluster in production, a team of 20+ people able to manage it, and 2 seniors as long-term leaders. Total cost: ~1.5M PLN. Alternative (hiring 10 ready Kubernetes engineers from market): ~4M PLN + turnover risk + lack of development culture.
FAQ
Can reskilling be done without reducing team productivity?
Realistically – no. Reskilling always means a period of reduced productivity because the employee is learning a new role. But you can minimize it:
- Spread learning over time (e.g., 20% work time for learning over a year instead of 100% for 3 months)
- Plan reskilling outside project peak
- Use job shadowing and rotations – employee learns new role working alongside expert
- Start with hybrid roles – e.g., developer learns DevOps, working 70% in development project, 30% in DevOps
How long should upskilling take to be effective?
Depends on scale of change. Small upskilling (new tool, new language feature) – 1-2 weeks. Medium (new framework, new platform) – 1-3 months. Large (new technology in same domain) – 3-6 months. If “upskilling” drags on longer than six months, it’s probably already reskilling.
What’s the risk that an employee will leave right after expensive training?
It’s a real risk. Minimization methods:
- Training agreements: employee commits to staying at company X months after training, otherwise returns costs (proportionally)
- Graduated investments: start with small training, observe engagement, invest more in people who show commitment
- Employer branding: employees stay not only for money, but for development culture. If you run development programs well, people stay longer
- Statistically: internally developed employees have 20-30% lower turnover than organizational average
Does upskilling/reskilling work for all employees?
No. About 15-30% of attempts end in failure – for various reasons (lack of aptitude, life plan changes, burnout, personal problems). Therefore:
- Collect voluntary declarations – people forced to learn rarely learn effectively
- Test predispositions – before expensive reskilling, it’s worth conducting tests, small trial projects
- Monitor progress – if after 3 months there are no effects, better to end the program than invest further
Are certifications necessary in the upskilling/reskilling process?
Depends on industry and role. In cloud areas (AWS, Azure, GCP), security, networking – certifications have great market value and validate knowledge. In programming – much less (practice and portfolio count). But certifications have additional value in internal development context:
- They provide clear goal and learning structure
- They motivate (exam is a deadline)
- They validate knowledge level (person with AWS Solutions Architect certification has minimum guarantee)
- They increase employer branding (company with 20 certified specialists looks better than without)
How to measure ROI from upskilling/reskilling programs?
There are many metrics, depending on goals:
- Time to productivity: how much time passed from program start to moment when employee works independently in new role/with new skillset
- Retention rate: do program participants stay at company longer than average
- Project success rate: do projects with trained people participation end successfully (on time, within budget, with high quality)
- Internal mobility: how many employees changed roles internally vs externally (good internal mobility = lower turnover)
- Cost avoidance: how much you saved developing people internally instead of recruiting from market
- Employee satisfaction: NPS, engagement surveys – do people feel company invests in their development
How to convince management to invest in upskilling/reskilling instead of hiring?
Use business language, not HR:
- Time: Upskilling delivers results faster than recruitment (2-3 months vs 6-9 months)
- Cost: Upskilling is 3-4x cheaper than hiring
- Risk: Retention rate of internal talents is higher than external
- Continuity: People developed internally know context, culture, processes – no risk of “wrong hire”
- Employer branding: Companies investing in development have easier recruitment and lower turnover
Present specific case with numbers: “Instead of recruiting 3 DevOps engineers (year cost: 1.5M PLN + 6 months onboarding + 25% departure risk), we can train 3 of our sysadmins (cost: 200k PLN, 4 months to productivity, 10% failure risk). ROI: 1.3M PLN savings + lower risk.”
Summary: build a strategy tailored to context
There’s no single answer to the question “what pays off more” – upskilling, reskilling, or hiring. The answer is always: it depends. It depends on how much time you have, what your budget is, what competencies you need, what your team is like, what the labor market is like in your region.
But certain patterns repeat:
Upskilling is the default, most cost-effective strategy – whenever you can, invest in developing current people within their current roles. It’s cheapest, fastest, least risky. And it builds a development culture that attracts talent and reduces turnover.
Reskilling is a transformation strategy – when you need to fundamentally change team competencies because the company is changing technological direction. More expensive and longer than upskilling, but still better than mass hiring.
Hiring is a complement, not a foundation – hire seniors, experts, leaders who will be mentors for internal talents. But don’t build entire strategy on recruitment, because it’s the most expensive and risky approach.
Best organizations combine all three strategies in a thoughtful way – 70% upskilling, 20% reskilling, 10% hiring. They build a continuous learning culture where development is the norm, not the exception. And thanks to this, they win the competition for talent – because the best people want to work where they can develop.
If you need help designing a competency development strategy for your team, contact EITT. We’ll conduct skills gap analysis, help choose appropriate development paths, build training programs, and support their implementation. Because competency development isn’t a cost – it’s an investment that pays back many times over in the form of better teams, better products, and better business results.
Ready to plan an IT competency development strategy at your company?
Contact EITT – we’ll conduct a free consultation, identify competency gaps, and help choose the best approach: upskilling, reskilling, or hybrid strategy. We work with over 500 experts and conduct over 2,500 training courses annually – we have the experience to help you build a team ready for the future.
Read Also
- Neurodiversity Myths in the Workplace: What Hiring People with Autism, ADHD, and Other Neurological Differences Really Means
- Ban on Pay Confidentiality Clauses – New Employee Rights to Discuss Salaries
- Objective Compensation Criteria – How to Build a Fair Pay System?
Develop Your Skills
This article is related to the training Team management and development - key competencies. Check the program and sign up to develop your skills with EITT experts.