Data on this page: Based on 48,000+ candidate responses, Oct 2025 to Apr 2026
Last updated: April 2026
What Candidates Say About Codility
Every engineer who completes a Codility assessment is invited to share anonymous feedback on their experience. We have collected more than 1.25 million responses. Those responses shape how we design assessments, build our coding environment, and measure fairness across the platform. This page shows what candidates tell us and what we do with it.
say the assessment content is fair.
n = 45,925
rate their experience Good or Excellent.
n = 46,863
responses in the last 6 months.
1.25M+ total
Technical assessments work best when they reflect how engineers actually work: real environments, practical problems, and a process that respects the candidate’s time and ability. That is the standard we hold ourselves to, and it is one we are constantly working to raise.
The engineers taking assessments today are the same people who will be designing hiring processes tomorrow. Every candidate is a member of the engineering community. How they experience an assessment shapes how they think about technical hiring for years afterwards. Getting this right matters, and the only way to know whether we are getting it right is to ask.
So we ask them. After every assessment, candidates receive an anonymous survey. We ask about the environment, the content, the clarity of instructions, and whether the process felt fair. We have been collecting this data for years, and the dataset now exceeds 1.25 million individual responses.
What follows is what they tell us.
How do candidates rate the Codility assessment experience?
Across the most recent six months of survey data, candidates consistently rate the Codility assessment experience positively. 9 in 10 say the assessment content is fair, and 87% rate their overall experience as Good or Excellent. These findings are consistent across demographic groups.
Monthly satisfaction trend (12 months)
Overall experience rated Good or Excellent. In mid-2025, satisfaction sat in the 82-85% range. Since October 2025, scores have stabilised at 85-87%, a sustained step up. The October peak (91.3%) and August dip (80.4%) reflect normal monthly variance. The shaded area marks the current six-month reporting window.
How candidates responded
Fairness is consistent across demographic groups
Our candidate feedback survey optionally collects demographic data (separately from the assessment itself). Across every group surveyed, fairness ratings remain above 83%.
By gender: “Assessment content seemed fair”
FEmale candidates
N = 9,007
93%
MAle candidates
N = 32,422
91%
By ethnicity: “Assessment content seemed fair”
+5pp
Candidates who have used other assessment platforms rate Codility higher on every metric. On skills evaluation and fairness, the difference is 4 to 5 percentage points. On real-world relevance, it is 8 points higher. Having a frame of reference makes a difference.
Satisfaction holds up, even when candidates score low
The biggest employer branding risk sits with candidates who do not advance. This data shows that even among candidates who rated their own performance lowest, the majority still rate the experience positively.
Percentage of candidates rating overall experience as Good or Excellent, by self-assessed performance score.
What candidates say in their own words
Numbers tell part of the story. The open-text responses tell the rest. We receive thousands of written comments from candidates each quarter. A selection:
The assessment was practical, well-structured, and closely aligned with real-world problem-solving. The tasks encouraged thoughtful data handling and transformation rather than rote coding, which made the exercise both engaging and meaningful.
The focus remained on tackling the questions themselves, rather than dealing with restrictive setup or tooling issues.
The ability to switch tasks whenever is great. Oftentimes I hit a mind block and being able to see the problem again later with fresh eyes is helpful.
The tasks were relevant to real programming problems and focused on practical problem-solving rather than memorization. The platform was easy to use, the instructions were clear, and the time provided was reasonable.
Tasks in the real test were exactly up to the level of my skills. I did not even need to use a personal IDE for troubleshooting. Everything was done in the Codility environment and well within the allocated time window.
The automated tests gave helpful feedback and made it easy to validate each stage of the solution.
Anonymised candidate responses from Codility post-assessment feedback survey, Oct 2025 to Apr 2026.
What are candidates telling us to improve?
Over half of candidates say the assessment tasks need no improvement. For the rest, the most common requests are more examples, shorter descriptions, and content that feels closer to real-world engineering problems. We track these responses to prioritise what we work on next.
When candidates suggest improvements, here is what they ask for
These are not abstract complaints. They are specific, actionable feedback from working engineers. More examples, cleaner descriptions, better findability. Each one feeds directly into how our assessment science and content teams prioritise work.
We show this data because transparency requires it. If we only published the stats where we score well, this page would not be worth reading. The commitment is to show the full picture and demonstrate that we are actively working on the areas candidates flag.
How does Codility use candidate feedback?
Candidate feedback directly shapes assessment design, environment improvements, and content quality at Codility. The survey is not a box-ticking exercise. Responses are analysed regularly and feed into product decisions, content updates, and fairness monitoring. When candidates say something is not working, we treat that as a signal to investigate and act.
The feedback survey runs continuously. Every candidate who completes an assessment is invited to respond, and the data is reviewed as part of a regular cycle. When we update our technical manual (the document that captures all assessment science evidence), candidate feedback is one of the sections that gets refreshed.
There are specific ways this feedback has shaped the platform:
Content realism
Candidates consistently tell us they want assessment tasks that reflect real engineering work, not abstract algorithmic puzzles. This feedback has reinforced our investment in practical, project-style assessment content. When candidates say “add more practical use cases to make it more realistic,” we take that seriously because it aligns with what assessment science tells us about validity: the closer a test mirrors the actual job, the better it predicts performance.
Environment quality
Feedback on the coding environment has driven improvements to our IDE, including the introduction of a VS Code-based environment for live interviews. Engineers spend their working lives in VS Code. Asking them to demonstrate their skills in an unfamiliar environment introduces unnecessary friction. The feedback data confirmed what we suspected: candidates perform better and report higher satisfaction when the environment feels familiar.
11%
of candidates name VS Code as their reference IDE. It is the third most-mentioned coding environment among Codility candidates, behind only HackerRank and “none.” Candidates are not comparing us to competitors. They are comparing us to the tools they use every day.
Fairness monitoring
We track a fairness rating as a standard metric. Clients see their own fairness score in their analytics dashboard, and we use the aggregate data across the platform to monitor whether any assessment type, content area, or environment change is affecting perceived fairness. If a new piece of content triggers a drop in fairness scores, we investigate.
This feedback loop exists because assessment science requires it. A fair assessment is a valid assessment, and candidate perception is one of the signals that tells us whether we are getting it right.
What does Codility do to make assessments fair?
Codility employs I/O psychologists and assessment scientists who design, validate, and monitor every assessment for fairness. This includes content validation, reliability measurement, and adverse impact analysis conducted in partnership with clients. The evidence is documented in a technical manual that is updated every six months.
Fairness in technical assessment is not a feature you ship once. It is an ongoing practice that requires dedicated expertise, structured methodology, and continuous monitoring.
Codility maintains an in-house assessment science team led by I/O psychologists. This team is responsible for ensuring that every assessment meets the standards expected by organisations that take hiring seriously: content validity, scoring reliability, and demographic fairness.
What adverse impact analysis means
Adverse impact is a term from employment law that refers to whether an assessment produces meaningfully different outcomes for different demographic groups. If male candidates consistently score higher than female candidates on the same assessment, that is a potential adverse impact issue that needs investigation.
Codility conducts adverse impact analysis in partnership with clients. Because our assessments are modular (clients configure their own task combinations), adverse impact must be measured at the local level, using the client’s own candidate population and demographic data. This is not something that can be done generically across the platform with a blanket statement.
What we can say is this: across every adverse impact analysis we have conducted with clients, we have rarely seen meaningful differences in scores across gender or ethnicity. This is an important finding, and it is one reason why our clients trust us in regulated industries where adverse impact is a legal requirement, not just best practice.
The technical manual
If we are going to ask organisations to make hiring decisions based on our assessments, we have an obligation to show our work. That is what the technical manual does.
The manual captures the evidence behind every assessment at Codility: content validation, scoring reliability, candidate feedback analysis, and fairness evidence including adverse impact findings. It is updated every six months and shared with prospective and existing clients who want to understand how the platform is built and monitored.
This is the kind of document that I/O psychologists, legal teams, and procurement reviewers look for when evaluating whether an assessment is professionally developed and defensible. Most technical assessment platforms do not maintain one. Many do not employ assessment scientists at all.
The reason this matters goes beyond compliance. Maintaining a technical manual signals how Codility operates: more like a professional assessment publisher than a tech platform that happens to have coding challenges on it. Technical assessments determine who gets hired and who does not. The science behind them should be treated with the same seriousness as the engineering that builds them.
Why does assessment realism matter for candidates?
Realistic assessments produce better signal for employers and a better experience for candidates. When the coding environment, task structure, and tools mirror actual engineering work, candidates can demonstrate what they genuinely know rather than what they have memorised. This is better for everyone: fairer for candidates, more predictive for hiring teams, and more defensible if the process is ever challenged.
Engineers do not solve algorithmic puzzles for a living. They navigate codebases, debug unfamiliar systems, make architectural trade-offs, and collaborate with other engineers. An assessment that tests something fundamentally different from the job is not just a poor candidate experience. It is bad science. The closer a test mirrors the actual work, the better it predicts whether someone will succeed in the role.
This is not a philosophical position. It is a finding supported by decades of research in industrial-organisational psychology. Assessment validity, the technical term for whether a test actually measures what it claims to, increases with fidelity to the target job.
Real coding environments
Our interview product uses a VS Code-based IDE. Candidates get the editor they use every day, with the features they rely on: IntelliSense, file trees, terminal access. They are not fighting an unfamiliar tool while trying to demonstrate their skills.
Practical tasks over algorithmic puzzles
Our assessment content is moving away from abstract algorithm challenges towards tasks that reflect genuine engineering problems. When a candidate tells us “add more practical use cases to make it more realistic,” that is not just feedback. It is validation that we are heading in the right direction.
Candidates are engineers, not test subjects
This is the point that matters most, and the one the industry gets wrong most often. The person sitting a Codility assessment today may be the VP Engineering evaluating assessment platforms in three years. Every interaction shapes how they think about technical hiring.
We do not treat candidates as objects to be screened. We treat them as engineers whose time and effort deserve respect. That means a clear, fair process. It means an environment that feels like real work. It means feedback, not silence. And it means being honest about what we are measuring and why.
Frequently asked questions
How many candidate feedback responses has Codility collected?
More than 1.25 million individual responses. The survey runs continuously, with every candidate invited to respond anonymously after completing an assessment. Data is reviewed regularly and used to inform assessment design, environment improvements, and fairness monitoring.