How to compare two specialists without a CV, photo or job title: 7 signals that actually matter

The OECD notes that employers still often use job titles, years of experience and qualifications as proxies for what a person can do. At the same time, the absence of a common, precise skills language makes capabilities harder to compare. [1]

In its 2026 report, the OECD describes skills-first hiring as an approach in which people are assessed on skills they can demonstrate, while formal qualifications and prior experience play a complementary role. The OECD also cautions that evidence on effective implementation remains limited and is still largely based on early organisational experience. [2]

For that reason, the seven signals below should not be treated as a scientifically validated scale. They are a practical comparison framework built from principles in research and guidance on work-related assessment, skills and evidence.

The U.S. Office of Personnel Management describes job analysis as the foundation of assessment and selection decisions: first understand the tasks, the competencies required, and the relationship between them. [3]

In practice, before opening two specialist profiles, write down:

• the problem that needs to be solved,
• the 3-7 skills that are genuinely essential,
• the tasks that will be performed,
• the required level of independence,
• the important constraints,
• the result you would consider sufficient.

This way, the criteria exist before you see the person, rather than being invented after one profile creates a stronger first impression.

The first question should not be: who has more skills?

A better question is: does this person have the skills needed to solve this specific problem?

A list of 40 skills is not necessarily better than six that closely match the task. The OECD stresses the need to break roles down into clearly defined skills rather than infer capability only from job titles. [1]

Compare not the number of labels, but how well the person's skills cover the requirements of your problem.

The statement "I can analyse data" is a weaker signal than seeing work in which that skill was actually used.

Depending on the profession, stronger evidence may include:

• a work sample,
• a project description,
• a prototype,
• a code excerpt that can be shown legally,
• an analysis,
• a report,
• a publication,
• campaign material,
• a result confirmed by a credible source.

OPM describes work samples as tasks or activities that mirror work performed on the job. [4]

The closer the evidence is to the real task you plan to assign, the more information it gives you about fit.

The same skill can be used under very different conditions.

Designing the interface of a simple informational website and designing a process inside a financial system with multiple user roles may require some of the same skills, but they are not the same problem.

Check, among other things:

• scale,
• type of user or customer,
• number of dependencies,
• responsibility for decisions,
• quality requirements,
• time or regulatory constraints,
• the need to work with other specialists.

A similar context does not guarantee success, but it helps estimate the size of the gap between previous experience and the new task.

The fact that a project appears in a portfolio does not tell you what the person actually did.

Compare:

• what they were personally responsible for,
• which decisions they made,
• what they completed independently,
• what was created jointly with the team,
• what part of the result can reasonably be connected to their work.

A specialist who precisely describes the boundaries of their own responsibility gives you better evidence for evaluation than someone who presents the entire outcome of a multi-person project as their own.

Experience from several years ago can still be valuable, but its relevance depends on the field.

In areas that change quickly because of technology, recency may matter a great deal. In other professions, foundational principles, domain knowledge or experience with a particular class of problem may remain useful for much longer.

Do not use one rigid age limit for evidence. Ask instead:

does the way this work was performed still correspond to the conditions, tools and requirements the specialist would face today?

A result helps distinguish a list of activities from the actual effect of the work.

It may include, for example:

• a delivered product,
• a shorter process,
• fewer errors,
• improved usability supported by research,
• delivery within defined constraints,
• a publication,
• a solution that was actually implemented,
• an outcome that can be confirmed by a client or another source.

Be careful with causal claims. If a business metric improved after a team project, that does not automatically mean one person caused the entire change.

The most useful signal connects the result with a clearly described scope of personal responsibility.

The strongest specialist on substance is not always the best choice for a particular engagement if the practical conditions do not align.

Separately from competence, compare:

• availability,
• possible start date,
• service scope,
• pricing and billing model,
• working model,
• communication language if relevant,
• ability to work in the required time zone,
• constraints on both sides.

Price is not a measure of competence, and availability is not evidence of quality. They are, however, real feasibility criteria and should therefore be assessed separately.

There is no basis for turning these seven signals into a universal numeric score that supposedly identifies the better specialist with precision.

A simple comparison card is more useful:

Strong signal - there is direct, relevant evidence.
Partial signal - evidence exists, but the context or scope differs from what is needed.
Insufficient information - there is not enough information to judge.

The key principle is:

missing information is not the same as a poor result.

If the information matters, obtain it instead of automatically assigning the lowest assessment.

Suppose you need to improve the initial setup process in a B2B application because users struggle to complete their first configuration.

Specialist A shows many visually attractive digital product projects. The material does not explain which parts they personally completed, how they investigated user problems, or what resulted from the changes.

Specialist B shows fewer projects, but one concerns a similar process. They describe their responsibility, show the research process, part of a prototype, how a decision was reached, and present the final outcome as a team result.

Based on this information, Specialist B provides more evidence related to the specific problem. That does not mean they are a better specialist in general. It means only that, at this stage, there is a stronger basis for assessing their fit for this particular task.

OPM describes work samples and simulations as methods in which a person performs tasks that reflect actual activities carried out in the job. [4]

CIPD also discusses skills-based tasks, including work samples and simulations, and states that they should resemble real job tasks as closely as possible. [6]

If you need an additional check, choose a small, representative task, not an unpaid piece of a real project with substantial business value.

The task should measure what you actually need, use clear criteria, and be comparable for everyone being assessed.

A structured interview uses the same predetermined questions in the same order and evaluates responses using the same rating scale and standards. OPM says this helps give candidates equal opportunities to provide information and supports accurate, consistent assessment. [5]

CIPD recommends asking candidates the same questions in the same order and evaluating responses against pre-agreed criteria. [6]

In practice, rather than questioning one person about technical decisions and the other about personality, ask both about the same problems, decisions and constraints.

Research on anonymous applications suggests that reducing the visibility of identity information can, in some settings, reduce discriminatory effects at an early stage of selection. Results are not uniform, however, and some studies have found unintended effects. The IZA review explicitly discusses both potential benefits and limitations of anonymisation. [7]

So the correct conclusion is not:

"hiding a name removes bias".

A better conclusion is:

"limiting some identity information at the first stage can help focus the comparison on task-related criteria, but by itself it does not guarantee an unbiased or better decision".

In real collaboration, not every piece of information can remain hidden indefinitely.

At a later stage you may need, among other things:

• information needed to enter into a contract,
• billing information,
• identity verification,
• information required for legal or security reasons,
• direct communication between the parties.

The point is therefore not permanent anonymity, but the order in which information is disclosed: first the data needed to assess professional fit, and identifying information when it is genuinely needed for the next stage.

In the European Union, Article 5 of the GDPR includes the principle of data minimisation: personal data should be adequate, relevant and limited to what is necessary for the purposes for which it is processed. [8]

This does not automatically mean that less data is always better. The scope of data needs to match the specific purpose and lawful basis for processing.

For an initial comparison of specialists, a useful practical question is therefore:

are a photo, age or full identity necessary at this stage to assess the ability to perform the specific work?

1. You review profiles first and invent the criteria afterwards.

2. You count skills instead of checking whether they match the problem.

3. You treat an attractive project as proof that one person created everything.

4. You compare one person through a portfolio and the other through a conversation.

5. You treat missing information as evidence of weak competence.

6. You treat price, years of experience or a famous employer as a direct measure of quality.

7. You create one numeric score without explaining the evidence behind it.

For both people, answer the same questions:

1. How well do their skills cover the requirements of the problem?
2. How strong is the evidence that those skills were actually applied?
3. How similar were the previous context and level of complexity?
4. Do we know what the person was personally responsible for?
5. Is the evidence sufficiently current for this field?
6. Is the result described credibly and can it be checked at least in part?
7. Do the working conditions make the engagement genuinely feasible?

For every point, also record what you do not know. The missing information is often exactly what the next question should address.

The comparison should answer this question:

who gives us stronger grounds to expect good performance on this specific task under these specific conditions?

It does not answer who is a more valuable person, who has the better career, or who is the better specialist in every possible situation.

That distinction matters. Two highly capable people may be the right choice for completely different problems.

A CV, job title, university, famous employer or photo can shape a first impression, but none of them alone answers the most important question:

can this person solve my problem, and do I have enough evidence to believe that?

A better order of evaluation is:

problem -> required skills -> evidence -> context -> responsibility -> result -> working conditions.

Only then should you add the information needed for conversation, verification and entering into a working relationship.

This approach does not remove uncertainty. It does make it easier to see what the decision is actually based on.

[1] OECD, A Skills-First Labour Market - Building a common skills language, 2026
Open source

[2] OECD, A Skills-First Labour Market - Promoting skills-first hiring and talent management, 2026
Open source

[3] U.S. Office of Personnel Management - Job Analysis
Open source

[4] U.S. Office of Personnel Management - Work Samples and Simulations
Open source

[5] U.S. Office of Personnel Management - Structured Interviews
Open source

[6] CIPD - Selection Methods, 2024
Open source

[7] IZA World of Labour - Anonymous job applications and hiring discrimination
Open source

[8] Regulation (EU) 2016/679 - GDPR, Article 5, EUR-Lex
Open source

Methodological note: some sources concern formal employee recruitment. This article uses their principles for task analysis, skills assessment, work samples and structured comparison as a reference point when choosing an independent specialist or service provider as well. This does not mean that every recruitment procedure transfers directly to every service relationship.