Choosing job posting data: a checklist for recruitment agencies

A long list of results is not necessarily a good result. If a consultant has to discard most of the suggested roles, a substantial part of the research work remains. When choosing job posting data for business development, do not just ask how many adverts a solution contains. Ask which of them will actually help in your daily work.

For candidate-led business development, the key question is: Based on their qualifications, would this candidate be a serious hiring option for the employer? A similar job title alone is not enough.

That does not predict whether someone will be hired. But experienced recruiters can assess whether a role is professionally plausible, which requirements are met and what needs clarification before an introduction. That judgement should be at the centre of a provider evaluation.

First clarify the purpose: market search or candidate matching?

Job posting data makes open roles and hiring companies searchable. If you want to develop a market segment in Germany or Austria, for example, you need relevant data and suitable filters. You do not need a particular candidate in mind at this stage.

Candidate matching adds another task: the roles need to fit a specific person's experience and qualifications and be prioritised meaningfully. Current data and good matching are therefore two separate questions of quality. Neither replaces the other.

Test the solution against the work you intend to use it for. If your focus is candidate marketing, an impressive demonstration of filters is not enough. The roles suggested for your candidates need to be convincing too.

If you are specifically considering an alternative to index Anzeigendaten, see our comparison of talentkiwi and index Anzeigendaten, in German. It covers documented features, differences in focus and points to check.

1. Does the data cover your actual market?

Start with the countries, regions and roles in which your team makes placements. A large overall number of adverts tells you little about how well your particular segment is represented.

During the trial, search for some vacancies you already know from your work. Can you find them? Can you narrow the results meaningfully by role and region? Also check whether you discover relevant companies beyond your usual contacts.

Ask the provider about the sources and coverage of its data and which countries are included. Finding individual adverts does not establish complete market coverage. For your decision, the first question is whether the data supports your agreed search cases well enough to be useful.

2. Are the roles still being advertised?

Even a relevant match loses its value if the vacancy is no longer available. Freshness should therefore be something you can verify in the results, not just a claim in a product description.

Open the original adverts for a sample of results. Are they still accessible? Is the position still being advertised, or is it marked as filled? Are the employer, location and responsibilities consistent with the entry in the solution?

A publication date provides context, but does not by itself confirm that a vacancy is still open. Equally, an older date does not automatically mean the role has been filled. Ask how the provider handles changes and removed adverts.

Also examine how duplicate adverts are handled. Does the same role appear several times through different sources? For daily work, the number of distinct opportunities matters more than the length of the list.

3. Do the first ten matches fit the candidate professionally?

For a matching solution, this is the central practical test. Use a candidate whose experience you know well and review the first ten results. If there are fewer, assess those available rather than artificially filling the list.

Place each role into one of three groups:

• Professionally plausible: The essential requirements align with demonstrated experience. A closer review or introduction appears worthwhile.
• Needs clarification: The role could fit, but the CV or advert is missing crucial information.
• Not suitable: An essential requirement or practical condition rules it out.

Briefly record the reason in each case. This turns “I like the matches” into a review that others can follow. Do the responsibilities genuinely fit, or just the title? Does the level of responsibility align with the candidate's experience? Are mandatory qualifications missing?

Useful prioritisation puts professionally relevant roles near the top. Also review a few results further down the list. Otherwise, you will see unsuitable suggestions at the top but miss good opportunities that the system may have ranked too low.

4. Does the explanation help you decide?

A fit score can make the ranking easier to interpret. On its own, however, it does not explain why a role fits. More useful is an explanation that connects specific requirements to demonstrated experience and keeps missing information visible.

Choose a result you can assess confidently yourself. Do the stated reasons agree with the CV and job advert? Are differences identified? Does an unverified skill remain an open question rather than being attributed to the candidate?

A fit score is not a probability of being hired. Nor should a fixed threshold determine which roles a consultant sees without first checking whether it is appropriate. You need several candidates and searches to judge whether the ranking meaningfully supports your professional assessment.

5. How much work remains before you have a useful selection?

Result quality and usability go together in daily work. A suitable role is of little help if its context is difficult to access or the same information has to be transferred repeatedly.

Test the complete process: add a candidate or define a market segment, review the results, open the original advert and record relevant roles. Can you see the employer and publication date? Can you find contact people where available? Is it clear when this information is missing?

Also record how long it takes to reach a professionally useful selection. Search speed is only one part of that. The effort required to identify unsuitable results and gather information for the next step matters too.

A practical plan for a two-week trial

A trial using your own candidates is more informative than a demonstration based entirely on prepared examples. During talentkiwi's two-week free matching trial, you could follow this approach:

1. Choose search cases. Use several candidates from your actual recruitment specialism: someone with a typical career history, a specialist and someone with a more complex CV. Record the essential requirements before searching.
2. Assess the initial selection yourself. Review the first ten matches for each candidate using the same method. Document suitable roles, open questions, reasons for exclusion and the effort involved.
3. Check the original adverts. Verify whether the interesting vacancies are still current and whether several results describe the same position.
4. Compare assessments with a colleague. Discuss differences using the CV and advert. This helps distinguish a matching issue from different expectations about the candidate.
5. Review again at the end. Repeat selected searches and assess whether the results remain useful throughout the trial. When comparing providers, use the same candidates and requirements over the same period wherever possible.

A sample like this does not produce a universally applicable success rate. It does, however, show whether the solution works for your tasks and where its limits lie.

Your team's professional judgement makes the final decision

The strongest basis for a decision is a set of specific search cases with clear assessments: which roles would your consultants pursue? Which need more information? Which would they reject, and why?

When current adverts, professional fit and clear explanations come together, a results list can become a reliable basis for action. The article on candidate marketing in permanent recruitment explains how to use that information in client conversations.

It is not the biggest list that needs to convince you, but the roles your team can actually work with.