Methodology

Data source

The market data comes from the Nexus project feed: publicly visible project and job listings for freelancers. New and changed listings are imported continuously and analysed automatically (skills, seniority, work location, contract type, hourly rate). The analysis uses an AI model and can be wrong in individual cases.

A listing is counted once. The same listing often appears with several recruiters; a project counts as a duplicate if it has the same title, the same skills and the same category and appears shortly after the first one.

Listings from the last 90 days in eight IT categories are counted: software development, DevOps & cloud, AI & machine learning, data & analytics, IT infrastructure & security, SAP & ERP, QA & testing, and project management & consulting. Listings without a recognised category are missing, so the numbers are a lower bound.

The numbers are calculated once a week (on Mondays) as a snapshot. Skill pages and category pages show the state of that snapshot ("As of: week …") and do not change until the next one.

Metrics

• Listings: the number of listings in the 90-day window that mention the skill or category.
• Demand share: the share of all IT listings in the window. It is the key metric and differs from the plain count: if the data set grows overall, a skill's share can fall even though more listings exist.
• Trend: the change in demand share compared with the 90 days before (previous quarter). A trend only appears if the earlier period has at least 15 listings.
• Risers: skills with at least 30 listings in both periods, a positive trend and at least 10 more listings than in the previous quarter.
• Minimum basis: a skill page exists from 30 listings. A category page is only indexed if the category has at least 100 listings and at least 3 skills with their own page. Partial figures such as the remote share or the contract types appear only from 30 data points that make up at least 30 % of the listings.

Hourly rates

Only real hourly rates stated in listings are shown (as a field, as an hourly budget or in the text). Converted annual salaries are not mixed in, because as employee figures they are systematically lower than freelance rates.

The median and the middle 50 % (the range between the 25th and the 75th percentile) only appear from 30 hourly rates. Below that, the page says "not enough data" or the section is left out.

Only a small share of listings state a rate at all. The median therefore describes the listings that state a rate, not the whole market.

Verdict and market fit

The market-fit score ranges from 0 to 100 and combines four sub-scores: skill fit, recency, positioning and scarcity. Your profile is compared with the listings that match your focus (your most important skills and your seniority).

Score, sub-scores and findings are calculated by fixed rules, not estimated by an AI. An AI model only words the short rationale of the verdict and may use nothing but numbers and skills from the calculation.

The verdict sorts the score into four bands: from 80 "Strong market fit", from 65 "Solid", from 40 "Needs improvement", below that "Critical". If a sub-score lacks reliable market data (for example on how long it takes until a listing disappears from the market), it counts as neutral.

Limits and liability

Listings are not contracts. The analysis shows what the market is looking for, not what is actually awarded or paid. Direct awards and non-public requests are missing.

The data may contain distortions, for example through the choice of job boards and recruiters covered, through repeated postings or through errors in the automatic analysis.

All information is provided for information purposes. It is no guarantee of contracts, utilisation or hourly rates and does not replace individual advice.

Data protection and your CV

The numbers on skill and category pages are calculated from listings, not from uploaded CVs. How your CV is processed, passed on to service providers and deleted is described in the privacy policy.