An average with nothing in the middle
A market where 15% of roles are remote sounds like a market in transition — some employers moved, most did not, and the number drifts upward each quarter. The distribution says something else entirely. Sort the 41 cities by the share of listings flagged remote-eligible and the middle of the range is nearly empty: 17 cities sit at exactly zero, 14 sit at half or more, and only ten fall anywhere between 1% and 49%.
The two groups do not look like each other. The highest shares belong to Tunis at 100%, Doha at 92%, Casablanca at 91%, Cairo at 86% and Lisbon at 83%. The zeros belong to London, Paris, Berlin, Amsterdam, New York, San Francisco, Tokyo and Singapore. The cities that built the office economy report the least remote hiring; the cities hiring into a global talent pool report the most.
| City | Remote-flagged | Note |
|---|---|---|
| Tunis | 100% | every listing flagged |
| Doha | 92% | |
| Casablanca | 91% | |
| Cairo | 86% | |
| Lisbon | 83% | |
| Copenhagen | 73% | |
| Milan | 70% | |
| Birmingham | 69% | |
| London | 0% | not one listing flagged |
| Paris | 0% | |
| Berlin | 0% | |
| New York | 0% | |
| San Francisco | 0% | |
| Tokyo | 0% | |
| Singapore | 0% |
Zero does not mean nobody works from home
This is where a number has to be read rather than quoted. The flag counts listings that say so — an employer marking a role remote-eligible in the posting itself. It does not detect a hybrid policy mentioned at interview, a team that works from home three days a week by custom, or a contract that allows it without advertising it.
So a 0% city is not a city without remote work. It is a city where no employer in the current snapshot wrote it down. That is a claim about disclosure, and disclosure is exactly what a job seeker is deciding on: a role that does not say is a role you cannot filter for, cannot plan around, and will not learn about until late. The market is not split between remote and office. It is split between employers who commit in writing and employers who leave it open.
The skills employers name are not the ones you would guess
The same discipline applies to skills. Ranked by how many listings name them, the top two are not tools: Leadership appears in 69 listings and Communication in 40. Together that is more than Python, TypeScript, Adobe Photoshop, Adobe Illustrator and Accounting combined — 109 against 67.
This is not an argument that technical skill is optional, and the ranking is partly an artefact of how listings are written: "Python" appears where a job needs Python, while "leadership" is written into postings across every function. But it is a fair account of what employers put in the text, and it says something plain about a CV that lists only tools.
- Leadership — 69 listingshuman skill
- Communication — 40 listingshuman skill
- Python — 26 listings
- Project Management — 14 listingshuman skill
- Adobe Photoshop — 12 listings
- TypeScript — 10 listings
- Accounting — 10 listings
- Adobe Illustrator — 9 listings
Concentration, and four days of movement
Supply is not spread evenly across employers either. Stripe alone holds 174 live roles, followed by Datadog at 130, Databricks at 85, Okta at 82 and Block at 77. Those five account for roughly 14% of every open role in the aggregate — five names out of thousands of hiring companies.
The pool is also moving quickly. Our previous edition, four days earlier, read 3,579 openings; today it reads 3,864. That is 285 roles, an 8% rise in four days — a reminder that any snapshot of a job market, including this one, describes a day rather than a season.
What to do with this
If you are searching: treat a missing remote flag as missing information, not as a no, and ask early. Treat a city-level percentage as a measure of local disclosure norms, not of how people actually work there.
If you are hiring: in a market where 17 cities report zero, writing the arrangement into the posting is close to free differentiation. Candidates filter on what is stated. Nothing that is left unsaid gets filtered in.
Every figure here is a page you can open. Start with the market by city and read your own market.
Earlier editions
- · 8 min readYouth unemployment looks stable. Nine million more young people are neither working nor studying.The global youth unemployment rate went from 12.3% in 2023 to 12.4% in 2025 - a tenth of a point, and the ILO projects it flat through 2027. Over the same two years the number of young people not in employment, education or training rose by nine million, to 257 million. Both come from the same report. They are not in conflict: they count different people.Read the analysis →
- · 7 min readBritain lost 19,000 vacancies in a year. Firms with under ten staff lost 18,000 of them.UK vacancies fell 2.7% over the year to 707,000. At businesses with 1 to 9 employees they fell 16.1%, to 95,000 - six times the rate of the market as a whole. That band holds about an eighth of all vacancies and accounts for almost the entire annual decline. The headline is a small number sitting on top of a very uneven one.Read the analysis →
- · 7 min readChanging jobs still pays more. Two trackers cannot agree how much.For July 2026, ADP put the annual raise for people who changed jobs at 7.0% and for people who stayed at 4.4% - a gap of 2.6 points, the widest in a year. The Atlanta Fed, reading a different survey the same month, put changers at 4.4% and stayers at 3.6% - a gap of 0.8. Both are real. 4.4% is the mover's raise in one measure and the stayer's raise in the other.Read the analysis →