Dice Market Snapshot, Counts You Can Actually Add Up
Pricing
from $0.01 / 1,000 facet count returneds
Dice Market Snapshot, Counts You Can Actually Add Up
The United States tech job market as dice.com counts it, taken in one request. Every count says whether it splits the board cleanly, double counts adverts, or leaves some out, because on dice.com all three happen and none of them is marked.
Pricing
from $0.01 / 1,000 facet count returneds
Rating
0.0
(0)
Developer
Gregory Bolshakov
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
9 days ago
Last modified
Share
Dice market snapshot
The United States tech job market as dice.com counts it. One request, about 21 rows, and every count says what you are allowed to do with it.
dice.com's counts do four different things
On 16 September 2026 the board held 111,699 adverts. Its own facet counts added up like this:
employerTypesummed to 111,700. It splits the board cleanly.employmentTypesummed to 119,226, over by 7,527, because an advert can be tagged full time and contract at once.workplaceTypessummed to 110,891, short by 808, because some adverts carry no workplace type at all.postedDatesummed to 33,215, and that number means nothing. Today sits inside Last 3 Days, which sits inside Last 7 Days.
Nothing on the page marks any of this. Four groups of counts that look identical, and only one of them can be charted as a breakdown of the board.
Every row carries sumsToTheSearchTotal, anAdvertCanHaveMoreThanOne, doesNotCoverEveryAdvert
and valuesAreNestedNotSeparate, so you can tell which you have before you add anything up.
Two counts you cannot filter on
dice.com publishes isRemote with 10,706 adverts and workFromHomeAvailability with 450. Both
come with counts. Send either one as a filter and the board comes back unfiltered, all 111,699
adverts, with no error.
workplaceTypes is the one that works. cannotBeFilteredOn marks the other two.
The 450 and the 10,706 are not a mistake in one of them either. They count different things, on the same board, on the same day.
Two totals
The board reports totalResults of 111,699 and totalJobCount of 122,744 on the same
response. Both are in the run summary. The facet counts are drawn against the first one, which
was checked by adding them up rather than assumed.
For a time series
Every row is stamped with takenAt, searchUrl and the filters in force, so snapshots of the
whole country and of one city sit in the same table without merging.
A facet that falls to zero keeps its row with a count of 0. A facet dice.com stops publishing has no row. A failed run writes nothing rather than rows of zero, so a gap in the series looks like a gap.
What it costs
$0.00001 per run and $0.00001 per row. A snapshot of the whole board is one request and about 21 rows, and that does not grow as the board grows, because the counts are published rather than counted.
Asking for a place costs one extra request. dice.com answers a place it cannot resolve with the whole country, so this run checks the total against the unfiltered board and stops rather than describing the United States under a label that says Austin.
No keyword search
dice.com's robots.txt disallows /jobs?q*, so this Actor never sends a keyword. Narrow the
snapshot with place, contract type, workplace, who is advertising, posted date or visa
sponsorship.