Funding Announcement Press Release Template and Statistics (2026)

Methodology

What the data is. The study covers published press releases distributed across 10 newswires. After the filters below, the sample holds 178,863 releases from 9 of those wires, published between September 2, 2021 and September 29, 2026. 99.7% of the sample was published on or after May 1, 2026, so the figures describe mid 2026 releases above all. This edition was updated on 30 September 2026 and replaces a smaller first pass that could only use releases carrying a stored publish date.

Recovering publish dates. 97,329 records in the snapshot have a body but no stored publish date. For these we read the first date written as month, day, and four digit year (for example September 29, 2026 or Sept. 29, 2026) in the first 1,000 characters of the release text, which is normally the dateline. English month names and their common abbreviations are recognized; other languages are left out. A recovered date is kept only if it is a real calendar date, falls on or before the day the record was stored, and is no more than 30 days older than that day. 93,223 dates (95.8%) were recovered this way. For the one wire whose web addresses carry the publication date, 94.6% of recovered dates matched that address exactly and 5.0% were one day off. Recovery rate by wire (wires are labeled by their share of the final sample).

Inclusion and exclusion. The starting snapshot held 188,912 records. We removed 5,742 with an empty body, 4,106 with no stored or recoverable date, 0 with a date outside January 1, 2021 to the snapshot date (September 29, 2026), and 201 with a body under 50 words. That leaves 178,863 releases, 85,642 dated by a stored publish date and 93,221 by a recovered date.

Weekday sample. Weekday figures use every release in the sample, dated by its stored publish date where present and by its recovered date otherwise. No wire supplies more than 41.2% of this sample. The headline weekday figures pool all releases. As a check, the same shares with each wire of at least 1,000 releases weighted equally are Monday 19.3%, Tuesday 22.0%, Wednesday 18.6%, Thursday 20.6%, Friday 13.8%, Saturday 3.0%, Sunday 2.8%.

Hour of day sample. Hour figures use only the 77,773 releases with a real publish time, meaning a stored publish time or a dateline time written with an explicit Eastern zone (4 releases). The time a record was stored is never used as a publish time. This sample is heavily skewed, because 90.5% of it comes from a single wire; the hour figures therefore describe that wire above all. With each wire of at least 1,000 time stamped releases weighted equally, the busiest hour is still 8 a.m. to 9 a.m. Eastern (16.0%).

How figures are aggregated. Every share is the number of releases with a feature divided by the number of releases in the group. Word counts strip any HTML and count whitespace separated tokens. Headline length counts visible characters after decoding HTML entities. Medians and percentiles use the nearest rank method. Stored publish times are in UTC and converted to US Eastern time with the America/New_York time zone rules, so daylight saving time is handled (UTC minus 4 hours in summer, minus 5 in winter). Records stamped exactly at midnight carry a date but no time, so they are left out of the hour of day figures and counted on their calendar date for the weekday figures.

How elements are detected. A quote is a quoted passage of at least 20 characters in a release that also uses an attribution verb such as said, says, added, or noted. A phone number is a North American ten digit pattern or an international number starting with a plus sign. A dateline is a place name followed by a full date within the first 600 characters. An About boilerplate is a line that starts with the word About followed by a capitalized name. An image is any attached image other than a generic wire placeholder, measured only for the 83.3% of releases whose records carry image information; an image whose file name contains the word logo is counted as a logo. An ALL CAPS headline has at least 10 letters and no lowercase letters. Duplicate headlines are identical after lowercasing and collapsing spaces.

Release types. Types come from transparent headline keyword rules, applied in the order below. A release takes the first type it matches, and anything that matches none is counted as other.

1. Investor litigation notice (excluded from the five types). A release is counted when its headline contains class action, shareholder alert, investor alert, securities fraud, securities litigation, lawsuit, lead plaintiff, investigation of, investigation on behalf, investors who, or investors with.
2. Acquisition. A release is counted when its headline contains acquires, acquired, to acquire, acquisition of, completes acquisition, completes the acquisition, merger, to merge, merges with, buyout, or takeover.
3. Funding. A release is counted when its headline contains raises, raised, secures, secured, closes, closed, lands, or nabs followed within 60 characters by a currency amount, million, billion, funding, seed, or Series A to F; or contains funding round, seed round, seed funding, pre-seed, a Series A to F round, venture funding, growth funding, growth equity, or strategic investment. Headlines that also mention a contract, order, award, dividend, guidance, outlook, deal, results, reports, earnings, gala, fundraiser, charity, or donation are not counted.
4. Executive appointment. A release is counted when its headline contains appoints, appointed, appointment, names, named, hires, hired, promotes, promoted, elevates, welcomes, joins, taps, elects, or elected, and also a role word such as CEO, CFO, COO, CTO, CMO, chief, president, vice president, VP, director, chair, head of, board, general counsel, or officer. Headlines about lists, awards, honors, finalists, or winners are not counted.
5. Partnership. A release is counted when its headline contains partners, partnered, partner with, partnership, partnering, teams up, team up, teaming, collaboration, collaborates, collaborate, alliance, or joins forces.
6. Product launch. A release is counted when its headline contains launches, launch, launched, launching, unveils, unveiled, unveil, introduces, introduced, introducing, debuts, debut, debuted, rolls out, rolled out, now available, or available now. This type covers product, service, and program launches.

1,170 releases (3.4% of those matching any of the five rules) matched more than one rule and were assigned by that order.

Limitations. Wires are not named. The largest wire supplies 41.2% of the full sample and 90.5% of the hour of day sample. Recovered dates come from the text of the release, so a release whose dateline differs from its actual distribution day is counted on the dateline day. Some wires republish releases that also appear on other wires, which adds cross wire duplicates. The detection rules are pattern based and will miss some elements and catch some false positives; for example, the word leading also matches ordinary uses such as leading to. Body text can include small amounts of non-release text such as navigation labels, which slightly inflates word counts. Release types use the headline only. Any figure based on fewer than 200 releases is flagged or left out.